Robustness of control laws implemented in visual based target tracking system
Lee, Han Chuan.
Kaminer, Issac I.
Dobrokhodov, Vladimir N.
MetadataShow full item record
The previously developed guidance law implemented onboard the Small Unmanned Aerial Vehicle (SUAV) relies exclusively on the information from the image processing software and allows the performance of coordinated SUAV guidance and vision-based target tracking and motion estimation. This enables "passive only" coordinated tracking of noncooperative targets. An analysis of the system performance shows that the developed target tracking law demonstrates poor range holding capability when the target performs evasive maneuvers. Therefore, a new guidance law has been formulated by resolving SUAV dynamics with respect to the moving target frame, as opposed to the inertial frame in previous formulation. This simple modification results in theoretically achievable perfect range holding capability for the price of requiring the target motion information to be known. As a result, this new modification is based on the assumption of known target states, which in turn requires an implementation of a target motion estimator. An obvious tradeoff in performance of the "passive only" and "estimator based" target tracking systems is investigated in this thesis under realistic conditions including target loss events. This work extends previous results by investigating the performance of both guidance laws to the variation in target velocity and frequency of tracking loss events. The results obtained are based on the high fidelity 6DOF simulation implemented in SIMULINK, and analyzed using the multi-criteria optimization methodology introduced in the previous work. The results show that both guidance laws suffer predictable degradation in performance when subject to the external disturbances and tracking loss events. However, in the absence of tracking loss events, the new guidance law suffers less degradation in performance as compared to the old guidance law. When "frequency" of tracking loss events is low (less than 12%), the new guidance law is still able to provide better performance than the old guidance law As the "frequency" of tracking loss events increases further (between 12% to 25%), the performance of the new guidance law starts rapidly degrading, converging to that of the initial system; the target estimator is no longer able to provide a good prediction of the target velocity and heading to the guidance law.
Showing items related by title, author, creator and subject.
Adaptive Vision-Based Guidance Law with Guaranteed Performance Bounds for Tracking a Ground Target with Time-Varying Velocity Ma, Lili; Cao, Chengyu; Hovakimyan, Naira; Dobrokhodov, Vladimir; Kaminer, Isaac (2008-08);This work extends the earlier results of authors on vision-based tracking of a ground vehicle moving with unknown time-varying velocity. The follower UAV is equipped with a single camera. The control objective is to ...
Ma, L.; Cao, C.; Hovakimyan, N.; Dobrokhodov, V.; Kaminer, I.I (AIAA, 2010-05);This work discusses vision-based tracking of a ground vehicle moving with unknown time-varying velocity. The follower unmanned aerial vehicle is equipped with a single camera. The control objective is to regulate the two- ...
Romano, M.; Agrawal, B. (2004);The dynamics equations of a spacecraft consisting of two bodies mutually rotating around a common gimbal axis are derived by the use of the Newton–Euler approach. One of the bodies contains a cluster of single-gimbal var ...