2019 年 8 巻 12 号 p. 501-506
In cyber-physical systems, the reliable data communication between the physical systems and computer systems are inevitable, and it is an essential issue to protect the systems from jamming attacks that obstruct wireless communication. Moreover, for practical uses of the systems, Quality of Service (QoS) is also important. In this paper, we consider a mobile robot and propose an anti-jamming mobile control system that uses QoS-based reinforcement learning. We show that the control system can find a policy that provides planning with a short path while avoiding jamming attacks.