This paper presents a method of source seeking for multi-agent unmanned aerial vehicle (UAV) systems in an environment with unknown obstacles. A distributed algorithm relies on localized measurements and neighbor-to-neighbor interactions to enable the group of UAVs to navigate to the location of the source. Each agent takes a scalar measurement of a signal emanating from the source, and the direction of motion is determined by estimating the gradient of the signal. The direction vector to the source is agreed upon by the agents in a coordinated manner to ensure aggregate motion toward the source of interest. Additionally, the agents avoid obstacles that are located between their initial position and the source without a priori knowledge of the number or positions of the obstacles. To ensure robust obstacle avoidance, a hybrid control method is used. The agents maintain a specified formation to improve observability during the gradient estimation process. Theoretical results for the algorithm are presented through simulation of the proposed source seeking method for a group of UAVs with single-integrator dynamics.
Distributed Source Seeking and Robust Obstacle Avoidance Through Hybrid Gradient Descent
01.03.2019
4671154 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Emergency Obstacle Avoidance Based on Gradient Descent Distance for Self-driving Vehicles
British Library Conference Proceedings | 2022
|Emergency Obstacle Avoidance Based on Gradient Descent Distance for Self-driving Vehicles
Springer Verlag | 2022
|Robust Hybrid Control for Unknown Obstacle Avoidance
British Library Conference Proceedings | 2009
|Robust Hybrid Control for Unknown Obstacle Avoidance
AIAA | 2009
|