An adaptive behavior decision model of mobile robot based on the neuromodulation

Abstract

How to effectively improve the behavior decision ability in unknown environment is a great challenge for mobile robots. Traditional methods suffer from the low efficiency and large computation load. Motivated by the regulation effect of serotonin and dopamine on human behavior, a novel adaptive behavior decision model for mobile robot is proposed in this work. This model integrates two kinds of adjustment factors that mimic the effects of two neuromodulators, serotonin and dopamine, to regulate the behavior decision of a mobile robot in unknown environment. Adjustment factor $$\mathbf {p}$$ is designed to simulate the function of the serotonin and make the robot avoid obstacles effectively, while the adjustment factor $$\mathbf {r}$$ is designed to simulate the function of the dopamine and make the robot approach the target quickly. Both of the adjustment functions can be adaptively regulated with the robot’s movement. Static and dynamic simulation results show that this neuromodulatory model has excellent collision avoidance capacity which can avoid the obstacles fluently, and the trajectory obtained tends to be optimal. Moreover, it is relatively simple, thus can effectively reduce the computation time and the load capacity, and greatly improve the running efficiency of the mobile robot in unknown environment.