VOL. 78/MAR. 1997 - Mitsubishi Electric · ferred to as Moore’s Law. Today, 64Mb DRAM devices are...

MITSUBISHIELECTRIC

VOL. 78/MAR. 1997

Automotive Electronics Edition

TECHNICAL REPORTS

FFFFFOREWORDOREWORDOREWORDOREWORDOREWORD

To the Special Edition on Automotive Electronics .................................. 1by Masanobu Tosa

OOOOOVERVERVERVERVERVIEWVIEWVIEWVIEWVIEW

The Present and Future Trends in Automotive Electronicsby Dr. Shoichi Washino ................................................................................ 2

A Car Navigation System with Traffic InformationCapabilities ................................................................................................. 4by Yoshisada Mizutani and Toshiro Sogawa

A Digital Audio Broadcast Receiver ......................................................... 7by Hiroaki Kato and Ken’ichi Taura

Vehicle Distance Interval Control Technology ....................................... 10by Kazumichi Tsutsumi and Shigekazu Okamura

A Drowsiness Detection System ............................................................. 13by Kenji Ogawa and Mitsuo Shimotani

Motor-Drive Control Technology for Electric Vehicles ......................... 17by Kazutoshi Kaneyuki and Dr. Masato Koyama

An Electric Power-Steering System ........................................................ 20by Takayuki Kifuku and Shun’ichi Wada

Technologies for Gasoline Engine Control andECU Miniaturization ................................................................................. 24by Yoshinobu Morimoto and Takanori Fujimoto

Starting and Charging Systems for Fuel-EfficientEngines ..................................................................................................... 28by Toshinori Tanaka and Akira Morishita

NEW PRODUCTS ................................................................................. 31

MITSUBISHI ELECTRIC OVERSEAS NETWORK

Mitsubishi Electric Advance is publishedquarterly (in March, June, September, andDecember) by Mitsubishi Electric Corporation.Copyright © 1997 by Mitsubishi ElectricCorporation; all rights reserved.Printed in Japan.

CONTENTS

Editorial Advisors

Editor-in-Chief

Akira Yamamoto

Jozo NagataToshimasa UjiTsuyoshi UesugiSatoru IsodaMasao HatayaGunshiro SuzukiHiroshi ShimomuraHiroaki KawachiAkihiko NaitoNobuo YamamotoShingo MaedaToshikazu SaitaHiroshi TottoriMasaaki Udagawa

Yukiyasu Watanabe

Masaaki UdagawaPlanning and Administration Dept.Corporate Engineering, Manufacturing &Information SystemsMitsubishi Electric Corporation2-3-3 MarunouchiChiyoda-ku, Tokyo 100, JapanFax 03-3218-2465

Tsuyoshi UesugiStrategic Marketing Dept.Global Operations GroupMitsubishi Electric Corporation2-2-3 MarunouchiChiyoda-ku, Tokyo 100, JapanFax 03-3218-3597

Orders:Four issues: ¥6,000(postage and tax not included)Ohm-sha Co., Ltd.1 Kanda Nishiki-cho 3-chomeChiyoda-ku, Tokyo 101, Japan

●Vol. 78/Mar. 1997 Mitsubishi Electric ADVANCE

Automotive Electronics Edition

Vol. 78 Feature Articles Editor

Editorial Inquiries

Product Inquiries

The background photograph for our coverwas taken at our laboratory simulatingnatural environments. This is used toassess safety factors. Fifty meters long, itcan simulate dense fog, torrential rain, andintense lighting under strictly controlledconditions

A Quarterly Survey of New Products, Systems, and Technology

TECHNICAL REPORTS

*Managing Director Masanobu Tosa is in charge of the Automotive Equipment Group.

Forewordto the Special Edition on Automotive Electronics

Masanobu Tosa*

Advances in electronics and communications technologies are respon-sible for making our society increasingly information oriented.Now that cars are being fitted with portable telephones and other

equipment for inter and intra-car communications, they are no longer largelyisolated from these trends and are becoming ever more closely integratedwith society, sharing in its information and communications infrastructure.This is typified by research into intelligent transport systems (ITSs). MitsubishiElectric is engaged in a wide range of ITS research and development projects,including semiconductor devices, in-car equipment, millimeter-band andoptical communications equipment, and the technologies for system opera-tion.

ITS-related equipment includes navigation systems, which are now com-ing into widespread use. Other important topics include the developmentand implementation of technologies for automating the actual processes ofdriving a car and improving safety. The corporation’s research and develop-ment for navigation systems is aimed at enhanced functionality, includingmaking them suitable for all geographical areas of the different countries,improved man-machine interfaces, etc. Again, in automatic driving andsafety technology, we are committed to R&D for the sensors and controltechniques that are essential for reliable systems that optimize the car’s basicperformance capabilities.

In a society that is so heavily dependent on the car, we cannot ignoreenvironmental concerns. Here, we can make a contribution not only bymanufacturing in-car equipment in factories that are operated responsiblybut also by making it smaller and lighter, and by optimizing the variouscontrol functions. ❑

March 1997 · 1

TECHNICAL REPORTS

*Dr. Shoichi Washino is with the Industrial Electronics & Systems Laboratory.

The Present and Future Trends inAutomotive Electronics

by Dr. Shoichi Washino*

Over the three decades from the 1960s to the1990s, the motor vehicle industry has gradu-ally introduced electronic technology to performmany of the control functions previously per-formed mechanically. This report surveys de-velopments in electronics technology and theirpractical applications.

BackgroundThree factors have stimulated motor vehiclemanufacturers to adopt electronic control in aincreasing number of motor vehicle systems:advances in semiconductor technology, morestringent environmental regulations, and userdemand for enhanced performance and conve-nience. In addition, the popularity of car navi-gation systems reflects the growing demand forinformation technology in the automotive en-vironment.

Table 1 shows a chronology of advances inautomotive electronics in the context of relatedtechnologies. Breakerless electronic ignitionswere first introduced in the 1960s and soonproved their reliability. Gradually electroniccomponents began to replace other mechanicalcomponents. The 1970s saw the introduction

of electronic fuel-injection systems and markedthe beginning of a rapid changeover from me-chanical to electronic control technologies.Electronic fuel injection showed that not onlycould electronic components substitute formechanical components but that electroniccontrol systems could radically alter the waythat motor vehicles perform.

In the 1980s, electronic fuel-injection systemswere supplemented by additional electronics forcontrolling ignition timing and exhaust gas recir-culation (EGR) systems. Electronic control sys-tems began to take on a comprehensive characterreflected in a new term: engine control systems.Manufacturers also started using electronics forABS and other chassis control applications. The1990s have brought car phones, navigation sys-tems and other information equipment into theautomotive environment.

Technological Advances in ElectronicSystems

FINE-PATTERN LITHOGRAPHY. Improvements insemiconductor processing technologies bring aquadrupling of device integration scale approxi-

Tec

hnol

ogy

Tre

nds

Aut

omot

ive

elec

tron

ics

Table 1 Chronology of Automotive Electronics Technology.1960 1970 1980 1990 2000

Events affecting Expo ’70 in Osaka First and second oil crises Bubble economy High yenJapan Automotive exhaust gas regulations Diesel exhaust gas regulations

Hardware DRAM 256kB 1Mb 16Mb 64Mb 1GbMicroprocessors 4-bit 8-bit 16-bit 32-bit 64-bit

Software Languages Fortran, Cobol PL/1 Pascal,ADA, C

Personal computer operating systems CP/M MS-DOSWindows 3.1 Windows 95

Communications Phones for First data transmission services Car phones, Cellular phones PHS Compact satelliteelectrical trains and digital telecommunications Fax machines, Packet transmission terminals

Information Vehicle information andcommunicationsystems for navigation

Chassis ABS Suspension, 4WD, Distance interval controlPower steering 4WS systems

Engine Electronic ignition Electronically controlled AT, Electronic combustion control,Electronic fuel injection Electronic valve timing control,

Per-cylinder knock control

Body Intermittent wipers Automatic air-conditioning controlDrive computer Keyless entry

1960 1970 1980 1990 2000

2 · Mitsubishi Electric ADVANCE

TECHNICAL REPORTS

mately every three years in a growth curve re-ferred to as Moore’s Law. Today, 64Mb DRAMdevices are commercially available, and labora-tories are already manufacturing samples of 1GbDRAMs. In practical terms, this higher inte-gration gives manufacturers the ability to offersophisticated control capabilities at a low costthrough use of application-specific integratedcircuits (ASICs) and micromachined semicon-ductor sensors. We can anticipate that manu-facturers will soon introduce system-on-chip(SOC) devices combining microprocessors,memory, software and peripheral circuitry in asingle device.

MICROSYSTEM TECHNOLOGIES. Microsystems in-corporate sensors, microactuators and micro-processors on a single chip, allowing entirecontrol functions to be implemented in a singledevice. Micromachining technology has alreadygiven us semiconductor acceleration and pres-sure sensors. Improvements in sensor accuracyand reductions in size and cost will follow asmicromachining and surface processing tech-nologies continue to evolve. The near future isexpected to bring intelligent sensor devices in-tegrating microprocessors, memory, ASIC tech-nology and sensor elements on a single chip.

MATERIALS AND DISPLAY TECHNOLOGY. Advancesin materials and processing technologies prom-ise to bring about significant developments inmonolithic microwave ICs (MMICs), room-temperature infrared sensors, blue-wavelengthsemiconductor lasers and optical amplifiers. Ad-vances are also expected in the liquid-crystal andplasma panels used to display information in carnavigation systems and other applications.

STANDARDIZED SOFTWARE. Faster microproces-sors and cheaper memory are giving embeddedsystem designers latitude to build more sophis-ticated systems. The high development costsof these systems remain as a significant ob-stacle. Future applications will most likely re-quire a combination of standardized andcustomized software to realize complex systemswith the reliability required of motor vehicle

control systems at an acceptable cost.

NETWORKING TECHNOLOGY. Active work on au-tomotive LAN standards is underway, withstandards for three automotive LAN systems(Class A, B and C) currently under development.We can expect that future vehicles will also belinked to wide-area networks such as theInternet through cellular phones, leased linesor other means. The tremendous popularity andutility of the global positioning system (GPS)suggests a promising future for other satellite-based applications.

Comprehensive Control Systems

INTEGRATED TRAVEL CONTROL SYSTEMS. We canenvisage a comprehensive travel control sys-tem integrating vehicle ECUs and travel moni-toring sensors over an automotive LAN. Suchsystems could be enhanced by sensors moni-toring roadway conditions and the distanceinterval between vehicles.

INTELLIGENT TRANSPORT SYSTEMS. In Novem-ber 1995, participants from Japan, Europe, NorthAmerica and other nations met in Yokohama,Japan, to discuss the future of intelligent trans-port systems (ITS) and other means to facilitatesafe, smooth transportation using intelligentcommunication between vehicles and the road-way. We can anticipate that the roadway infra-structure will eventually be enhanced withtelecommunication capabilities and corre-sponding equipment will be developed for mo-tor vehicles.

