Comparison of Gradient Estimation in Cooperative Multi-Agent Source Seeking
Agent-based source seeking problem is addressed in this paper. This problem is relevant in, e.g., in hazardous gas leak in a chemical disaster. In the cooperative search, agents develop a formation to effectively search the source by communicating one to another via a communication topology. The source, or the search target, is represented by a scalar field y which might describe a temperature level, hazardous concentration of substances or vapor. Every agent has the information on its own position and the value of y at any instance. The agents are identical modeled as single and double integrator. Consensus filter is used to control the agent formation and comparison three types of gradient estimation are employed to search the source. Experiments show that the proposed schemes give good performance to solve cooperative search for source seeking problem