A study on predictive information transfer for CSMA/CA-based autonomous decentralized control

摘要

This paper deals with autonomous decentralized control of mobile robots in CSMA/CA (Carrier Sense Multiple Access with Collision Avoidance)-based wireless networks. There is a problem that the quality of consensus control deteriorates due to packet collisions caused by hidden terminals and overlapping backoff times, and the loss of transmission opportunities within a control period due to the backoff times. In this paper, we propose a predictive information transfer method in which each agent not only transmits its own position information, but also simultaneously transfers predicted position information obtained from adjacent agents to achieve better control input. Simulation results show that the proposed method can improve the consensus control quality.