Localisation and navigation of a robot platform using laser scanner and Adaptive Monte Carlo Localisation (AMCL) for an indoor horticultural application

Abstract

Automation and robotics in horticulture have the potential to replace manual work in repetitive activities, thereby achieving economic advantages. Reliable, sensor-based localisation and navigation in indoor environments would allow easy integration of robotic solutions into existing installations. The aim of this article is to evaluate this localisation accuracy of a robot developed for indoor applications in horticulture. For this purpose, 10 target positions are approached with a robot platform in a test environment and the robot end position and orientation are recorded by a motion capture system (MCS). The result shows that the target positions were reached with an average distance of 45.1 mm and an average angular deviation of 0.4°.