Research Of Driving Characteristics Of The Fixed Axle Transmission And The Second-order Planetary Steering Tracked Vehicle Based On Statistical Learning

With the continuous research and development of military unmanned vehicles,the unmanned transformation and unmanned platform research based on the fixed axle transmission and the second-order planetary steering tracked vehicle can maximize the effectiveness of the decommissioned equipment,and have a very significant advantage in the complex cross-country environment and special military use.In addition,the fixed axle transmission and the second-order planetary steering tracked vehicle have complex nonlinear relations and strong coupling characteristics between various systems.The steering system is relatively simple and uncertain.So in order to achieve better vehicle driving control effect,it can not only take into account the characteristics and motion laws of the vehicle platform itself.It is also a combination of pilot manipulation experience and characteristics.The research of driver characteristics is of great significance for improving the accuracy of intelligent vehicle control.In this paper,based on the statistical learning method,the driver control characteristics of the vehicle are studied and the driver model of the vehicle state switching is established to realize the accurate state prediction in the vehicle running process.On the basis of in-depth analysis of the mechanical structure and driving characteristics of the vehicle,the importance of state switching and pilot study on the motion control of tracked vehicles is proposed in this paper.A real vehicle data acquisition system is set up to collect a large number of experienced driver’s experience data under the real scene and establish the corresponding database.Based on the GMM,the steering control movement of the vehicle is extracted and the corresponding base element switching sequence is extracted.The graph theory method is used to express the vehicle transverse longitudinal basic element.State and local sequence diagrams for handover and composite relationships.Then,this paper uses the multiple GMM and SVM to establish the vehicle transverse longitudinal state switching model based on the state switching guidance of sequence diagram,.The classifier is trained for each state switching node and the predicted state sequence is combined,and the optimal state parameters of each node are selected.Improve the accuracy and accuracy of the model.Finally,the switching behavior test of each vehicle is selected to verify the state switching model reliability.Based on the vehicle test platform which has been built in Beijing Institute of Technology,combined with the steering characteristics of the tracked vehicle,the pilot test is carried out on the real cross-country road and the result of the test is analyzed.