Optimizing the Scheduling of Localization Sensors for an Intelligent Robot by Reinforcement Learning

This paper studies the scheduling of on-board localization sensors for an intelligent robot navigating an indoor environment, aiming to reduce power consumption and extend operational lifetime. We propose a novel offline scheduling method that transforms the problem into a Markov decision process model, solved via reinforcement learning. The precomputed scheduling policy dynamically selects sensor combinations based on real-time resource availability and environmental dynamics, adaptively balancing localization precision and computational efficiency. Field tests with a mobile robot in real-world environments demonstrate the practical effectiveness of the proposed approach.