Mitsubishi Electric is conducting R&D on allthe key hardware, software and networkingtechnologies presented in this report. The cor-poration is committed to the development ofautomotive electronics for applications rangingfrom stand-alone components to comprehen-sive control systems as well as new-generationsystems with telecommunication capabilities. ❑

March 1997 · 3

TECHNICAL REPORTS

*Yoshisada Mizutani and Toshio Sogawa are with the Sanda Works.

Table 1 Media Used to Distribute VICS Information

Medium Data rate Remarks

FM multiplex broadcast Approx. 50KB/5 minutes per station Wide-area data distribution to all vehicles in broadcast service area

Radio beacons Approx. 8KB per beacon Installed along freeways to inform drivers of conditions ahead

Infrared beacons Approx. 10KB per beacon Installed above roadways to inform drivers of conditions ahead

A Car Navigation System with TrafficInformation Capabilitiesby Yoshisada Mizutani and Toshio Sogawa*

Mitsubishi Electric has developed a car naviga-tion system capable of displaying traffic infor-mation from Japan’s Vehicle Information andCommunication System (VICS). This articlereports on functions that overlay roadway mapswith symbols showing traffic jams and otherdata, and driver-guidance functions using sketchmaps and synthesized voice announcements.

Car Navigation Using VICSJapan’s Vehicle Information and Communica-tion System (VICS), in operation since April1996, transmits realtime information on trafficjams, accidents, roadway obstructions and otherpertinent phenomena to vehicles using a com-bination of FM multiplex broadcasts, radio bea-cons and infrared beacons (Fig. 1, Table 1). Thecar navigation system we developed superim-poses this information on the main map dis-play. The information is also available forindicating traffic jams and other obstructionson sketch maps of the planned route, and forsearching for alternative routes.

VICS provides three levels of information in-tended for car navigation systems of varyingtypes. Level 1 consists exclusively of text infor-mation. Level 2 consists of simplified mapsshowing traffic jams and obstructions.

Level 3 provides detailed maps and includesfour types of information. Traffic jam data showsthe precise location of jams and the correspond-ing transit delays. Accident data describe acci-dents and other phenomena that obstruct traffic,and indicate the severity of the obstruction.Parking information shows the locations andnames of roadway service areas and parkingareas, and their degree of crowding. Travel timedata reports the transit times between specificpoints.

Superimposed DisplayDrivers can manually request the display oflevel 1 and level 2 data. Level 3 data includes anumerical code indicating the roadway to whichit applies. This allows the car navigation sys-tem to selectively display information relevantto the vehicle’s planned or actual route.

Figs. 2 and 3 show how these data are includedin the map display. In Fig. 2, traffic jam, road-way obstructions and parking data are shownvisually. Traffic jams are shown by a chain ofarrowheads that indicate the length of the jamand the direction of traffic affected. The arrowsare shifted slightly to the left to avoid interfer-ing with the lines that indicate the roadwayand recommended route. The arrows alwaysappear beneath the recommended route for bet-ter legibility.

Fig. 3 shows a map display with a text win-dow for detailed accident and roadway obstruc-tion data. The driver can also call up detailedparking availability data. The side-by-side for-mat is easy to read and does not interfere withnormal use of the navigation system.

Guidance by Sketch Map and VoiceMost car navigation systems provide route guid-ance using a combination of enlarged intersec-tion drawings and synthesized voices. The draw-

Roadway observation

VICS center

Infraredbeacons

Radiobeacons

FM multiplexbroadcasts

Car navigation system

Data acquistion

Data processing andediting

Data distribution

Data utilization

Fig. 1 The VICS system. The arrows show thedirection of information flows.


TECHNICAL REPORTS

the demands of driving permit drivers to view thenavigation screen only briefly, simplified displaysare needed so that drivers can assimilate infor-mation rapidly and without strain. This led us toinvestigate drivers’ mental representations of road-way networks—which we refer to as cognitivemaps—and to design the user interface to presentinformation in a manner that is consistent withthese representations.

Cognitive maps are characterized by quanti-zation of direction and distance between sig-nificant points, representation of links betweenpoints by straight lines and the mental conver-sion of graphic images to propositional images.Figs. 5 and 6 show the simplified sketch mapswe developed for route guidance. Fig. 5 shows amap of an intersection where a turn is required,and Fig. 6 an overview of the planned route. Inboth cases only essential information is pro-vided. The direction and distance informationbetween points is quantized, and the path be-tween points is represented by straight lines.

The intersection maps usually show roadwayangles in increments of 45 degrees. A 30-degreequantization is used for more complicated in-tersections where 45-degree quantization wouldcause roadways to overlap. Intersections wherethe roadways are at a slight angle are representedusing right angles. This simplification is consis-tent with drivers’ cognitive mapping and improvesmap recognition.

The sketch maps are always oriented northup. Intersections are joined by straight linesusing quantized distances and the roadway

ings indicate the path of travel, the distance tothe next turn and nearby landmarks (Fig. 4). Since

Fig. 2 Traffic information superimposed on a map display.

Infrared beaconRadio beacon

FM multiplex broadcast

Parking full

Lane obstructed

Parking available

Roadway traffic jam

Freeway traffic jam

Fig. 4 Detailed drawing of a guidance point.

Fig. 3 Detailed information on obstructed lanes.

Lane Obstruction Alert

1 laneblocked

Subwayconstruction

March 1997 · 5

TECHNICAL REPORTS

Fig. 5 Sketch map of a guidance point. Fig. 6 Sketch map of a planned route.

angles near the intersections are maintained.Table 2 shows how the task of conveying in-

formation to the driver is distributed betweenthe screen and synthesized voice output. Thesystem first provides the driver with advancednotice of an upcoming turn. A synthesized voiceannounces the distance to the turn and the di-rection the driver will be required to turn. Thisreduces the mental workload by enabling thedriver to prepare for the action. As the driverapproaches the turn, the screen displays a sketchmap of the intersection and the synthesizedvoice advises the driver of the direction to turn.

The driver can also set up the system so thatinformation requests are answered by both syn-thesized voice and screen display.

Through skillful integration of roadmaps, super-imposed traffic information, simplified maps atturns on the planned route and synthesized voiceannouncements, this car navigation system pro-vides drivers with information in easily assimi-lated form with a minimal mental workload.Further development is planned to enhance thesebenefits for improved driving safety and conve-nience. ❑

Table 2 Information Content of Sketch Maps and Synthesized Voice Guidance

Time Screen Voice

About 1km before turn No display

1. Location specifier: transit, destination, ferry port,freeway on-ramp, freeway off-rampExample: ‘Nearing transit point. Take freeway off-ramp’

2. DirectionExample: ‘Left turn’

3. DistanceExample: ‘1km ahead’

Immediately before turn

1. Location specifiers: intersection shape, intersection 1. Location specifier: transit, destination, ferry port,name, landmarks freeway on-ramp, freeway off-ramp

2. Route identifier Example: ‘Nearing transit point. Take freeway off-ramp’

3. Vehicle position indicator 2. Direction

4. Numerical distance-to-turn display Example: ‘Left turn’

Table 2 Information Content of Sketch Maps and Synthesized Voice Guidance


TECHNICAL REPORTS

A Digital Audio Broadcast Receiver

by Hiroaki Kato and Ken’ichi Taura*

*Hiroaki Kato is with the Sanda Works and Ken’ichi Taura with the Imaging Systems Laboratory.

FM broadcasts are subject to multipath inter-ference that causes impairment of sound qual-ity when the signal is received by a movingvehicle. To solve this problem and provide au-dio broadcasting of compact disk quality, theEuropean Eureka project developed a digital au-dio broadcasting (DAB) system. DAB publicservices have been operational and on air in theUK since September 1995, and are expected tospread widely throughout Europe by fall 1997.

The DAB SystemThe DAB system uses ISO/MPEG1 audio cod-ing for source coding and orthogonal frequencydivision multiplexing (OFDM) for modulation.

MPEG1 is a highly efficient source codingtechnique. It allows bit rate reduction by uti-lizing a psychoacoustic model of the humanear, thus preserving the subjective quality ofthe audio signal. The audio frequency range isdivided into 32 sub-bands of equal width. Everysub-band is allocated a bit capacity based onthe difference between the maximum signallevel and audible threshold, which is calculatedfrom the threshold of audibility at silence andthe masking threshold.

OFDM is a multicarrier scheme which trans-fers high-speed data using a number of orthogo-nal carriers with tight frequency spacings. Whencombined with error correction code, it is calledcoded OFDM (COFDM). Each carrier is modu-lated by π/4 differential quadrature phase shiftkeying (DQPSK). Even if some of the carriersare degraded by frequency-selective fading dueto multipath propagation, the data of these car-riers can be recovered from the data of the re-maining carriers by performing error correction.

Fig. 1 shows the transmission frame formatof the DAB system. Each frame consists ofmultiple symbols. The first frame is a null sym-bol which is the duration of no radio frequency(RF) signal transmission, and is used for framesynchronization. The second frame is a phasereference symbol for QSPK demodulation,which is followed by two symbol groups. Oneis a fast information channel (FIC), and the otheris a main service channel (MSC). The FIC trans-mits program related information and the multi-

Transmission frame

Synchro-nizationchannel

Fastinformationchannel (FIC)

3 symbols

Main service channel(MSC)

72 symbols

Phase reference symbol

Null symbolGuard interval

OFDMsymbolduration

∆t Ts

Fig. 1 Transmission frame.

plex configuration information of the MSC. TheMSC transmits audio programs and various dataservice information.

Each symbol consists of a useful symbol du-ration (Ts) and a guard interval (∆t), in which,part of the time, a signal of Ts is cyclically re-peated. As long as the multipath propagationdelays do not exceed the duration of the inter-val, no inter-symbol interference occurs and nochannel equalization is required. This enablesa broadcast network to be extended virtuallywithout limitations by operating all transmit-ters on the same radio frequency; known as asingle-frequency network (SFN).

Table 1 lists four transmission modes avail-able in the DAB format, each of which supportsa bandwidth up to 3GHz.

Mode I Mode II Mode III Mode IV

No. of radiated1,536 384 192 768

carriers

Frame duration 96ms 24ms 24ms 48ms

Null symbol1,297µs 324µs 168µs 648µs

duration

Useful symbol1,000µs 250µs 125µs 500µs

duration

Guard interval 246µs 62µs 31µs 124µs

Nominal freq.

