Villanova Autonomous Surface Vehicle: Research Challenges ......Vehicle: Research Challenges with...
Transcript of Villanova Autonomous Surface Vehicle: Research Challenges ......Vehicle: Research Challenges with...
Villanova Autonomous Surface Vehicle: Research Challenges with
RoboBoat and RobotX Competitions Anderson Lebbad and C. Nataraj
Center for Nonlinear Dynamics & Control Department of Mechanical Engineering Villanova University, Villanova, PA 19085
April 10, 2014
Competitions: RoboBoat
∗ 7th Annual Competition ∗ Held at the Founders Inn
and Spa, Virginia Beach ∗ 8-13 July 2014 ∗ Boat must fit within a 6ft
x 3ft x 3ft box ∗ Max weight of 140 lbs
Competitions: RobotX
∗ First ever competition ∗ Only 15 Teams accepted to
compete, 3 from each of 5 different countries
∗ Held at the Marina Bay, Republic of Singapore
∗ 20-27 October 2014 ∗ Standard surface vessel
(14ft catamaran) ∗ Max payload of 300 lbs
∗ Demonstrate Navigation and Control (Mandatory) ∗ Underwater Search and Report ∗ Symbol Recognition and Docking ∗ Observation ∗ Obstacle Detection and Avoidance
Tasks
∗ Craft must autonomously maneuver between two sets of buoys
∗ Buoy gates are marked by red left buoy, green right ∗ Linear course ∗ Start gate will have a GPS point provided, end gate
must be detected autonomously
Demonstrate Navigation and Control
∗ Craft must locate an underwater device emitting an acoustic signal
∗ 40m x 100m search area ∗ Attached to an anchored buoy (will be visible for
RoboBoat, may or may not be for RobotX) ∗ GPS coordinated of the search area will be given
Underwater Search and Report
∗ Three docking bays, each with a unique symbol on a placard affixed to it ∗ Triangle, Circle and Cruciform
∗ One of the three symbols will be designated as the symbol for that particular day
∗ Craft must approach the dock and enter the bay of the designated symbol
∗ Placards may be moved from bay to bay between runs
Symbol Recognition and Docking
∗ Craft must detect and report a sequence of colors emitted ∗ Emitter will be on a white buoy no more than 3 m above
water for RobotX ∗ Emitter will be underwater near the deployment dock
for RoboBoat
∗ Each color will be displayed for 500 ms, and the sequence will be repeated after a 2 second delay
Observation
∗ There are 3 gates on either side of a field of obstacles ∗ Craft must enter and exit through designated gates,
specified the night before ∗ Gates are between four buoys, specified by color ∗ Red / White / White / Green
∗ GPS coordinates are given for the four corners of the field
Obstacle Detection and Avoidance
∗ GPS ∗ Compass ∗ Inertial Measurement Unit ∗ Video ∗ LIDAR ∗ Hydrophones
Sensors
∗ Speedgoat Target Machine ∗ Realtime ∗ xPC Target ∗ Simulink
∗ Arduino Microcontrollers ∗ PandaBoard ∗ Lightweight Communications and Marshalling
Computers, Controllers, and Communication
Finite State Machine
∗ Ordered Subsystems ∗ Task Oriented ∗ Separable for Isolated
Testing
Video
∗ Crucial to observation task, symbol recognition and gate identification for obstacle detection / avoidance
∗ Color spaces ∗ RGB ∗ YCbCr
Video: Challenges
∗ Challenges ∗ Over Saturation ∗ Reflections ∗ Background Noise
Video: Sample Image
Video: Color Mask
Red Binary Mask Green Binary Mask
Video: Submatrix
Red Binary Mask Green Binary Mask
Video: Blob Analysis
Video: Extent Filter
Binary Mask Blob has been Filtered Out
Video: Detect Closest
LIDAR
∗ Scanning laser that measures distance
∗ Has a 270 degree field of view
∗ 4 readings per degree ∗ 60 m range
∗ Targets can be detected and selected via LIDAR ∗ Known possible color states of targets ∗ Probability of the target being each of the finite
states can be found ∗ Bayes Filter
Color Analysis
∗ Develops a map based off of sensor information ∗ LIDAR
∗ Localizes the craft in the map ∗ Continuously updates the map as the craft moves
through the environment ∗ Critical for Obstacle Detection and Avoidance
Simultaneous Localization and Mapping
∗ Sensor suite for acoustic task ∗ Pattern Recognition Algorithm for Symbol
Recognition and Docking ∗ Sensor suite for detecting underwater LED array for
(RoboBoat) Observation ∗ Path planning algorithm for Obstacle Detection and
Avoidance
Still in Development
Team Members
∗ Current Students ∗ Anderson Lebbad – MSME ’14 ∗ Nick DiLeo – MSME ’14 ∗ Chida Mahananda – MSEE ‘14 ∗ J. Wes Anderson – MSME ‘15 ∗ Edward Zhu – BSME ‘15 ∗ Dylan DeGaetano – BSME ’15 ∗ Michael Benson – BSME ‘15 ∗ Gin Cheng – BSCPE ‘15 ∗ Gus Jenkins – BSME ‘16 ∗ Priya Shah – BSME ‘16
∗ Previous Students ∗ Joseph Denny – MSME ‘14 ∗ Ryan Holihan – BSME ‘12 ∗ Shahriar Khan – MSEE ’13 ∗ Lester McMackin – BSME ‘13 ∗ Mike Weber – BSCPE ‘13 ∗ Ralph Sullivan – MSME ‘12
∗ Advisors ∗ Dr. C. Nataraj ∗ Dr. Garrett Clayton ∗ Dr. Cedrick Kwuimy ∗ Dr. Helen Loeb ∗ Dr. DongBin Lee
∗ Acknowledgements ∗ Mr. Chris Townsend