Empirical Study on Premises Monitoring Algorithm Implementation in Mobile Robotic System

Mobile robotic systems have gained major attention from the whole global community for their unique features and ability to automate routine processes. Such systems can be implemented to optimize the monitoring process in dynamic dangerous or unknown environments. In this paper, we provide the algorithm for premises monitoring by a group of custom mobile robots. The main legitimate agents' task is to detect all intruder agents, located in the premise. The proposed approach is based on the optimization task, which requires legitimate agents to perform monitoring according to the optimization criteria. To test the proposed algorithm, we designed a physical testing ground, which includes models of the premises with rooms and a narrow corridor, legitimate agents, and an intruder. To assess the algorithm implementation effectiveness on mobile robots to detect the intruder, we compared it using two types of sensors: on-board camera with the computer vision application, and ultrasonic range finders. Experiment results showed that despite classification errors, occurred in the monitoring process, computer vision application provided faster and flexible detection than using ultrasonic range finders.