Research Of Domestic Mobile Robot SLAM System Based On Muti-sensor Fusion

Mobile robot technology combines the technology of motion control, image processing, sensor fusion, path planning, and wireless communications. SLAM (Simultaneous Localization and Mapping) technology is the key towards building autonomous robots. Multi-sensor information fusion for mobile robot navigation system can coordinate multiple sensors to solve the problem of nonlinearities and uncertainties in robot obstacle avoidance system, thus is of great significance to improve the robot navigation accuracy. Through the environmental information collected by the CCD vision sensor and ultrasonic sensors, this paper achieved building a real-time mobile robot with high accuracy and robustness of obstacle avoidance and navigation. This paper uses a high-performance S3C6410ARM11processor as the core hardware part instead of traditional PC, which saves power consumption and cost, enhances the mobility; the actual performance of the system is quite satisfactory. The main work and research achievements are reflected in the following aspects:1. Described domestic and international research results of the study, compared the two mainstream SLAM algorithms.2. Developed multi-sensor fusion system based on ARM11hardware design. The system is consisted of6ultrasonic sensors and1CCD vision sensor, and controlled by the ARM processor, which performs image acquisition and processing.3. Completed the migration of Linux system and developed application software based on embedded Linux.4. In this paper, a fuzzy PID control algorithm is introduced as navigation control algorithm. Finally, this paper performed EKF-SLAM Matlab simulation analysis.The system used ultrasonic sensors and visual sensor, combined with the optical encoder on drive motor, using EKF based multi-sensor information fusion algorithm, built ARM-Linux embedded development environment, and transplanted USB camera and the wireless network adapter driver. Embedded Web server Boa is used based on B/S mode, achieved real-time updates of the image and dynamic output of web pages. Simulation and experimental results show that the system can achieve multi-sensor fusion under Linux, as well as obstacle avoidance algorithm and path planning. The system is able to complete the autonomous navigation of mobile robots in domestic environment and motion control with high reliability and practicality.