Congestion Strategies for Clustered Central Place Foraging
Apurva Bharaswadkar, Vida Vakilian, Brian Thoms, Jason T. Isaacs · 2021
In central place foraging tasks, multiple robots search for and gather resources in an environment. The robots then proceed to deposit the collected resources in a single central location. The performance of central place foraging approaches is reduced due to congestion around the central collection point. The congestion problem is made worse in the case where resources are distributed in clusters whereby several robots collect resources in the same area and deliver those resources to the collection point along the same path. The approach proposed here seeks to alleviate this congestion problem through simple path planning strategies that reduce the number of inter-robot collisions. Path Planning And Collision Avoidance Algorithm For Clustered Central Place Foraging (PPCA-CCPFA) addresses congestion by detecting possible inter-robot collisions and finding alternate collision free paths for each robot. We compare our approach to the Distributed Deterministic Spiral Search Algorithm (DDSA). This approach provides a notable increase in the performance of DDSA in cases where resources are distributed in a single cluster. A simulation study was conducted using the swarm robotics simulation tool ARGoS to measure the effectiveness of the proposed approach as measured by the number of resources collected per unit time and by the number of inter-robot collisions per unit time.