Localization of Unknown Networked Radio Sources Using a Mobile Robot with a Directional Antenna

In this paper, we present algorithmic developments for a single mobile robot equipped with a directional antenna to localize unknown networked radio sources. We assume that the robot only senses radio transmissions at the physical layer and that the network has a carrier sense multiple access (CSMA)-type medium access control (MAC) layer. The total number of radio sources are assumed unknown. We first formulate the localization problem and then propose a particle filter-based localization algorithm. The algorithm combines a new CSMA model and a new directional antenna model. The new CSMA model provides network configuration data, such as the network size and the probability of collision, when listening to network traffic. The directional antenna model enhances the efficiency of robot motion. The new combined sensing model is capable of handling transmission collisions during localization. The overall algorithm runs in O(nm) time for n particles and m radio sources at each step. The numerical results show that the algorithm can localize unknown networked radio sources effectively.