375MHz 1.5GHz 3GHz 1.5GHzrange (formobilereception)

Suitable Single-freq. Terrestrial/ Satellite Terrestrialcommunication terrestrial satellite communi- communi-applications networks communications cations

cations

Table 1 General DAB Specifications

March 1997 · 7

TECHNICAL REPORTS

A DAB ReceiverFig. 2 shows a block diagram of a DAB receiverdeveloped by Mitsubishi Electric. The RF blockcan receive both VHF and L-band signals, andconverts them to intermediate frequency (IF)signals. The IF signals are converted to digitalsignals by an A/D converter (ADC). The digitalsignals are then converted to in-phase andquadrature components and fed into the OFDMdemodulator. Here, the orthogonal carriers arederived from the time signal on a symbol basisby means of a fast Fourier transform (FFT) pro-cessor. After this, the data is reproduced by aDQPSK modulator from each carrier. A Viterbidecoder executes the time and frequency inter-leaving and error correction for the reproduceddata. An MPEG1 audio decoder expands thecoded audio data to linear PCM audio data,which is then converted to analog audio sig-nals by a D/A converter (DAC). A 16b general-purpose digital signal processor (DSP) is utilizedfor the synchronization process of the receivedsignal to fulfill the requirement of frequencyand timing accuracy.

FFT is executed on a useful symbol duration.Each carrier has a peak position where all oth-ers have zero-crossings if a carrier-spacingequals 1/Ts (=fo). This orthogonal relationshipprevents inter-carrier inteference.

Frequency offset of a local oscillator of a RFblock disturbs the orthogonal relationship. Thiscauses degradation of the carrier-to-noise (c/n)ratio due to inter-carrier interference, and re-sults in degradation of the error rate. We mea-sured the bit error rate in relation to thefrequency offset of local oscillators. In the caseof Mode I transmissions, the bit error rate isdegraded markedly at an offset larger than 50Hz.Considering this, our objective was a frequencysynchronization accuracy of 25Hz, which cor-responds to 0.1ppm precision in a 220MHz sig-nal.

Frequency synchronization is generally con-trolled as follows. The local oscillator is adjusted

Antenna

RF ADCI/O

genera-tor

OFDMdemodulator

Viterbidecoder

Audiodecoder DAC

Tuning

AFCDSP

Micro-computer

Audio output

KeyADC Analog to digital converterAFC Audio frequency controlDAC Digital to analog converterDSP Digital signal processingI/Q In-phase/Quadrature-phaseOFDM Orthogonal frequency division

multiplexingRF Radio frequency

Fig. 3 Constructive contribution of echoes.

Fig. 2 Block diagram of a DAB receiver.

to the frequency at which the phase differencebetween the phase of received data after de-modulation and the possible phase point of re-produced data by the receiver duringsynchronization becomes zero. However, if thephase difference due to frequency offset exceedsπ/4, an alias-lock on the frequency with errorscorresponding to a multiple of ±π/2 may occur.We therefore investigated control methods uti-lizing information related to received signallevel and phase of each carrier, which is ob-tained by FFT execution on the received signal.

As shown in Fig. 3, when a direct wave andreflective waves due to multipath propagationare received, intersymbol interference will oc-cur in the signals received by the RF block. Inthe DAB system a guard interval is insertedbetween successive symbols, enabling the re-production of signals free from multipath propa-gation by executing FFT for the duration of nointersymbol interference.

In DQPSK modulation, data is transmittedby the relative phase difference between adja-cent symbols. Therefore, FFT must be executedin the same position of each symbol duration.

These requirements can be fulfilled by execut-ing FFT from the end point of the guard inter-vals of the direct wave. We investigatedcalculating the reference point of the direct wavewithin the symbol duration from the impulse

∆t Ts ∆t

Symbol k Symbol l Symbol m

Echo 1

Echo 2

Echo 3

Inter-symbolinterference

FFTwindow

k l mm

m

m

m

l

l

l

k

k

k


TECHNICAL REPORTS

response of the transmission channel.

Field TestsWe carried out field tests in the UK (London)

for the purpose of verifying the reception per-formance of the newly developed DAB receiverand FIC data collection. Fig. 4 shows BBC trans-mitter locations and the radiation power inLondon and its suburbs. In the moving vehicletest around Surrey Research Park in Guildford,interference-free sound reception was con-firmed under the electric field conditions of Fig.5a. We also confirmed interference-free recep-tion under a multipath environment. In suchenvironments, reception of the same programsusing other FM broadcast receivers suffered frommultipath noise. We could therefore confirmthe superiority of the DAB system.

We also conducted a reception test at Dorking,midway between the Reigate and Guildfordtransmitters, to test reception at the fringe ofthe SFN transmitter service area, where fieldstrength becomes poor due to interferencecaused by each transmitter’s signal (Fig. 5b). Thefield strength was –80dBm, –90dBm whenpicked up on a receiver in a moving vehicle athigh speed. Under such conditions, the receivercould reproduce an audio program without au-dible noise. Through these tests, we confirmedthat a mobile DAB receiver requires a sensitiv-ity of –96dBm to cover the SFN service area thor-oughly.

During the above tests, the following FIC datawere collected; MSC configuration data, labeland language information, date and time infor-mation (including local offset), reconfigurationdata of service and subchannel organization.

We have developed a prototype DAB receiverand confirmed the superiority of DAB systemsthrough field tests and subsequent mobile re-ception tests. Considering the results of thesetests, utilizing BBC regular services, we hopeto develop a widely acceptable commercial DABreceiver with superior reception sensitivity anda variety of functions. ❑

Name Location Approx. power

Crystal Palace South Central London 4kW ERPBluebell Hill East of London 1kWAlexandra Palace North London 1kWReigate Hill South London 1kWGuildford Surrey (South) 1kW

Fig. 4 Transmitters in the BBC’s DAB network inLondon.

a) Spectrum waveform at Guildford

Fig. 5 Spectrum waveform measurements.

March 1997 · 9

b) Spectrum waveform at Dorking

Reigate Hill

Crystal Palace

Alexandra Palace

Bluebell Hill

Guildford

TECHNICAL REPORTS

Vehicle Distance IntervalControl Technology

by Kazumichi Tsutsumi and Shigekazu Okamura*

*Kazumichi Tsutsumi and Shigekazu Okamura are with the Himeji Works.

Technical advances can make the task of driv-ing a vehicle safer and more pleasant. This re-port introduces a distance interval controlsystem for vehicles that supplements the capa-bilities of conventional cruise control. The sys-tem uses a laser radar system to identifypreceding vehicles, processes images from aCCD camera to identify one’s own lane of travel,finds the closest vehicle in the lane and thencontrols the throttle to establish a safe follow-ing distance. The system combines two thor-oughly understood technologies: active laserradar and passive image processing.

ConfigurationThe system, illustrated in Fig. 1, consists of alaser radar system, a CCD camera, sensors forspeed, steering wheel angle and other automo-tive parameters, a throttle actuator, an electroni-cally controlled automatic transmission, analarm buzzer and a controller.

Laser Radar HeadFig. 2 shows the laser radar head’s scanningmechanism. On receiving a signal from thecontroller, a motor-driven cam scans a mirrorback and forth under microprocessor control.Light emitted by the laser diode is directed by afixed mirror to the scanning mirror and pro-jected forward from the car creating a fan-shapedbeam. Reflectors on vehicles and other objectsreturn the beam through the receiving lens to aphotodiode. The microprocessor measures thetime required for the reflection to return anduses that information to compute the distanceinformation, which it sends to the controller.

A dirty front windshield, rain, snow or waterkicked up by the preceding vehicle can all de-grade the performance of a laser radar system.We studied these effects in a roadway environ-ment simulation facility (Fig. 2) and installed asensor near the laser diode that detects a dirtywindshield and reports this to the micropro-cessor.

Preceding Vehicle Detection AlgorithmA single scan by the laser radar system collectsdistance data in 80 directions. Reflectors on cars

Camera

AlarmThrottleThrottle

Automatictransmission

Controller

Laserradar

Sensors

(vehicle speed,steering wheelangle, etc.)

and trucks or along the roadside result in nu-merous adjoining data points. Several steps arerequired to resolve this data into individual re-flectors and then to identify vehicles. Fig. 4 il-lustrates this process.

The solid boxes in the figure show wheremeasurement points have been grouped to iden-

Receiving lens

Moving reception mirrorCam Laser diode

Transmission lens

Fixed mirror

Transmission lens

Moving transmission mirror

Reception windowPhotodiode

Fig. 2 Laser radar optics.

Fig. 1 The system configuration.

Fig. 3 Natural environment simulation facility.


TECHNICAL REPORTS

tify individual reflectors. Next, position andmovement data are used to associate reflectorsinto larger groups that correspond to vehiclesor other objects. Steering displacement data isthen used to identify other vehicles travelingin the same direction.

The next stage of processing uses lane infor-mation extracted by processing an image fromthe CCD camera. Vehicles in the lane of travelare identified, and the distance to the nearest isreturned as the vehicle distance interval.

Lane Detection AlgorithmThe image processing system analyzes CCDcamera images of the roadway to identifypainted lane markers. These lines may be soiled,worn away or completely absent. The systemdeals with these possibilities by first identify-ing edges in the image, and then examining theedges to identify lines marking the lane of travel.The detection algorithm identifies the lane oftravel by searching for boundaries that are stablewith respect to time for vehicle speeds over acertain threshold. The lane width is used asauxiliary criteria, since lane widths tend to fallwithin a particular range.

Once the left and right lane edges are found,

the region in between them is known, makingit possible to determine if a particular vehicleis in one’s own lane of travel. Fig. 5 shows aprocessed image used for lane marker detection.

ActuatorsThe throttle actuator, shown in Fig. 6, consistsof an electric-motor-driven diaphragm vacuumpump, two solenoid valves and a vacuum-driventhrottle actuator. The vacuum generated by thediaphragm pump is supplied to the actuator’svacuum chamber. The electric motor and theswitching of the two solenoid valves controlthe throttle position.

Vehicle Distance Interval ControlThis section describes several operationalscenarios.

If no preceding car is detected, vehicle speedis determined by the cruise control, which

March 1997 · 11

Precedingvehicle

Reflectiveobject

Scanningangle

Own vehicle

Preceding vehicle

KeyDetection pointGroupVehicle candidate

Fig. 4 The vehicle detection algorithm.

Throtttle valve

Pulley assemblyWire cableVacuum hose

Diaphragm pump

DC motor

Accelerator pedal

Releasevalve

Controlvalve

Fig. 6 Accelerator actuator.

Fig. 5 Lane detection.

TECHNICAL REPORTS

maintains the vehicle at a driver-selected speed.As the car approaches another vehicle, the

vehicle speed is used to determine an appropri-ate following distance, and the throttle andtransmission are adjusted to establish and main-tain this distance. Initial control parameters arebased on the vehicle speed and relative speedsof the two vehicles. If the car is traveling at100km/h and is approaching another car at arelative speed of 20km/h, throttle-controlleddeceleration begins at a distance of about 85m.Deceleration continues until the speeds of thetwo vehicles are matched at 80km/h and a 45mfollowing distance is established. If the relativespeed of the two vehicles is excessively high,that is, if the car is approaching another car tooquickly to be handled by throttle control, thethrottle will close and an alarm buzzer willsound to signal the driver to brake.

