matthew deans - research

Bearings-only localization and mapping

My thesis research is on combining odometry and monocular vision in order to simultaneously produce metric maps of the environment and localize the robot. I have been investigating methods such as nonlinear optimization (bundle adjustment) kalman filtering unscented filter covariance intersection projective factorization particle filters the variable state dimension filter hierarchical mapping strategies My most current work deals with the variable state dimension filter framework and makes a few improvements. I am submitting an ICRA paper, which should be posted here soon. More information on my SLAM related work can be found here.

Antarctic Meteorite Search

I worked on the Antarctic Meteorite Search project for a few years. Our team converted the Nomad rover to be capable of carrying out an autonomous search for meteorites on the ice fields in the Antarctic. My work on the project included hi-res camera & pant-tilt calibration omnidirectional camera hardware and software estimating the position of science targets from images localization and mapping helping to specify the science autonomy software architecture being a member of the field team in 1998 and 1999 For details about this project see the main project site at http://www.ri.cmu.edu/projects/project_238.html.

Visual Servoing

I spent my first summer in grad school working at NASA Ames research center. I worked with researchers there to implement visual servoing for terminal guidance on the Marsokhod rover. I worked on hardware specification and integration and helped develop a visual servoing library that the group still uses. For details about this project see my description.

M. Deans, Maximally informative statistics for localization and mapping, RIACS Technical Report 01.25, October 2001. [abstract] [gzipped ps(71k)] [pdf(192k)]

M. Deans, Maximally informative statistics for localization and mapping, Submitted to IEEE ICRA '02, 2002. [draft]

M. Deans and M. Hebert, Experimental comparison of techniques for localization and mapping using a bearings-only sensor, International Symposium of Experimental Robotics, ISER 2000 (2000). [abstract] [gzipped ps(220k)] [pdf(600k)]

M. Deans and M. Hebert, Invariant filtering for simultaneous localization and mapping, Proceedings 2000 ICRA. IEEE International Conference on Robotics and Automation pp 1042-7, April 2000. [abstract] [gzipped ps(46k)] [pdf(134k)]

M. Deans, Robust, Efficient, Simultaneous Localization and Mapping using Bearings Only Sensing, Thesis Proposal, Robotics Institute, Carnegie Mellon University, February 2000. [abstract] [gzipped ps(82k)]

Lee, P., W.A. Cassidy, D. Apostolopoulos, D. Bassi, L. Bravo, H. Cifuentes, M. Deans, A. Foessel, S. Moorehead, M. Parris, C. Puebla, L. Pedersen, M. Sibenac, F. Vald s, N. Vandapel, and W. L. Whittaker, Search for Meteorites at Martin Hills and Pirrit Hills, Antarctica, 30th Lunar and Planetary Science Conference, Houston, Texas, 1999.

R. Blum, M. C. Deans, Distributed Random Signal Detection with Multibit Sensor Decisions, IEEE Transactions on Information Theory, Vol. 44, No. 2, pp 516-524 (1998). M. C. Deans, An Adaptive Algorithm for Distributed Detection System Design, Lehigh University Master's Thesis, Lehigh University, Bethlehem, PA, 1996.

M. C. Deans, R. S. Blum, An Adaptive Algorithm for Distributed Detection System Design, Sensor Fusion and Networked Robotics VII, Proceedings of the SPIE 2589, pp 172-179 (1995). [abstract] [gzipped ps(53k)]

K. R. Elder, Hao-Wen Xi, M. Deans, J. D. Gunton, Spatiotemporal Chaos in the Damped Kuramoto-Sivashinsky Equation, p. 702 in AIP Conference Proceedings 342, CAM-94 Physics, Cancun, Mexico 1994, Arnulfo Zepeda (Ed.) AIP Press 1995.