Design And Implementation Of Laser Radars Based Object Detection System For Intelligent Vehicle

The traditional target detection methods are usually based on one single sensor such as laser radar or the visual sensor, which can not provide sufficient environment information to the intelligent vehicle system. And the visual sensor is also inappropriate for the system which needs high real-time performance and accuracy.In this paper, the laser radars data fusion methods were analyzed. Using the grid voting method and the target feature method, we designed a laser radars target detection system to detect the dynamic targets and static targets, which improves the real-time performance and reliability of the target detection. The main contribution of paper is organized as follows:1. The background and development of laser radar target detection technology were summarized in this paper. We introduced the basic principles of multi-sensor fusion technology and target detection technology, and analyzed the current problems and proposed the research targets.2. Considering that the target detection of intelligent vehicle needs all-around coverage, high accuracy, high real-time performance, we designed a voting fusion rule and a grid voting method for laser radars data fusion. The experiment results showed that this method can meet the requirements of intelligent vehicle target detection, and has high real-time performance.3. Because that the above-mentioned grid voting method can not detect dynamic targets, we proposed a laser radars target fusion method based on the target feature to solve this problem. To realize multi-sensor data fusion and target detection, the geometric features and motion features of targets were respectively obtained by cluster matching and target tracking, which improve the ability of dynamic target detection. In addition, we also used the ICP algorithm to solve the problem that system can not accurately get motion features of targets at a such situation that GPS signal is weak in room and tunnel.4. In order to meet requirements of the practical application, we designed a laser radars target detection system for intelligent vehicle, and then introduced each subsystem in detail. Finally, the experiment results of real vehicle testing showed that this system could real-time detect surrounding targets accurately.