Object detection and tracking using sensor fusion and Particle Filter

Abstract

This paper presents a moving object tracking system with a Particle Filter algorithm. A software tool is developed to track an unknown moving object in a sensing region occupied by other dynamic objects. Several components are used to determine objects, to self-localize, and to match the determined objects iteratively in conjunction with the previously determined objects. Each object is labeled with a unique identification number. Main sensor is a Laser Imaging Detection and Ranging (LIDAR) to sense the objects, Inertial Measurement Unit (IMU) is used to localize the ego-vehicle and wheel odometer is used to improve the accuracy of positioning. The Particle Filter algorithm predicts self-position, utilizing the data received from both the IMU and the odometer. Performance and detection accuracy tests are carried out using various sized objects, as well as different environmental settings in order to conduct a comparison analysis for the gathered data.