The Study Of Unmanned Vehicles Traveling Strategy Based On Multi-sensor Data Fusion

In order to improve road safety, modern unmanned vehicle system is widely use various types of sensors to obtain environmental information required for the vehicle. Faced with increasingly complex physical environment, traditional single sensor due to not meet the real-time and quickly provide a high accuracy and reliability of the monitoring and location information,nor have the autonomy, and strong ability to adapt to environmental change, anti-strong interference and high cost performance and other requirements in order to influence the development of unmanned vehicle driving strategy. We all know that to develop good driving strategies need to be monitored through a lot of environmental information on key parts of the vehicle driving state, and thus be able to better understand the performance of the entire vehicle, and develop timely and appropriate adjustment strategy to improve the reliability and safety of the vehicle sex.Therefore, this paper is explain how a more rational use of the sensor from the perspective of information on vehicle operating status detection and multi-sensor data fusion technology for unmanned vehicles traveling under the related research strategy. First, the unmanned vehicle based on environmental information collection system has been set up for each sensor location; once again, due to the complex environment of information, so that the uncertainty of the information obtained by the sensor with a certain degree, this paper consider all fusion algorithm on the basis of class, take advantage of the processing results D-S evidence theory to fuse a good deal of uncertainty problem. Finally, D-S evidence theory itself can not handle the uncertainty and problems of high conflict, this paper presents the theoretical matrix analysis corrected the problem; combined with the corresponding guidelines adopted by the fusion of the results obtained above, we can analyze decisions and develop unmanned vehicles travel policies.By actual road carried unmanned vehicles to experiments, the proposed algorithm can accurately represent the uncertainty prior knowledge of the driver, to ensure in case of failure of one or more sensors still have good fault tolerance.