Multi-sensor Fusion Localization and Navigation Algorithm for Hotel Service Robots in Complex and Dynamic Environments

Abstract

With the rapid development of intelligent hotel industry, service robot has become the core tool to improve service efficiency and quality. However, in a highly traffic-friendly and dynamic hotel environment, they often face issues such as positioning and navigation errors. Traditional single-sensor localization solutions are susceptible to light, enclosure, and electromagnetic interference and cannot meet high accuracy and durability requirements in complex scenes. This paper first developed a multi-sensor fusion system integrating laser radar, visual camera, inertial measurement unit (IMU) and ultra-wide band (UWB). Secondly, an improved adaptive Kalman filter based sensor data fusion algorithm is proposed to solve the data synchronization problem under dynamic noise. Finally, a dynamic route planning strategy was developed to integrate semantic information about the surrounding environment. The proposed method has a completion time of about 42s, an average speed of 1.90m/s, a completion time of about 55s and a speed of 1.45m/s for LiDAR-IMU, a completion time of about 72s and a speed of 1.14m/s for LiDAR-only. The navigation efficiency of the proposed method is better.