Alice: Road Following

Prev: MHE

Course Home

Next: Packet-based Estimation

This lecture will present a practical example application for the estimation tools presented in the Monday and Wednesday lectures. The example application is road estimation and tracking for off-highway navigation. I'll present implementation details, resources and results from live testing of an extended Kalman filter implementation, and provide resources and preliminary results for implementing moving horizon estimation to solve the same problem.

Lecture Materials

Lecture presentation slides forthcoming.

Reading

• Cf. (extended) Kalman filter reading materials from Monday 4/17/06 lecture.
• Cf. Moving Horizon Estimation reading materials from Wednesday 4/19/06 lecture.
• Model-Based Estimation of Off-Highway Road Geometry using Single-Axis LADAR and Inertial Sensing, Lars B. Cremean and Richard M. Murray, To appear, Proc. of 2006 International Conference on Robotics and Automation

Additional Resources

• The Open Source Computer Vision Library (OpenCV) has a well-documented C++ class that implements the Kalman filter, and a host of image processing tools. See http://www.intel.com/technology/computing/opencv/ and http://opencvlibrary.sourceforge.net/
• An Octave (MATLAB clone) toolbox for implementing nonlinear receding horizon control and moving horizon estimation is available at http://jbrwww.che.wisc.edu/home/tenny/nmpc/. The release is somewhat old (2003); be sure to get the CVS version to avoid some known bugs.