Self-Adaptive Collective Motion of Swarm Robots

Abstract

Collective motion is a fundamental operation of robot swarms by which a group of robots move from a source to a destination in a cohesive way (i.e., connectivity is preserved during these movements). However, the collective motion of robot swarms along preplanned paths has not been well studied. In this paper, we propose self-adaptive collective motion algorithms for swarm robots in 3-D space. Using the proposed collective motion algorithms, robots are able to move along a preplanned path from a source to a destination while satisfying the following requirements: 1) the robots use only one-hop neighbor information; 2) the robots maintain connectivity of the network topology for information exchange; 3) the robots maintain a desired neighboring distance; and 4) the robots are capable of bypassing obstacles without partitioning the robot swarm (i.e., member loss). Our basic idea is to introduce a guidance force and a topology force into the system. The guidance force is used to guide the robots to their destination along the preplanned path. It ensures that the robots continue to move until they reach their destination. The topology force is used to maintain a “good” topology of the robot swarm, such as maintaining connectivity of the network topology and the desired distance between neighboring robots. The resultant of the guidance and topology forces determines the movement of a robot. We develop collective motion algorithms for three cases: 1) no obstacles or leaders; 2) no obstacles with a leader; and 3) with obstacles (with and without a leader). Extensive simulations are conducted to evaluate performance of the proposed algorithms. The simulation results show that: 1) our algorithms meet all the requirements; 2) our algorithms are resistant to GPS errors and robot failures; and 3) self-adaptive control of our algorithms makes network topologies more stable and significantly saves travel time of swarm robots. Note to Practitioners —We propose self-adaptive collective motion algorithms that enable swarm robots to move along a preplanned path from a source to a destination in 3-D space. Our algorithms use only one-hop neighbor information and operate without central controllers. The algorithms are designed to be self-adaptive in the sense that robots are able to dynamically determine proper moving parameters, based on their environments and statuses. With the proposed algorithms, swarm robots are able to: 1) maintain connectivity and a desired neighboring distance during movement; 2) bypass obstacles without member loss (i.e., the robot swarm is partitioned); and 3) be resistant to GPS errors and robot failures. We address both cases of with a leader and without a leader. Simulation results show effectiveness of our algorithms which can be applied in applications of surveillance, search and rescue, mining, agricultural foraging, autonomous military units, and distributed sensing in micromachinery or human bodies.