Landmark-based navigation of mobile robots in manufacturing

Abstract

Landmark based navigation of autonomous robots has been widely used in manufacturing industry. Such a navigation strategy relies on identification and subsequent recognition of distinctive environment features or objects that are either known a priori or extracted dynamically. This process is inherently difficult in practice due to noise in sensors and changes in the real world. The paper presents a navigation algorithm that locates the robots and updates landmarks in a dynamic manufacturing environment. A key issue being addressed is how to improve the localization accuracy for mobile robots in a continuous operation, and how to initialize and recalibrate the robot position when necessary. Kalman filter algorithms are adopted to integrate odometry data with scanner data to achieve the required robustness and accuracy. Kohonen neural networks are used to recognize landmarks in the process of initialization and re-calibration of the robot position.