BICLARE: bio-inspired collaborative lightweight algorithm for robust swarm exploration

Abstract

Having access to a floor plan upon arrival enables first responders to act swiftly and efficiently. This study investigates the use of robotic swarms for autonomous environment mapping, with the aim of generating maps before emergency personnel arrive at the scene. We introduce BICLARE, a lightweight exploration algorithm inspired by ant-colony pheromone-based communication and navigation. BICLARE employs a confidence model to estimate cell occupancy within a map and optimizes target selection based on travel time and battery estimates to improve efficiency. The algorithm is designed to run efficiently on low-cost hardware. The performance of BICLARE is validated through simulated experiments in diverse and dynamic environments, demonstrating accurate mapping and robust coverage even in the presence of noise and uncertainty. Field trials further confirm the feasibility of the approach on low-cost COTS devices.