Two-Stage Multi-Robot Task Allocation Algorithms in Local Communication Scenarios

Abstract

In robot emergency rescue scenarios, it is common for communication between robots to be restricted, allowing interaction only within localized communication ranges. However, commonly proposed task allocation algorithms with weak communication models often focus on communication quality while neglecting communication distance. Consequently, this paper introduces a Bernoulli communication model incorporating distance information as the communication model. Subsequently, a two-stage distributed task allocation algorithm is proposed based on this communication model. In the convergence stage, each robot utilizes the K-means algorithm to determine the target task group and employs a distributed bee algorithm to select the target task. In the dispersion stage, robots use an improved distributed genetic algorithm to allocate remaining tasks, exchanging optimal solutions with other robots, thereby achieving conflict-free autonomous task allocation. Finally, real-time simulations are conducted to validate that this algorithm effectively resolves the multi-robot task allocation problem within the localized communication model.