Regular Meeting Minutes
April 5, 2008, 10:30AM
- Introduction to Bayesian Particle Filters
- BotFest ‘08
4. Update on Open AHRMS
5. Show and Tell
Introduction to Bayesian
Particle Filters – Professor
Barton C. Massey,
Authors E-mail - firstname.lastname@example.org
Authors Website - http://web.cecs.pdx.edu/~bart/
1. What's the problem?
Why do we need to know state of Robot? Platform Control, Accurate Navigation
One Solution - Bayesian Particle Filtering (BPF)
A method of Sensor Fusion – Robot State Space tracking
BPF is the HOT NEW Alternative to Kalman Filtering – Last 10-15 years
a) Deal with sensor unreliability by adding redundant sensors -
b) Kalman - linear normal distribution estimation of maximum probability position
Linear algebra - lots of high level math required
Improvements - EKF, UKF
c) Bayesian Particle Filtering - Computational intensive
3. Bayesian Particle Filtering - How does it work?
State of vehicle - s
g: dead reckoning
h: measurement function
Need to deal with multiple probabilities from Dead Reckoning and sensor measurement. In order to find robot location you must maximize the product of all state probabilities. Sounds Simple, Right? à Space is large
Basic Process for BPF:
0) Create copies of vehicle state, Si. Set all probabilities to 1 (Wi = 1)
1) Propagate each state copy, Si, forward one time of sample.
2) Calculate the probability of each state (Si) given the sensor reading (h) and probability, Wi.
3) Pick and Si to use right now based on maximum probability product, Wi[max]
4) Resample Si. Use a weighted random selection of states to continue duplication and propagation.
5) Repeat for next time sample.
Wi' = h(Si)*Wi [normalize]
Bayesian Particle Filter is a method of simulated evolution. BPF is a useful machine learning technique.
Drawback - Computationally intensive - Floating Point can be a plus
BotFest '08 Indoor Competition – May 17th at the
BotFest ’08 Outdoor Competition – May 25th at the
New PARTS Website – New Content Management System
OpenAHRMS is the Open Source project to develop an Attitude and Heading Reference and Measurement System for Personnel Robotics. Hardware layout is nearing completion. The current design will provide USB connectivity to Host controller.