A Localization Approach in a Distributed Multi-Agent Environment

Agents in a multi-agent system need to localize themselves and know the positions of the other agents for navigation and good coordination. In the outdoor environment, GPS can solve this problem; however indoors and in the GPS-denied situations, alternate arrangements are necessary. Most of such methods use landmarks or vision-based solutions, and are quite complex and expensive. The objective of this research is to design a simple, yet effective, sensing system for approximate localization of all the agents (or at least those in close vicinity) in a multi-agent environment. The proposed technique uses optical transmitters and receivers available with each agent. An agent transmits a signal to prompt the other agents about the "region" in which it is present. All other agents detect this signal and mark the regions (boundaries) where this agent is detected. When all other agents have marked the boundaries for the transmitting agent, the information gathered is exchanged and intersection region is calculated. The final subset region gives the approximate location of the agent that had transmitted the signal. MATLAB (R) simulations show the effectiveness of the proposed localization approach. The proposed system is scalable and does not require landmarks; more precise localization is possible depending on the requirements.