Abstract | ||
---|---|---|
In this paper we consider a probabilistic model of a pursuit-evasion game on Rn, in which neither pursuer nor evader positions are known with certainty. Both parties are represented by normal distributions that evolve according to a Kalman filter as new sensor readings (observation from overhead camera/satellite images) are obtained. The objective is to design the control commands issued by the pursuer (which is executed noisily). The control commands issued by the evader are unknown - only sensor measurements are given. Even with such limited knowledge we prove boundedness of a distance between the pursuer's distribution and the evader's true distribution (one that takes into account the evader's control commands). Our simulation results support the claimed guarantees. |
Year | DOI | Venue |
---|---|---|
2014 | 10.1109/IROS.2014.6942544 | IROS |
Keywords | DocType | ISSN |
Kalman filter,pursuer distribution,Kalman filters,overhead camera,control commands,sensor readings,pursuit evasion game,satellite images,normal distributions,evader true distribution,sensor measurements,probabilistic model | Conference | 2153-0858 |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Chanyoung Jun | 1 | 0 | 0.34 |
Subhrajit Bhattacharya | 2 | 462 | 36.93 |
robert ghrist | 3 | 446 | 32.46 |