If the preceding car moves out of the way, thesystem returns control to the cruise control andthe speed returns to the preset value.

This vehicle distance interval control systemsupplements the capabilities of conventionalcruise control. We believe that the system willfind wide acceptance among drivers because thetwo systems work well together, and becauseoperation is substantially the same as that for a


Target distance

Distance

Relative speed

Vehicle speed

Throttle actuatorTargetspeed Cruise

control

Fig. 8 Vehicle speed control logic.

cruise control. In the future, smoother controlwill be achieved through closer integration withthe cruise control system. These improvementswill require continued advances in the tech-nologies for sensing the vehicle environment. ❑

Dis

tanc

e In

terv

al (

m)

100

50

85m

45m

Relative speed20km/h

Relative speed0 km/h

Vehicle speek (km/h)

Fig. 7 Initial control.

TECHNICAL REPORTS

*Kenji Ogawa is with the Himeji Works and Mitsuo Shimotani with the Industrial Electronics & Systems Laboratory.

by Kenji Ogawa and Mitsuo Shimotani*

Systems that detect when drivers are becomingdrowsy and sound a warning promise to be avaluable aid in preventing accidents. This ar-ticle reports on a drowsiness detection systemdeveloped at Mitsubishi Electric. The systemanalyzes facial images of the driver to deter-mine blinking behavior, which it uses as ameasure of driver alertness. An accuraterealtime drowsiness detector prototype wasmanufactured and tested in actual vehicles.

Waning Alertness ExperimentsWe studied the relationship between alertnessand blinking behavior by testing ten subjectsin the driving simulator illustrated in Fig. 1.The subjects had to drive through nighttimefreeway scenery, and while doing so, underwenttwo tests. The first, a tracking test, was to con-trol the steering and accelerator to keep a cur-sor aligned with the image of the precedingvehicle. The second, a reaction time test, wasto press a button on the dash the moment a redsquare appeared on the screen.

A VCR recorded changes in blinking behav-ior as the subjects grew drowsy. Also monitoredwere the subjects’ brainwaves, eye movementsand other physiological indicators. The subjectsevaluated their own subjective state of alert-ness by pushing buttons for slightly sleepy,moderately sleepy and very sleepy. The resultsof the reaction time test were also recorded.

Test Results and AnalysisWe tested ten subjects (eight male, two female)ranging in age from 21 to 59. Nine of the tensubjects experienced waning alertness.

We analyzed the relationship between alert-ness and blink duration. Fig. 2 shows the rela-tionship between decreasing alertness and blinkduration. For each minute, the figure shows ahistogram for various blink durations, reactiontimes and subjective drowsiness evaluations.Long blink durations became more frequent asthe subjects began to report feeling drowsy. Atthe same time, alertness waned, causing reac-tion times to extend to 2s and longer. Table 1

Facial image

Telemetrytransmitter

Telemetryreceiver

DAT data recorder

VCR VCR

System controller(Generate test and superimpose

images)

Test signal

Timing signals

Screen image

Image with cursorand mark Driving simulation

controller

Speaker

Steering angle and throttleopening signals

Drivingnoise

Precedingvehicle

Cursor

IR LED, CCD camera

Switch for reactiontime test

Switch signal

EOG

EEG

ECG

EDA

Respiration

Projection screen

KeyCCD Charge-coupled device EEG ElectroencephalogramDAT Digital audio tape EOG ElectrooculogramECG Electrocardiogram IR InfraredEDA Electrodermal activity

Fig. 1 The laboratory test apparatus.

A Drowsiness Detection System

March 1997 · 13

TECHNICAL REPORTS

is assumed to be alert, and a threshold for judg-ing long blink durations is established. (2) Thedegree of alertness (α) is periodically calculatedas follows:

α = number of long duration blinkstotal number of blinks .............. (Eq. 1)

The driver is judged drowsy when α falls belowa specified threshold.

Blink DetectionWe used a small CCD camera mounted in thevehicle instrument panel to capture images ofthe subject’s face while driving. We had to sur-mount several hurdles in developing an effec-tive method to measure blink duration. The firstwas dealing with rapid changes in solar illumi-nation during daytime driving. Sometimes adriver’s face is illuminated by direct sunlight;sometimes parts of the face are shadowed bypillars or a sun visor; and these illuminationpatterns change quickly. Second, artificial illu-mination must be provided in tunnels and dur-ing nighttime driving.

The nighttime illumination problem was re-solved by using an invisible infrared lightsource. We rejected this approach for daytimeuse because the great number of LEDs requiredto provide appropriate daytime illumination lev-

Table 1 Elapsed Time Until Various DrowsinessIndications (min)

Subject Blink Subjective evaluationReaction time

Sex Ageduration Slightly Mod. Very over 2sover 0.5s sleepy sleepy sleepy

M 21 10 11 13 33 22

M 25 7 10 15 19 16

M 29 10 9 12 — 12

M 37 9 5 9 13 12

M 43 13 18 34 37 50

M 44 — 8 12 18 44

M 45 — — — — —

M 59 5 3 8 — 12

F 23 25 56 — — —

F 29 11 5 11 31 35

lists the test data, from which we drew the fol-lowing conclusions:

The average blink duration for alert subjectsvaries with the individual. Long blink durationsof a half-second or more correspond to subjec-tive evaluations of slightly or moderately sleepy.The one subject who remained completely alertrecorded no blink durations over a half-second.Blink durations of a half-second or longer werealways present when reaction times over twoseconds were recorded.

Alertness Inference AlgorithmKnowing that average blink durations differ foralert subjects, and that longer blink durationsare associated with decreasing alertness, led usto develop an alertness inference algorithmbased on the following principles: (1) Statisticson blink duration are collected when the driver


Fig. 2 Results of analysis for subjects

No.

of s

witc

h pu

shes

Time (min)Time (min)

Rea

ctio

n tim

e (m

s) 4,000

3,000

2,000

1,000

010 20 30 10 20 30

0

1

2

3

2520

1510

5

0

5,03

3 -

2,03

3 -

5,00

01,

033

- 2,

000

867

- 1,

000

633

- 66

756

7 -

600

500

533

433

- 46

7

300

- 33

336

7 -

400

167

- 20

023

3 -

267

100

- 13

333

- 6

7

a) Histogram of blink behavior for a subjectwho feels drowsy.

700

- 83

3

0

30

20

10

40

50

60

Number of blinksper minuts

Elapsed time(min)

Blink duration (ms)

c) Subjective evaluationsb) Reaction times

TECHNICAL REPORTS

els would be expensive, would increase powerconsumption and might present a potential vi-sion safety hazard. Instead, we explored digitalfiltering methods that generate a reliable im-age of the driver’s eyes from a facial image cap-tured under variable daytime illumination.

DAYLIGHT BLINK DETECTION. Fig. 3a shows ablock diagram of the image filtering system.First applied is a tophat filter that removes darkregions that are short in the vertical direction.When a tophat filter is applied to the cameraimage of Fig. 3b, the eyebrows, eyes and nos-trils are removed, yielding the image of Fig. 3c.Fig. 3d, the difference between Figs. 3b and 3c,shows the eyebrows, eyes and nostrils. Abinarization filter applied to Fig. 3d yields Fig.3e. The filtering not only eliminates the shad-owing visible in Fig. 3b but also eliminates hair,background and other unnecessary details.

The binary image is then analyzed to moni-tor blink duration. Fig. 4a shows a steeper in-clination of the eye corners with the eyes openthan in Fig. 4b when the eyes are closed. Mea-suring this quantity gives a robust indication

of the state of the eyes that is largely unaffectedby changes in distance between the driver andcamera.

NIGHTTIME BLINK DETECTION. Under low-illu-mination conditions, the reflection from theretina can be monitored to determine whetherthe eye is open or closed. Infrared LEDs alignedwith the optical axis of the camera lens illumi-nate the driver’s face (Fig. 5a). Infrared lightenters the pupil and is reflected by the retina(Fig. 5b). This reflection is captured by the cam-era (Fig. 5c). The driver’s face is dim, but thereflection from the pupil is bright. These brightreflections are absent when the eyelids are

a) Image filter circuit

Camera Tophat filter BinarizationBinary image

b) Input image c) Output of tophatfilter

d) Difference image e) Binary image

Fig. 3 Image filtering process.

a) Steep inclination ofeye corner wheneyes are open

b) Shallow inclinationof eye corner wheneyes are closed

Fig. 4 Method of daytime blink detection.

March 1997 · 15

TECHNICAL REPORTS

closed. Low-intensity illumination is sufficientto detect these reflections.

Charge-coupled devcice Infrared LED

Visible-light filtera) Camera

Pupil Eye

Reflection at retina

b) Retina reflection

c) Retina reflections visible in nighttimedriving image

Fig. 5 Method of nighttime blink detection

By using blink duration as a indicator of driveralertness, the electronic equipment and algo-rithms presented in this report provide a foun-dation for manufacturing accurate and inex-pensive drowsiness detectors. Future studiesusing vehicle-mounted prototypes will be usedto study the blink behaviors of a larger sampleas well as methods to support driver alertness. ❑


TECHNICAL REPORTS

Motor-Drive Control Technology forElectric Vehicles

by Kazutoshi Kaneyuki and Dr. Masato Koyama*

*Kazutoshi Kaneyuki is with the Himeji Works and Dr. Masato Koyama with the Industrial Electronics & SystemsLaboratory.

Because they emit no exhaust gases, electricvehicles are increasingly attractive to govern-ments and industries concerned with reducingatmospheric pollution. Mitsubishi Electric isapplying its experience in electric motors andpower electronics to develop drive systems forhigh-performance electric cars. This report de-scribes recent developments in this area.

Traction MotorsWe chose three-phase cage-rotor inductionmotors as best suited to electric vehicle driveapplications. Induction motors are cost-effec-tive, and are suitable in terms of size and weight,speed of rotation, efficiency, controllability andreliability. The durable rotor and high-speedoperation under easily implemented field-weak-ening control enable the use of large reductiongear ratios that limit maximum torque require-ments, allowing the incorporation of smallermotors driven by compact low-current invert-ers.

We selected water cooling over forced-airmotor cooling because it offers superior cool-ing performance and supports compact, light-weight motor designs. Our analysis showed thatuse of water cooling permitted weight reduc-

Fig. 2 Construction of the controller.Fig. 1 Construction of an induction motor for

electric vehicle use.

Water jacket Stator Rotor

Water jacket outlet Water jacket inlet

Smoothing condenser

Circuit boardBusbarIntelligent powermoduleHeatsink

March 1997 · 17

tions of 20% and size reductions of 30% as com-pared to forced-air designs, while the powerconsumption for cooling dropped by 75%. Useof a single water-cooling system for the motorand the inverter’s power module permitted fur-ther size reductions. Another advantage of wa-ter cooling is that the motor can be sealed morecompletely, which facilitates dust- and weather-resistant design.

