Probabilistic Robotics Bayes Filter Implementations Particle filters.
-
date post
18-Dec-2015 -
Category
Documents
-
view
223 -
download
0
Transcript of Probabilistic Robotics Bayes Filter Implementations Particle filters.
![Page 1: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/1.jpg)
Probabilistic Robotics
Bayes Filter Implementations
Particle filters
![Page 2: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/2.jpg)
Sample-based Localization (sonar)
![Page 3: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/3.jpg)
Represent belief by random samples
Estimation of non-Gaussian, nonlinear processes
Monte Carlo filter, Survival of the fittest, Condensation, Bootstrap filter, Particle filter
Filtering: [Rubin, 88], [Gordon et al., 93], [Kitagawa 96]
Computer vision: [Isard and Blake 96, 98] Dynamic Bayesian Networks: [Kanazawa et al., 95]d
Particle Filters
![Page 4: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/4.jpg)
Importance Sampling with Resampling
),...,,(
)()|(),...,,|( :fon distributiTarget
2121
n
kk
n zzzp
xpxzpzzzxp
)(
)()|()|( :gon distributi Sampling
l
ll zp
xpxzpzxp
),...,,(
)|()(
)|(
),...,,|( : w weightsImportance
21
21
n
lkkl
l
n
zzzp
xzpzp
zxp
zzzxp
g
f
![Page 5: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/5.jpg)
Importance Sampling with Resampling
Weighted samples After resampling
![Page 6: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/6.jpg)
Particle Filters
![Page 7: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/7.jpg)
)|()(
)()|()()|()(
xzpxBel
xBelxzpw
xBelxzpxBel
Sensor Information: Importance Sampling
![Page 8: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/8.jpg)
'd)'()'|()( , xxBelxuxpxBel
Robot Motion
![Page 9: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/9.jpg)
)|()(
)()|()()|()(
xzpxBel
xBelxzpw
xBelxzpxBel
Sensor Information: Importance Sampling
![Page 10: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/10.jpg)
Robot Motion
'd)'()'|()( , xxBelxuxpxBel
![Page 11: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/11.jpg)
1. Algorithm particle_filter( St-1, ut-1 zt):
2.
3. For Generate new samples
4. Sample index j(i) from the discrete distribution given by wt-
1
5. Sample from using and
6. Compute importance weight
7. Update normalization factor
8. Insert
9. For
10. Normalize weights
Particle Filter Algorithm
0, tS
ni 1
},{ it
ittt wxSS
itw
itx ),|( 11 ttt uxxp )(
1ij
tx 1tu
)|( itt
it xzpw
ni 1
/it
it ww
![Page 12: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/12.jpg)
draw xit1 from Bel(xt1)
draw xit from p(xt | xi
t1,ut1)
Importance factor for xit:
)|()(),|(
)(),|()|(ondistributi proposal
ondistributitarget
111
111
tt
tttt
tttttt
it
xzpxBeluxxp
xBeluxxpxzp
w
1111 )(),|()|()( tttttttt dxxBeluxxpxzpxBel
Particle Filter Algorithm
![Page 13: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/13.jpg)
Resampling
• Given: Set S of weighted samples.
• Wanted : Random sample, where the probability of drawing xi is given by wi.
• Typically done n times with replacement to generate new sample set S’.
![Page 14: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/14.jpg)
1. Algorithm systematic_resampling(S,n):
2.
3. For Generate cdf4. 5. Initialize threshold
6. For Draw samples …7. While ( ) Skip until next threshold reached8. 9. Insert10. Increment threshold
11. Return S’
Resampling Algorithm
11,' wcS
ni 2i
ii wcc 1
1],,0]~ 11 inUu
nj 1
11
nuu jj
ij cu
1,'' nxSS i
1ii
Also called stochastic universal sampling
![Page 15: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/15.jpg)
Start
Motion Model Reminder
![Page 16: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/16.jpg)
Proximity Sensor Model Reminder
Laser sensor Sonar sensor
![Page 17: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/17.jpg)
17
![Page 18: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/18.jpg)
18
![Page 19: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/19.jpg)
19
![Page 20: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/20.jpg)
21
![Page 21: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/21.jpg)
22
![Page 22: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/22.jpg)
23
![Page 23: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/23.jpg)
24
![Page 24: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/24.jpg)
25
![Page 25: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/25.jpg)
26
![Page 26: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/26.jpg)
27
![Page 27: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/27.jpg)
28
![Page 28: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/28.jpg)
29
![Page 29: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/29.jpg)
30
![Page 30: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/30.jpg)
31
![Page 31: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/31.jpg)
32
![Page 32: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/32.jpg)
33
![Page 33: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/33.jpg)
34
![Page 34: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/34.jpg)
35
![Page 35: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/35.jpg)
36
Sample-based Localization (sonar)
![Page 36: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/36.jpg)
37
Initial Distribution
![Page 37: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/37.jpg)
38
After Incorporating Ten Ultrasound Scans
![Page 38: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/38.jpg)
39
After Incorporating 65 Ultrasound Scans
![Page 39: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/39.jpg)
40
Estimated Path
![Page 40: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/40.jpg)
Using Ceiling Maps for Localization
![Page 41: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/41.jpg)
Vision-based Localization
P(z|x)
h(x)z
![Page 42: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/42.jpg)
Under a Light
Measurement z: P(z|x):
![Page 43: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/43.jpg)
Next to a Light
Measurement z: P(z|x):
![Page 44: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/44.jpg)
Elsewhere
Measurement z: P(z|x):
![Page 45: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/45.jpg)
Global Localization Using Vision
![Page 46: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/46.jpg)
47
Robots in Action: Albert
![Page 47: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/47.jpg)
48
Application: Rhino and Albert Synchronized in Munich and Bonn
[Robotics And Automation Magazine, to appear]
![Page 48: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/48.jpg)
Localization for AIBO robots
![Page 49: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/49.jpg)
50
Limitations
•The approach described so far is able to • track the pose of a mobile robot and to• globally localize the robot.
•How can we deal with localization errors (i.e., the kidnapped robot problem)?
![Page 50: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/50.jpg)
51
Approaches
•Randomly insert samples (the robot can be teleported at any point in time).
• Insert random samples proportional to the average likelihood of the particles (the robot has been teleported with higher probability when the likelihood of its observations drops).
![Page 51: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/51.jpg)
52
Random SamplesVision-Based Localization936 Images, 4MB, .6secs/imageTrajectory of the robot:
![Page 52: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/52.jpg)
53
Odometry Information
![Page 53: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/53.jpg)
54
Image Sequence
![Page 54: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/54.jpg)
55
Resulting Trajectories
Position tracking:
![Page 55: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/55.jpg)
56
Resulting Trajectories
Global localization:
![Page 56: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/56.jpg)
57
Global Localization
![Page 57: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/57.jpg)
58
Kidnapping the Robot
![Page 58: Probabilistic Robotics Bayes Filter Implementations Particle filters.](https://reader031.fdocuments.us/reader031/viewer/2022032201/56649d245503460f949fa64c/html5/thumbnails/58.jpg)
60
Summary
• Particle filters are an implementation of recursive Bayesian filtering
• They represent the posterior by a set of weighted samples.
• In the context of localization, the particles are propagated according to the motion model.
• They are then weighted according to the likelihood of the observations.
• In a re-sampling step, new particles are drawn with a probability proportional to the likelihood of the observation.