Combined Vehicle Tracking and Obstacle Detection Based on D-S Evidential Reasoning

Abstract

Intelligent vehicle tracking can form vehicle formation or have more applications in AGV navigation vehicle. In the case of obstacle detection, when we use line-of-sight positioning and half-line-of-sight positioning, the difference of data obtained by these two methods is different, which can effectively identify obstacles. The predicted distance is divided into X direction and Y direction. The interval probability statistics of the difference between the two directions are carried out. The probability of obstacles is predicted separately by the difference between the two directions, and the D-S evidence reasoning algorithm is applied to the detection of obstacles. Experiments show that the combined positioning can not only control the positioning error within 1m to about 4cm, but also achieve 91.5% accuracy of obstacle recognition.