Towards a Low-Cost Robot Navigation Approach based on a RGB-D Sensor Network

Abstract

Motivated by the recent interest in the modernization of the traditional manufacturing facilities, this work focuses on the development of an automated warehouse where mobile manipulators autonomously navigate in the environment by exploiting additional spatial information gathered by a network of fixed RGB-D sensors. In detail, a navigation approach is proposed wherein both the mobile and fixed nodes composing the system are low-cost devices characterized by limited sensing and computing capabilities. A ROS2-based proof of concept of the outlined scenario is discussed and investigated with special regard to the practical challenges and the possible limitations. The results of some preliminary tests are reported to provide an intuition on the feasibility and potentiality of the designed approach in real-world industrial scenarios.