复杂环境下自主车辆地面物理特征的试验研究-机械工程专业论文.docx

复杂环境下自主车辆地面物理特征的试验研究 复杂环境下自主车辆地面物理特征的试验研究 EXPERIMENTAL RESEARCH FOR PHYSICAL CHARACTERISTICS OF SURFACE BASED ON AUTONOMOUS VEHICLE UNDER COMPLICATED ENVIRONMENT ABSTRACT The main purpose of this study is to improve the maneuvering performance of autonomous vehicles, then realize the navigation of intelligent vehicle and accurate prediction of maneuverability. This topic will explore such a detecting technology: Combining with the information of different physical characteristics and related information collected by the vehicle, considering the characteristics of different trial conditions, and then based on the analysis of the specific situations, we deeply discussed the interactions between vehicle and the surfaces, and we also discussed how to collect the physical characteristics information. Therefore we designed and developed a MATLAB-based Bayesian network terrain classification program. The realization of this program is a foundation for the experimental research. The vibration test system this paper selected is making up of acceleration sensor and a collaborative SQLab II multichannel measurement system. In this paper, we presented a single and definite vibration test environment, and we regulated many representative trial terrains such as asphalt and grass which basically include various terrains in outdoor environment, and satisfying the needs of vibration test. We specified the vibration test data processing system is mainly composed of ArtemiS analysis software and MATLAB-based power spectral density of processing. By ArtemiS vibration analysis software, we obtained the time-domain data of vibration ， and then we used MATLAB to analyze the power spectral density, realized frequency-domain process of time-domain signals. In this paper, we researched the Bayesian network structure, briefly summarized the basic theory of Bayesian network structure and learning algorithm of Bayesian networks, and we also pointed out the advantages and d