Fig. 1 shows the construction of the induc-tion motor that we developed for this applica-tion. The stator features a low-loss core materialand highly heat-resistant insulators. High-den-sity windings are used with lightweight hous-ing materials that yield motors one-thirdsmaller and lighter than general-purpose induc-tion motors of the same output. The water-cool-ing unit was placed on the upstream side of themotor.

Motor-Drive ElectronicsThe drive electronics consist of an inverter thatdrives the motor and a controller that controlsthe inverter’s output voltage waveform. The sys-tem was specially developed to satisfy the spe-cific requirements of electric vehicles.

An inverter generally includes power devices,a drive circuit, a protection circuit, a smooth-ing condenser and a snubber circuit. We sim-plified the inverter by using high-performanceintelligent power modules (IPMs) based on in-

TECHNICAL REPORTS

Fig. 4 Motor efficiency curves.

iqr and φdr which suit the torque command.Fig. 3 shows the block diagram of a vector

control system implementing this algorithm.Commands for iqr and φdr are computed basedon the maximum efficient drive conditions andthe torque command determined by the fluxand current command calculator. The second-ary flux vector estimator and primary voltagecommand calculator then perform vector con-trol to ensure that ids and iqs follow the corre-sponding commands.

Fig. 4 compares the efficiency of the new vec-tor control system at various torque and speedranges against that of a conventional controlsystem. The efficiency improvement is espe-

sulated-gate bipolar transistors (IGBTs) thatintegrate the power devices, drive circuit andprotection circuit. We eliminated the snubbercircuit by positioning the flat busbar andsmoothing condenser to maximize noise can-celing effects. We were also able to substan-tially reduce the size of the IPM by closelymatching the characteristics of the motor andIPM.

The controller includes a torque commandprocessor that computes the optimum torquecommand for current driving conditions, and amotor control unit that supplies the three-phasemotor-drive voltage for maximum efficiencybased on the torque command, voltage and cur-rent values. The controller is implemented as asingle-chip microprocessor with motor and con-trol constants, and other parameters carried inflash memory. This design allows the use of astandardized controller in a variety of applica-tions.

Taken together, these design choices allowedus to reduce the size of the drive electronics by50% and the weight by 45% as compared tocontrollers employing discrete IGBTs.

Highly Efficient, Fast-Response VectorControl TechnologyWe employed a proprietary slip-frequency-con-trolled vector-control algorithm to implementfast-response vector control for the inductionmotor. This method, which employs a modelof the primary circuit of the induction motor,infers the secondary-circuit flux vector from themotor speed and primary current, then conductsfeedback control on the motor’s magnetizingcurrent and torque current using the d-q coor-dinates that rotate synchronously with the fluxvector. This design results in smoother motoroperation with fast torque-control response atspeeds ranging from a standstill to the motor’smaximum rating. The vector-control algorithmhas been proven in induction motor drive sys-tems for steel rolling mills and numerous otherapplications requiring precise torque control andfast response.

The motor-drive conditions for maximum ef-ficiency are determined as follows. From themotor circuit equation, we know that the ratioof the q-axis component of the secondary cur-rent (iqr) to the secondary flux vector amplitude(φdr) matches a specific variable determined bythe motor constants and the primary frequency.We also know that the induction motor outputtorque (τm) is proportional to the product of iqr

and φdr. Now, while satisfying the above maxi-mum efficiency conditions, we can select the

Flux andcurrent

commandcalculator

Primarycurrent

commandcalculator

Primaryvoltage

commandcalculator

PWMinverter

Secondaryflux vectorestimator

Induction motor

Tm*φdr*

idr*

φdr^ω

ids*

iqs*

θ ids iqs

Vus*Vvs*Vws*

ius

ivs

ωm

Fig. 3 Block diagram of a highly efficient vectorcontrol system.

a) Highly efficient vector control

b) Conventional vector control

Output torque (p.u.) Speed (rpm)

100

Eff

icie

ncy

(%) 80

60

40

20

02.0

1.51

0.50

5,00010,000

15,000

Output torque (p.u.) Speed (rpm)

100

Eff

icie

ncy

(%) 80

60

40

20

02.0

1.51

0.5 5,00010,000

15,000

0


TECHNICAL REPORTS

cially pronounced in the low-torque range,where an improvement of 35 percentage pointswas recorded. Since vehicles are frequently op-erated in the low-torque range, this improve-ment promises to bring a significant increasein cruising range. High efficiency means thatless power is dissipated in the motor as heat,which permits higher motor output and smallermotor dimensions.

Sensorless Speed Control TechnologyUse of vector control makes it possible to real-ize fast-response torque-control characteristics,however a weakness of vector control is that itrequires motor speed information, which is gen-erally provided by an external sensor. Numer-ous researchers are investigating sensorlessvector-control solutions [that promise greateraccuracy and reliability]. Recent advances inmicroprocessor performance and control tech-nology have led to sensorless vector-control sys-tems with improved performance that are nowbeing applied to general-purpose inverter-con-trolled motor-drive systems.

However another problem—poor torque linear-ity during low-speed regenerative braking—hasremained. In practical terms, this means that thecontrol system is unable to apply sufficient re-generative braking torque at low speeds.

We addressed this issue by developing a sec-ondary-circuit flux vector identifier employingmodel referenced adaptive control (Fig. 5). Thisidentifier matches the secondary flux vector(computed using a reference model of the in-duction motor’s primary circuit) to the second-ary flux vector identifier value (computed fromthe secondary-circuit model) using an adaptivealgorithm to adjust the motor speed (ωm) tothe value required by the secondary circuitmodel.

By experimenting with the adaptive algo-

Vds*

ω̂r

Vqs*ids

iqs

Reference model

Primary circuitmodel

Adaptivealgorithm

Secondarycircuit model

ω

φdr

φqr

φqr

φdr

^

^

θdt

+–

+–

2.0

1.0

–1.0

–2.0

–1.0–2.0 2.01.0

Output torqueτm (p.u.)

Torque command τm (p.u.)

rithm, we were able to improve the accuracy ofsecondary flux vector prediction under low-speed regenerative conditions. By replacing thesecondary flux vector estimator in Fig. 3 withthis identifier, we achieved the required perfor-mance of a sensorless vector-control system.Fig. 6 shows the torque-control characteristicsof this system under low-speed regenerative op-eration, indicating that torque linearity ismaintained.

Electric vehicles promise to lower atmosphericpollution by dramatically reducing vehicle emis-sions. Ongoing industry-wide R&D will gradu-ally bring the improvements in performance, costand convenience required for wide acceptance ofthis modern transportation technology. Mitsubi-shi Electric plans to continue researching anddeveloping motor-drive and other related tech-nologies for this application. ❑

Fig. 6 Torque control characteristics at 2% ofrated speed on a 200V, 60Hz, 1.5kW testmotor.

Fig. 5 Block diagram of a secondary flux vectoridentifier.

March 1997 · 19

TECHNICAL REPORTS

An Electric Power-Steering System

by Takayuki Kifuku and Shun’ichi Wada*

*Takayuki Kifuku and Shun’ichi Wada are with the Himeji Works.

Electric power-steering systems for larger ve-hicles have not been mass produced becausethe high-output motors required for this appli-cation add inertia and friction that cause anunacceptable loss of road feel. This report in-troduces a new motor control system that elimi-nates this problem.

Problems in High-Output Electric SteeringSystemsWhen a high-output electric motor is used toprovide steering assistance, the motor’s largeinertia and friction torque become significantwith respect to the tires’ self-aligning torque.The steering wheel return is sluggish and roadfeel suffers. We addressed this problem in twoways: by minimizing the motor’s friction torqueand by developing methods to compensate forthe motor’s inertia and friction torque.

A large motor also causes heat dissipationproblems in the drive circuitry due to the higherarmature current. We redesigned this circuit forlower loss.

Friction Torque ReductionThere are two main sources of friction torquein the permanent magnet DC motors used inthe application: mechanical losses in brushesand bearings, and magnetic losses. Larger mo-tors require bigger brush-to-commutator con-tact areas to maintain low electrical resistance,and the magnetic flux density is also higher.Both of these losses therefore increase with themotor output, causing larger friction torques.

Use of magnetic field analysis to optimize thecore design can reduce the magnetic losses. Fig.1 shows magnetic loss before and after this op-timization was carried out, both calculated andmeasured. A reduction of better than 60% wasachieved.

Compensation for Motor Inertia and FrictionTorqueAs shown in Fig. 2, the desired motor current iscalculated as the arithmetic sum of the power-assist current, and inertia, damping and fric-tion compensation currents. The motor controlcircuit applies this current, adjusting itself to

provide precisely the indicated value.The steering assistance control determines

the driver’s static steering effort. The output ofa torque sensor is phase compensated and thecharacteristics of Fig. 3 are applied to give thedriver positive road feel at higher speeds.

The motor operation is governed by

KT.Ia = TL + JM

.(dωM/dt) + DM.ωM + FM

.sgn(ωM)........ (Eq. 1)

where KT is the motor torque constant, Ia is thearmature current, TL is the load torque, JM isthe motor inertial moment, ωM is the motorangular velocity, DM is the motor’s viscous fric-tion coefficient and FM is the motor frictiontorque.

For the inertia compensation to cancel outthe disturbance torque caused by the motorinertial moment (JM), the motor inertia com-pensation current (IJ) must satisfy

IJ = KJ.(dωM/dt)..................................... (Eq. 2)

where KJ is a constant representing the inertiacompensation gain.

While inertia compensation improves steer-ing responsivity, it also causes damping prob-lems that affect steering stability, especially athighway speeds. The required damping com-pensation current (ID) is given by

Rel

ativ

e m

agne

tic lo

ssCalculated value Measured value

Before optimization

After optimization

1.2

1

0.8

0.6

0.4

0

Fig. 1 Magnetic loss.


TECHNICAL REPORTS

ID = −KD.ωM......................................... (Eq. 3)

where KD is a constant representing the damp-ing compensation gain.

The motor friction compensation current (IF)is given by

IF = KF.sgnωM..........................................(Eq. 4)

where KF is a constant representing the frictioncompensation gain.

The inertia, damping and friction compensa-tion all vary as a function of ωM, so some meansof detecting this quantity must be provided.Since ωM is proportional to the motor’s inducedvoltage, we can determine ωM from the inducedvoltage and substitute, allowing us to imple-ment the above control functions without theuse of an external sensor.

Reduction of Motor-Drive Circuit LossThe motor-drive circuit consists of powerMOSFETs in an H bridge circuit that is drivenby pulse width modulation (PWM) over a 20kHzcarrier. Fig. 4 shows two PWM-driven H bridgecircuits. Current control is simpler for the cir-cuit of Fig. 4a while the loss of the circuit ofFig. 4b is smaller. We obtained the advantagesof both by switching between the two configu-rations as appropriate.

