Resource Optimization Algorithm for Task Offloading of Service Robots with Position Prediction

摘要

With the wide application of service robots, task offloading of robots becomes a new research direction. Current studies do not fully consider the scenario of high robot density in some local areas, and the traditional schemes may lead to unreasonable allocation of offloading resources caused by robot congestion. In this paper, we propose an intelligent task offloading algorithm for service robots based on position prediction in the case of congestion. The overall idea is to predict the positions of pedestrians and robots by combining the social force model and the artificial potential field method. The predicted positions are used to indicate in advance the areas where pedestrians and robots may gather, and Q-learning is used to generate task offloading strategy for each robot. The algorithm can allocate the offloading resources in advance for the hot areas where pedestrians and robots gather, so as to avoid the problem that the service robots in this region cannot offload tasks when nearby base stations are fully occupied. Simulation results show that the proposed algorithm can solve task offloading in hot crowded areas well, and its delay is better than the other compared algorithms.