Fusing low-cost sensor data for localization and mapping of automated guided vehicle fleets in indoor applications

Abstract

This paper describes the joint use of external and on-board low-cost sensors for small to mid-size automated guided vehicles (AGVs) in varying indoor applications. Localization and mapping for navigation purposes of multi-scale vehicle fleets is achieved by combining fixed, one-time calibrated cameras with odometry information and cheap LIDAR (Light detection and ranging) and collision avoidance sensors, which are mounted on the vehicles. Our approach is scalable and applies to different domains, as logistics in production environments or hospital facilities. Experiments show that the proposed system concept, based on low-cost units, reaches sufficient accuracy in real-time and could therefore be a promising solution for the target applications.