An SOM-based algorithm with locking mechanism for task assignment

In this paper, an improved self-organizing map (SOM) approach is proposed to solve task assignment problem of multiple robots. The purpose is to control a team of robots to accomplish all tasks with a minimal consumption. Compared with conventional self-organizing map approach, there is important improvement in the proposed algorithm. A locking mechanism is proposed to avoid the output neuron oscillation phenomenon which leads to a too many number of iteration. To demonstrate the effectiveness of the proposed algorithm, three groups of experiments are conducted. The experimental results show that the improved algorithm decreases the total energy consumption of entire robot team efficiently, and have less number of iterations than conventional self-organizing map.