Simultaneous Fusion, Classification, and Traction of Moving Obstacles by Multiple Sensors Using Bayesian Algorithm Based on Fuzzy Dempster-Shafer Theory

In a near future, preventing the collisions with fixed or moving, alive, and inanimate obstacles will appear to be a serious challenge due to the increased use of Unmanned Ground Vehicles (UGVs). Lidar sensors and cameras are usually used in UGV to detect obstacles. We used a fuzzy algorithm in this study to examine the relationship between the obstacles, in which, the degree of confidence in each hypothesis is determined according to the design of the fuzzy system for extracting hypothesis. We proposed a novel approach in this paper, which uses a co-association matrix to gather all the information on targets and tracks generated for the Bayesian algorithm to predict and estimate the classes of obstacles and track them at sequential frames. To do so, the data generated by Lidar sensor and camera were fused by the extended Dempster-Shafer theory. This processed was used to assign the elements of the above mentioned matrix to, simultaneously classify and tract the obstacles in each frame. The simulation results of detecting obstacles revealed the benefits of the proposed method in classifying and tracking the obstacles. , we prepared the raw data and converted them into the probabilistic format to be used in the data fusion algorithms. We studied Lidar and camera raw data fusion layer and examined the process of deriving a hypothesis.