Fig. 5 shows the motor current control sys-

Ste

erin

gas

sist

ance

cur

rent

(A

)

Vehicle speed (km/h)

Steering torque (N·m)

0 2 4 6 8

10

20

30

40

50

0

40

80

120

Steering torque

Vehicle speed

Motor speed(estimated)

Frictioncompensation

Steeringassistance

control

Desiredmotorcurrent

calculation

Motorcuttentcontrol

Phasecompensation

Dampingcompensation

Inertiacompensation

Motor current(detected)

d/dt

dω/dt

Motor drive signal

Fig. 2 A block diagram of the electric power steering system.

Fig. 3 Steering torque vs. steering assistancecurrent curves.

March 1997 · 21

TECHNICAL REPORTS

tem. A power inertia (PI) controller implementsfeedback control while feed-forward control isimplemented by a composite approach. Feed-forward control improves the responsivity of themotor current and suppresses the ripple cur-rents associated with switching between thetwo PWM methods.

ResultsFig. 6 shows the (steering angle/torque) fre-quency characteristics measured by sweepingthe steering frequency with the vehicle travel-ing at 50km/h. Applying inertial compensation

causes a phase lead in the 1~ 3kHz region thatindicates enhanced responsivity. Fig. 7 showshow the steering wheel returns from lock whenreleased at a vehicle speed of 10km/h. The wheelreturns to center smartly when friction com-pensation is applied.

Through use of the inertia, damping and fric-tion compensation technologies presented here,electric power-steering systems that save space

Desired motorcurrent

Feed-forward

compensator

PI controller Motor

Detected motor current

+–

++

Fig. 5 A block diagram of the motor currentcontrol system.

Q1

Q2

l1 Q3

Q4l2

Q1 l1 Q3

M M

Q2 l2 Q4

Q1 ONOFF

Q4 ONOFF

Motorcurrent

l2 l1 l1l2 t

(Q2 and Q3 are off)

Q1 ONOFF

Q4 ONOFF

Motorcurrent

l2 l1 l2 l1 t

(Q2 and Q3 are off)

+—

+—

Fig. 4 PWM driven H bridge circuits.

40

30

20

0

10

1 2 3 4 5

1 2 3 4 5

0

–90

0

–180

Frequency (Hz)

Frequency (Hz)

Ga

in (

°N·m

) (d

B)

Ga

in (

°N·m

) (d

B)

Key— With inertia compensation- - Without inertia compensation

a) G ain characteri sti cs

b) Phase characteri sti cs

Fig. 6 Measured steering angle/steering torquefrequency characteristics.


TECHNICAL REPORTS

Fig. 7 Measured steering wheel return at low vehicle speed.

and improve fuel economy can be applied tolarge cars without sacrificing road feel. Theauthors would like to acknowledge the gener-ous cooperation of the many people who havecontributed to this project. ❑

a) Without friction compensation b) With friction compensation

Return complete

Time (s)

02 4 6

−180

−90

Ste

erin

g to

rque

(N·m

)S

teer

ing

angl

e (°

)

Time (s)

Hands off steering wheel

1

0

−2

−1

−3

−4

642

Time (s)

02 4 6

−180

−90

Ste

erin

g an

gle

(°)

Return complete

Ste

erin

g to

rque

(N·m

)

1

0

−2

−1

−3

−4

642

Time (s)

Hands off steering wheel

March 1997 · 23

TECHNICAL REPORTS

Technologies for Gasoline EngineControl and ECU Miniaturization

by Yoshinobu Morimoto and Takanori Fujimoto*

Car electronics have helped car makers satisfyemission requirements and boost engine per-formance. Their role is expanding to includeadditional functions and convenience featuressuch as customizing car operation to individualdrivers. This report surveys trends in enginecontrol technology and engine control units.

Engine Control SystemsFig. 1 illustrates requirements for engine per-formance and engine control systems.

Mitsubishi Electric has developed a knock-control system and implemented it in an 8-bitmicroprocessor. The low-cost system improvesengine operating characteristics.

Larger engines of over two liters are startingto incorporate lean-burn systems that keep thegas-air mixture as lean as possible for better fueleconomy. Mitsubishi Electric and Mitsubishi

*Yoshinobu Morimoto and Takanori Fujimoto are with the Himeji Works.

Motors have jointly developed a linear airflowsensor and a control system that respond to fluc-tuations in engine speed. The corporation hasdeveloped improved airflow feedback controland an exhaust gas recirculation (EGR) systemwith feedback control that help meet strict newemission control standards.

The corporation has also developed sophisti-

High outputFuel

efficiency

Smoothrunning

Low noiseand vibration

Lowemissions Low cost

Highreliability

Compactlightweight

Automotive engine

Airflow sensor

Intake maniforldpressure sensor

Engine Transmission

Catalyticconverter

Front oxygen sensor

Crankshaftposition sensor

Rear heated oxygen sensor

Catalyticconverter

Purge solenoid

Canister

Check valve

Vent solenoid

Bypass solenoidFuel tank

Pressure sensor forevapotion system

Engine control unit(ECU)

Tester

Malfunctionindicator light

Automatictransmissioncontrol unit

Fig. 1 Goals of automotive engine design.

Fig. 2 Mitsubishi Electric’s OBD-II onboard diagnostic system.


TECHNICAL REPORTS

88 90 92 94 96 98Year

88 90 92 94 96 98Year

100

10

1

100

10

1Rel

ativ

e ca

paci

ty

Rel

ativ

e tim

e

development in the C language is growing morecommon. Mitsubishi Electric has developed op-timized C compilers and program design guide-lines that result in compact executable code just1.3 times the volume of handcrafted assemblylanguage.

ECU Hardware TechnologiesMitsubishi Electric is working to minimizeECU size through studies of both circuit designand production technology. ECUs must be com-

cated on-board diagnostics that detect misfiresby fluctuations in engine speed. This technol-ogy has been incorporated in production ve-hicles since 1994. Fig. 2 shows an engine controlsystem employing these new diagnostics.

Current electronic engine control systemslimit engine torque during automatic transmis-sion shifts for smoother shift transitions. Morerecently, data from the engine is being used tooptimize the automatic transmission shift pat-terns. Mitsubishi Electric has developed and ismass producing engine control units (ECUs)that integrate engine and automatic transmis-sion control for compact and subcompact cars.

Engine Control TechnologyMore sophisticated control programs are beingdeveloped to further improve engine perfor-mance and fuel economy. Executing these pro-grams requires high-performance ECUs. SinceECU performance is limited by microcomputerperformance, Mitsubishi Electric has developedhigh-performance microcomputers for ECUapplications. ECU developers are rapidly chang-ing over from 8- to 16-bit microprocessors thatoffer faster computing and larger memory ca-pacity. Fig. 3 shows trends in ECU microcom-puter arithmetic speed and memory capacity.Table 1 lists typical microcomputer specifica-tions. Microcomputers with a large on-chipflash EEPROM are being utilized to supportshorter development time.

Software DevelopmentManufacturers have developed engine controlsoftware in assembly language due to limita-tions on microcomputer speed and memorycapacity. With the larger memory capacitiesavailable in 16-bit microcomputers, software

a) Memory size b) Speed for 16 x 16 bitmultiplication

Table 1 Typical specifications for ECUMicrocomputers

Table 1 Typical specifications for ECUMicrocomputers

Architecture 8 bit 16 bit

OTP ROM 60kB 62kB 92kB

RAM 1.5kB 2kB 3.25kB

Speed for 16-bit 7.8µs 2.3µs 2.4µsmultiplication

ADC (bits x channels) 10 x 12 10 x 8 10 x 16

Timer channels 24 13 23

Serial communicationinterface 2 2 3

Package100-pin 84-pin 136-pin

QFP PLCC QFP

Fig. 3 Trends in microcomputers for automotiveapplications.

Table 2 Environmental Conditions at Several ECU Installation Sites

Location Temperature Water exposure Vibration Electromagnetic compatibility

Passenger compartment −40 ~ 85°C Negligible 4 ~ 5G max. Protection required against noise from radio equipment

Engine compartment −40 ~ 110°CComplete water-

4 ~ 5G max. Protection required against ignition and relay noiseproofing required

Mounted on engine −40 ~ 120°C Water-jet protec- 35G max. Protection required against ignition and relay noisetion required

March 1997 · 25

TECHNICAL REPORTS

Table 3 ECU Miniaturization Technologies

Component mounting

Exclusive use of surface mount components

High-density component placement, highly reliable soldering ofQFP with 0.55mm lead pitch

High-temperature solder that endures engine-compartmentheat cycling

Bare-chip mounting using die bonding and wire bondingtechnologies

Components

Small-outline packages such as QFPs with pin counts of 100or more

Fine-line PCBs with conductor pitch of 0.25mm

Custom ICs that reduce component count

Development of power MOSFETs with low on-state resistancefor reduced power dissipation

Case

High-density, high-pin-count connectors for 100 pins or more

Lightweight molded case

Water, shock and heat-resistant molding

Table 3 ECU Miniaturization Technologies

pact enough for space-efficient engine compart-ment mounting despite the large number of con-nections they require. ECUs must also resisttemperature extremes, water and severe vibra-tion—and they must be cost-effective. Table 2lists environmental stresses associated with dif-ferent mounting locations.

Mounting TechnologiesFig. 4 shows trends in printed circuit board (PCB)and component mounting technologies. ECUsdesigned to resist the harsh extremes of directengine mounting will need to be fabricated bybare-chip mounting on a ceramic substrate.

Although much of the high-density assem-bly technology required for ECUs has alreadybeen pioneered for compact consumer electronicequipment, the operating environment and“mission-critical” nature of the ECU requireshigher levels of reliability. Table 3 lists mount-ing technologies for ECU miniaturization.

Power dissipation is another obstacle to ECUminiaturization. The temperature differencebetween power devices in the ECU’s powersupply and output circuits and the ambient tem-perature must be kept within 30°C. Highlyefficient power MOS transistors with reduceddissipation offer a solution.

Another important issue is connector size.The complexity of ECUs with integrated auto-matic transmission control can require connec-tion of as many as 110 to 120 conductors. ECUsize reductions therefore await development ofsmall, high-density connectors.

Mitsubishi Electric has adopted two approachesto ECU miniaturization. The first level takesmaximum advantage of existing double-sided,surface-mounting technologies. The secondlevel employs bare-chip mounting and wirebonding technologies to build ultracompactECUs for direct engine mounting. Fig. 5 showsECUs developed using these two approaches.

