Research On Automatic Driving System For Unmanned Ground Vehicles Based On Human-Simulated Intelligent Control

Unmanned ground vehicle will be an important force in the army’s future, as well as the vital platform in the realization of information and unmanned war. By loading different equipment, it can replace human to complete various military missions in nuclear-biological-chemical, radiated, explosive and other harsh complex battlefield environment. With the emergence and practice of information warfare, modern war will become fiercer and crueler with greater damages, and be easier to cause casualties. It is increasingly critical to reduce casualties, and to complete the equipment handling, supply and escort, medical evacuation, reconnaissance and surveillance, detection of mines and combined operations in wars in non-human contact way. Under such military needs, unmanned ground vehicle can gives full play to its military potential and operational effectiveness, and bears important dangerous military tasks, which is one of the important directions of modern army weapons and equipment development. It will take a vital operational role in deciding the future war success and grabbing the battlefield initiative.In the development of unmanned ground vehicle, vehicle automatic driving system is mainly responsible for the real-time and accurate specific motion control operation of unmanned ground vehicle. Its performance has been always one of the key factors that affect vehicle intelligence capabilities, which is also the basis for unmanned ground vehicle’s environment perception, path planning and decision-making and reasoning, and also one of the key systems of unmanned ground vehicle. With the continuous development and maturity of the relevant technologies, the future vehicle automatic driving system should be possessed of a higher level of adaptability and autonomous driving capability.Due to the complex environment and a nonlinear time-varying system of unmanned ground vehicle, it is hard to describe in precise mathematical model. The traditional control based on models is difficult to meet the design specifications of the system. While the thought of constant combination of human intelligence on the basis of traditional control theory is an important path to solve such control problem.Based on the above background, this paper designs and realizes a universal vehicle automatic driving system of human-simulated intelligent control. The main work and innovation includes the following aspects: (1) The modeling and simulation of automatic driving system for unmanned ground vehicle. A kinetic model of automatic driving system and a vehicle dynamic model with coupling characteristics are built. The two typical road tracks of right angle bend and8-shaped road are designed to make simulation study on vehicle dynamic response characteristics of the road tracks. The basic motion performance of vehicle dynamics model is analyzed. The result of simulation study illustrates that the vehicle’s PID control based on the theory of "preview-following" realizes relevant accurate tracking, and obtains satisfactory control results, which lay the foundation for the design and realization of automatic driving system and real vehicle test.(2) The design and realization of universal executive device and control system for automatic driving system. The executive device adopts modular design approach. Actuator modules and corresponded automatic control modules are added in steering wheel, clutch pedal, brake pedal, accelerator pedal and shift knob of unmanned ground vehicle. Each module is connected with upper decision system via CAN bus. Then the automatic control of the vehicle can be achieved.(3) A human-simulated intelligent control algorithm for unmanned ground vehicle based on the automatic driving system designed in this paper is proposed. A motion controller of automatic driving system based on this algorithm is designed on operational control level and parameter correction level, and experimental simulation study is made on the designed right angle bend and8-shaped road. The simulation results show that the control effect of robustness and control accuracy of human-simulated intelligent controller is superior to that of PED controller, which lays the basis for further real vehicle test.(4) The actual control effect of real vehicle based on automatic driving system is tested and verified. On the basis of the previous study, the principled sample machine of designed automatic driving system is installed on the vehicle, and the real vehicle test is made, including vehicle’s longitudinal control, lateral control and comprehensive control, and analysis of the actual control effect of unmanned ground vehicle on PID control and human-simulated intelligence is made through the obtained experimental data. The real vehicle test shows that the designed automatic driving system realizes the accurate control of the vehicle’s steering wheel, brake pedal, accelerator pedal and shift knob, thus realizing the purpose of vehicle automatic driving control. Also, the human-simulated intelligent control is superior to PID control. It has laid a solid foundation for the further research on the unmanned transformation of manned vehicle and the intelligent control of unmanned ground vehicle.