The final step in ECU miniaturization willoccur when all necessary circuitry is imple-mented in a single silicon device that can bebare-chip mounted on a PCB. Achieving thislevel of integration will require development ofstandard circuits and continued integration offunctions into custom ICs. ECUs comprisingseveral custom ICs are currently being devel-oped, and bare-chip mounting technologies will


Year 80 85 90 95

BoardsDouble-sided

epoxy-impregnatedpaper

Double-sidedglass epoxy

Four-layerglass epoxy

Six-layer glassepoxy

Four-layerceramic

Mountingtechnology

Discrete components,(axial, radial)

Discrete components,hybrid ICs

Insertedcomponents, surface-mount components,

hybrid ICs

Surface-mountcomponents, custom

ICs Bare-chip mounting

Fig. 4 Printed-circuit board and mounting technology for ECUs.

TECHNICAL REPORTS

March 1997 · 27

a) High-density mounting b) Very high-density mounting

Fig. 5 ECU construction.

High-density wiring

Automatedassembly

Glass epoxy PCB

Plastic cover with integratedconnector

Case with integral waterproof connectorMolded case

Bare-chip mounted ICs

Aluminum wirebonding

Glass-epoxy multilayer PCB

soon be ready for production.

Electronic technology contributes to cleanervehicle exhaust and better fuel economy. Im-provements in ECU design such as those de-scribed here will contribute to even bettervehicle performance and reliability. ❑

TECHNICAL REPORTS

Starting and Charging Systems forFuel-Efficient Engines

by Toshinori Tanaka and Akira Morishita*

*Toshinori Tanaka and Akira Morishita are with the Himeji Works.

Demand for further reductions in automotivefuel consumption and emissions is expected tocontinue for the foreseeable future. This reportdescribes the use of computer-aided engineer-ing (CAE) and advanced materials to build small,lightweight starters and alternators. It also de-scribes power generation control technologythat reduces alternator loading.

Magnetic Flux AnalysisThe copper windings and iron magnets areamong the heaviest parts of motors and genera-tors. Their weight can be reduced through im-proved electromagnetic design based onmagnetic flux analysis.

In general, the size of a motor’s electromag-netic components can be expressed as

Da2Lc = K.(1/g).(T/I).(1/√Rs)................. (Eq. 1)

where Da is the rotor diameter, Lc is the rotorlength, K is a constant, g is the reduction gearratio, T is the torque, I is the current and Rs isthe internal resistance (a function of the out-put). T/I is referred to as the “torque gradient.”This equation is used to determine the motorsize once the motor’s torque, output and reduc-tion ratio are decided. New designs and tech-nologies can influence the value of K, leadingto smaller size and weight.

Mitsubishi Electric, for example, achieved asubstantial reduction in starter motor size byintroducing a six-pole magnetic starter motorwith auxiliary poles (Fig. 1 (b)). The corpora-tion recently introduced three-dimensional fluxanalysis to supplement conventional two-dimensional analysis (Fig. 2 (d)), and also light-ened the solenoid that engages the motor byredesigning the plunger mechanism to matchthe plunger movement with the meshing of thepinion and ring gears.

Cooling DesignFig. 3 shows the trade-offs between alternatoroutput characteristics, permissible temperaturerise and cooling noise. These trade-offs are im-portant since the alternator rotates continuouslywhile the engine is running and represents a sig-

nificant load. We began our redesign efforts byconducting an analysis of thermal stress and air-flow characteristics. Since temperature affectsthe lifetime of the voltage regulator and recti-fier, we redesigned the heatsinks for better cool-ing. We conducted an airflow cooling analysisof the alternator and then an overall study ofcooling characteristics to develop a more effec-tive cooling design. We also made models forvisual studies of airflow characteristics (Fig. 4).

Structural AnalysisAnother priority of our research was to addressstructural design issues. We analyzed the struc-ture of many parts in the starter motor mecha-nism, optimizing shapes and reducing weightwithout sacrificing mechanical strength.

Plastic components have already proven thor-oughly reliable for the starter reduction gear,pinion shift lever and switch plunger hook. Wetherefore studied using plastics for various othercomponents as a means of reducing weight. Twocomponents currently under study are the al-ternator pulley and the rear starter housing.

Plastics suffer the drawbacks of shock andwear resistance that are inferior to metals andthey require special handling. Most housingsare made of lightweight diecast aluminum. Fur-ther weight reductions may be achieved by us-ing magnesium; however, problems related tomass production and cost remain to be solved.

Charging System ImprovementsControl of alternator power generation offersanother avenue to reduce the engine load. Here,we will introduce one solution; a load-response-controlled regulator developed at MitsubishiElectric.

The alternator generates electricity whenpower is being used by the headlights and otherelectrical loads. Power generation draws onengine torque and can cause rough idling andeven stalling. The load-response-controlledregulator maintains a stable load on the engineeven when electrical loads fluctuate (Table 1).Stable alternator behavior allows the use oflower idle speeds, thus contributing to bettergas mileage. Since headlamp output fluctua-


TECHNICAL REPORTS

March 1997 · 29

Fig. 1 Studies for starter motor improvement.

(a) Structural analysis ofplastic rear housing

(b) Magnetic fluxanalysis (c) Internal gear stress

analysis(d) Flange elasticity

analysis

(f) Pinion gear stressanalysis

(e) Housing structuralanalysis

Fig. 2 Studies for alternator improvement.

(a) Diode thermal stress analysis

(b) IC regulator thermal stress analysis

(c) Thermal analysis of airflow

(e) Modal analysis

(d) Three-dimensional magnetic flux modeling

TECHNICAL REPORTS

Table 2 Effect of Power Generation Control onEngine Speed

Tem-peraturedown

Note: Each value is ration,not absolute value.

Noiseup

Tem-perature up

Noise down

Tem

pera

ture

Output down

Tem-peratureup

Output upTem-peraturedown

Ventilation noise Output

Output up

Noise upOutput down

Noise down

Ven

tilat

ion

nois

e

110

105

100

95

90

95

100

105

105 100 95 90 110100

Fig. 4 A sample alternator with transparenthousing.

Fig. 3 Relationships between alternator output,temperature and ventilation noise(arbitrary units).


tions can occur as a side effect, the load-con-trolled regulator is disabled at cruising speeds.

Table 2 shows the measured benefits of theload-controlled regulator. The idling speed can

be reduced by about 50rpm, which converts toa 1.4% increase in fuel economy measured bythe 10.15 method.

Care in the design of even “minor” engine com-ponents such as the starter and alternator canbring substantial benefits to fuel economy. Theimprovements presented here are likely to beseen in vehicles manufactured in the near fu-ture. ❑

Electrical load

Battery voltage

Alternator output current

Engine speed

Conventional alternatorAlternator with LRC regulator

At idle At cruising speeds

td LRC activated LRC canceled

Idle speed controltime lag and overshoot

Keytd Response delay timeLRC Load-response control

Displacement

Time

Note: For a 4 cylinder 1.6l engine with ISC (idling setting Ne-800rpm).

Table 2 Effect of Power Generation Control onEngine Speed

Load Conventional LRC regulator regulator

Headlamps (high) −70rpm −30rpm

Brakes −30rpm −20rpm

Fan −90rpm −40rpm

Rear-window defogger −50rpm −30rpm

NEW PRODUCTS

Angular velocity sensors will form avital part of future sophisticatedvehicle stability control systems.Given the required accuracy andspeed of response, they enable acomparison between the actual an-gular velocity of a vehicle and theangular velocity that would be in-ferred from the road speed and theangle of the steering wheel. Thisprovides, for instance, an indicationof when the vehicle may be skidding.

Mitsubishi Electric has developeda high-resolution sensor based on avibration beam with two supports thatis compact, lightweight and inexpen-

sive. If constant vibration is

Vibration beam

Support wire

Damper

Piezoelectric devices for drivingand detection

Support base

Y

X

Z

An Angular Velocity Sensor

Radio with CD Player for the European Market

Model DH-6581 is a top-end radiowith an integrated CD player andradio data system (RDS) functionsdeveloped as dealer-installedoptional equipment for MitsubishiMotors vehicles sold in Europe.

In addition to previous RDS func-tions, Model DH-6581 supports pro-gram type (PTY) and enhancementof other network (EON) signals. PTYenables the user to select stations bygenre. EON supports automaticprogram and station selection andfunctions that enable the unit to serveas a traffic information receiver.

The automatic station search hasbeen enhanced by incorporating twotuning circuits, one that scans rap-idly, one slowly. Continuous received-signal noise-level measurements are

fed to a microcomputer that dynami-cally adjusts reproduction frequencycharacteristics for optimal reception.

The CD player features a newlydeveloped automatic servo adjust-ment mechanism that absorbs varia-tions in optical characteristics fromdisk to disk, ensuring stable play-back. A one-bit DAC with an eight-bitoversampling digital filter guaranteeshigh-quality CD sound, and repeat,random and other user playbackfunctions are provided.

The unit features a front panel withcurved lines and enlarged LCD forimproved legibility, and theft-preventive design. The operatingpanel can be removed, and the unithas a flashing LED radio/CD playervehicle alarm system indicator. ❑

Table 1 Specifications

Operating voltage8 ~ 16V

range

Storage temp. range −40 ~ 85°C

Operating temp. range −30 ~ 75°C

Angular velocity −100 ~ 100 deg/sdetection range

Static output voltage 2.5V typ. (at 25°C)

DC temperature drift 10 deg/s max.

Sensitivity20 ±1mV/(deg/s)over operating

(scale factor)temperature range

Resolution 0.2 deg/s max.

Linearity (at 25°C)0.5% of full-scaledeflection

External dimensions85 x 36 x 52mm(including mount)

Weight 100g

applied in a particular direction, theCoriolis force causes a displacementperpendicular to the direction of thevibration with a magnitude propor-tional to the angular velocity aroundthe vibration beam. A ceramic piezo-electric vibrator is used to applyvibration to the beam and to measurebeam displacement. The sensor alsoincludes circuitry that converts thedetector output to an angular velocitysignal usable for vehicle stabilitycontrol. ❑

The radio/CD player for the European market.

Photo of angular velocity sensor.

Construction of sensor element.

Out

put v

olta

ge (

V)

Clock-wise

Counterclockwise Angular velocity (deg/s)

–100 –20–60 0 20 40 60 80 1000

1

2

3

4

5

key— Operational device- - Failed device

Output characteristics.

March 1997 · 31

of knocking, opening the door tooptimized per-cylinder engine controlcapabilities.

silent, quick-response mechanism, along-stroke damper and a high-capacity memory buffer that enhanceshock immunity during playback.User playback functions includecueing, repeat and random play. Aone-bit DAC with an eight-bit over-sampling digital filter guaranteeshigh-quality Minidisc sound.

The tuner memory stores up to 12FM stations and six AM stations, and

This product, developed for the ’97model year, combines an AM/FMtuner, Minidisc player and high-power, four-channel amplifier into acompact unit occupying just one DINmounting slot. This unit and its com-panion CD player, also newly devel-oped, serve as the flagship productsof Mitsubishi Electric’s new line ofcar-audio equipment.

The Minidisc player features a new

Mitsubishi Electric has developed arevolutionary sensor that accuratelydetects engine misfires and the level

NEW PRODUCTS

An Ionic Current Detection System

The ionic current detection module.

Block diagram

Radio with Minidisc Player for the Japanese Market

Engine control module

Ignition timing signal

Ignition coil

Additional parts

Ionic current

Ionic current

Spark plug (ion probe)

Bias voltagesupplyCurrent-to-voltage

converter

Misfiredetection

Knockdetection

Misfire detection signal

Knock detection signal

Wav

efor

m s

hapi

ng

IONIC CURRENT DETECTION MODULE

The radio/Minidisc player for the Japanese market


has an autostore function that auto-matically registers strong stations. Amemory scan function rotatesthrough the preset stations in suc-cession.

The audio unit features a high-power four-channel amplifier, elec-tronic volume control, and RCA-typepreamp output jacks. Use of special-purpose audio capacitors maintainsthe studio quality of digital musicsources. Through MitsubishiElectric’s proprietary Diabus, theunit’s LCD screen and controls canbe used to operate other Diabusequipment including a CD changer,CD player, Minidisc changer and TVtuner.

The modern front housing incorpo-rates a dot-matrix LCD panel thatdisplays Minidisc title information inEnglish or Japanese kana. ❑

The system can monitor conditionsin the cylinders of an internal com-bustion engine by measuring thecurrents associated with the ioniza-tion that occurs along the flame frontduring the power stroke. The systememploys conventional spark plugs asthe ion probes and is easily adaptedto any ignition system. It consists ofan ignition coil with supplementalelectronics, a modified distributorcap (if distributor ignition is used)and an ionic current detection mod-ule.

The system contributes to opti-mized engine control by monitoringeach combustion cycle in each cylin-der and providing feedback to theengine control system. It requiresneither engine structure modificationsnor an external high-voltage biassupply. The detector outputs can beeasily integrated into the enginecontrol computer. ❑

Country Address TelephoneU.S.A. Mitsubishi Electric America, Inc. 5665 Plaza Drive, P.O. Box 6007, Cypress, California 90630-0007 714-220-2500

Mitsubishi Electronics America, Inc. 5665 Plaza Drive, P.O. Box 6007, Cypress, California 90630-0007 714-220-2500Mitsubishi Consumer Electronics America, Inc. 2001 E. Carnegie Avenue, Santa Ana, California 92705 714-261-3200Mitsubishi Semiconductor America, Inc. Three Diamond Lane, Durham, North Carolina 27704 919-479-3333Horizon Research, Inc. 1432 Main Street, Waltham, Massachusetts 02154 617-466-8300Mitsubishi Electric Power Products Inc. Thorn Hill Industrial Park, 512 Keystone Drive, Warrendale, Pennsylvania 15086 412-772-2555Mitsubishi Electric Manufacturing Cincinnati, Inc. 4773 Bethany Road, Mason, Ohio 45040 513-398-2220Astronet Corporation 37 Skyline Drive, Suite 4100, Lake Mary, Florida 32746-6214 407-333-4900Powerex, Inc. Hills Street, Youngwood, Pennsylvania 15697 412-925-7272Mitsubishi Electric Research Laboratories, Inc. 201 Broadway, Cambridge, Massachusetts 02139 617-621-7500

Canada Mitsubishi Electric Sales Canada Inc. 4299 14th Avenue, Markham, Ontario L3R 0J2 905-475-7728Mitsubishi Electronics Industries Canada Inc. 1000 Wye Valley Road, Midland, Ontario L4R 4L8 705-526-7871

Mexico Melco de Mexico S.A. de C.V. Mariano Escobedo No. 69, Tlalnepantla, Edo. de Mexico 5-565-6269

Brazil MELCO do Brazil, Com. e Rep. Ltda. Av. Rio Branco, 123, a 1507, 20040, Rio de Janeiro 21-221-8343MELCO-TEC Rep. Com. e Assessoria Tecnica Ltda. Av. Rio Branco, 123, a 1507, 20040, Rio de Janeiro 21-221-8343

Argentina MELCO Argentina S.R.L. Florida 890-20º-Piso, C.P. 1005, Buenos Aires 1-312-6982

Colombia MELCO de Colombia Ltda. Calle 35 No. 7-25, Oficinas No. 1201/02, Edificio, Caxdac, Apartado Aereo 1-287-927729653, Bogotá

U.K. Mitsubishi Electric U.K. Ltd. Travellers Lane, Hatfield, Herts. AL10 8XB, England 1707-276100Apricot Computers Ltd. 3500 Parkside, Birmingham Business Park, Birmingham, B37 7YS, England 21-717-7171Mitsubishi Electric Europe Coordination Center Centre Point (18th Floor), 103 New Oxford Street, London, WC1A 1EB 71-379-7160

France Mitsubishi Electric France S.A. 55, Avenue de Colmar, 92563, Rueil Malmaison Cedex 1-47-08-78-00

Netherlands Mitsubishi Electric Netherlands B.V. 3rd Floor, Parnassustoren, Locatellikade 1, 1076 AZ, Amsterdam 20-6790094

Germany Mitsubishi Electric Europe GmbH Gothaer Strasse 8, 40880 Ratingen 2102-4860Mitsubishi Semiconductor Europe GmbH Konrad Zuse Strasse 1, 52477 Alsdorf 2404-990

Spain MELCO Iberica S.A. Barcelona Office Polígono Industrial “Can Magi”, Calle Joan Buscallà 2-4, Apartado de Correos 3-589-3900420, 08190 Sant Cugat del Vallés, Barcelona

Italy Mitsubishi Electric Europe GmbH, Milano Office Centro Direzionale Colleoni, Palazzo Perseo-Ingresso 2, Via Paracelso 12, 39-6053120041 Agrate Brianza, Milano

China Shanghai Mitsubishi Elevator Co., Ltd. 811 Jiang Chuan Rd., Minhang, Shanghai 21-4303030

Hong Kong Mitsubishi Electric (H.K.) Ltd. 41st Floor, Manulife Tower, 169 Electric Road, North Point 510-0555Ryoden Holdings Ltd. 10th Floor, Manulife Tower, 169 Electric Road, North Point 887-8870Ryoden Merchandising Co., Ltd. 32nd Floor, Manulife Tower, 169 Electric Road, North Point 510-0777

Korea KEFICO Corporation 410, Dangjung-Dong, K unpo, Kyunggi-Do 343-51-1403

Taiwan MELCO Taiwan Co., Ltd. 2nd Floor, Chung-Ling Bldg., No. 363, Sec. 2, Fu-Hsing S. Road, Taipei 2-733-2383Shihlin Electric & Engineering Corp. No. 75, Sec. 6, Chung Shan N. Rd., Taipei 2-834-2662China Ryoden Co., Ltd. Chung-Ling Bldg., No. 363, Sec. 2, Fu-Hsing S. Road, Taipei 2-733-3424

Singapore Mitsubishi Electric Singapore Pte. Ltd. 152 Beach Road #11-06/08, Gateway East, Singapore 189721 295-5055Mitsubishi Electric Sales Singapore Pte. Ltd. 307 Alexandra Road #05-01/02, Mitsubishi Electric Building, Singapore 159943 473-2308Mitsubishi Electronics Manufacturing Singapore Pte. Ltd. 3000, Marsiling Road, Singapore 739108 269-9711Mitsubishi Electric Asia Coordination Center 307 Alexandra Road #02-02, Mitsubishi Electric Building, Singapore 159943 479-9100

Malaysia Mitsubishi Electric (Malaysia) Sdn. Bhd. Plo 32, Kawasan Perindustrian Senai, 81400 Senai, Johor 7-5996060Antah MELCO Sales & Services Sdn. Bhd. 3 Jalan 13/1, 46860 Petaling Jaya, Selangor, P.O. Box 1036 3-756-8322Ryoden (Malaysia) Sdn. Bhd. 2nd Fl., Wisma Yan, Nos. 17 & 19, Jalan Selangor, 46050 Petaling Jaya 3-755-3277

Thailand Kang Yong W atana Co., Ltd. 15th Floor, Vanit Bldg., 1126/1, New Petchburi Road, Phayathai, Bangkok 10400 2-255-8550Kang Yong Electric Co., Ltd. 67 Moo 11, Bangna-Trad Highway, Km. 20 Bang Plee, Samutprakarn 10540 2-312-8151MELCO Manufacturing (Thailand) Co., Ltd. 86 Moo 4, Km. 23 Bangna-Trad, Bangplee, Semudparkarn 10540 2-312-8350~3Mitsubishi Elevator Asia Co., Ltd. Bangpakong Industrial Estate, 700/86~92, Moo 6 Tambon Don Hua Roh, 38-213-170

Muang District Chonburi 20000Mitsubishi Electric Asia Coordination Center 17th Floor, Bangna Tower, 2/3 Moo 14, Bangna-Trad Highway 6.5 Km, 2-312-0155~7(Thailand) Bangkawe, Bang Plee, Samutprakarn 10540

Philippines International Elevator & Equipment, Inc. Km. 23 West Service Road, South Superhighway, Cupang, Muntinlupa, 2-842-3161~5Metro Manila

Australia Mitsubishi Electric Australia Pty. Ltd. 348 Victoria Road, Postal Bag No. 2, Rydalmere, N.S.W. 2116 2-684-7200

New Zealand MELCO Sales (N.Z.) Ltd. 1 Parliament St., Lower Hutt, P.O. Box 30-772 Wellington 4-569-7350

Representatives

China Mitsubishi Electric Corp. Beijing Office SCITE Tower, Rm. 1609, Jianguo Menwai, Dajie 22, Beijing 1-512-3222Mitsubishi Electric Corp. Shanghai Office Room No. 1506-8, Shanghai Union Building 100, Yanan-East Street, Shanghai 21-320-2010

Korea Mitsubishi Electric Corp. Seoul Office Daehan Kyoyuk Insurance Bldg., Room No. 1204 #1, Chongno 1-ka, 2-732-1531~2Chongno-ku, Seoul

Viet Nam Mitsubishi Electric Corp. Ho Chi Minh City Office 8 B2, Han Nam Officetel 65, Nguyen Du St., 1st District, Ho Chi Minh City 8-243-984

MITSUBISHI ELECTRIC OVERSEAS NETWORK (Abridged)

MITSUBISHI ELECTRIC CORPORATIONHEAD OFFICE: MITSUBISHI DENKI BLDG., MARUNOUCHI, TOKYO 100-8310, FAX 03-3218-3455

VOL. 78/MAR. 1997 - Mitsubishi Electric · ferred to as Moore’s Law. Today, 64Mb DRAM devices are...

Documents

Transcript of VOL. 78/MAR. 1997 - Mitsubishi Electric · ferred to as Moore’s Law. Today, 64Mb DRAM devices are...