Research On Key Technologies Of Intelligent Autopilot For Automobiles

In the context of artificial intelligence,the combination of the traditional automotive industry and AI has created a new research area of self-driving.In addition to autonomous driving from the structure of the car itself,self-driving can also be achieved by installing a autopilot on a conventional vehicle.Compared with traditional autopilot,this research in this paper adopts the overall scheme of humanoids.It can adapt to different cab environments through the design and control.It can realize self-adaptation of the driving process through its own automatic adjustment.The following studies were mainly conducted:Firstly,the background and significance of the research on self-driving is expounded,the domestic status quo of autopilot and abroad is analyzed,and a new type of autopilot for vehicle driving in abroad is briefly analyzed,and the contents of each chapter are arranged.Secondly,the shortcomings of the traditional autopilot is analyzed,and the scheme of using the multi-freedom humanoid shape in the intelligent autopilot is determined and the driver's behavior is fully imitated.Based on the overall scheme,the system autopilot composition system is analyzed.The base subsystem,joint subsystem,control subsystem,energy subsystem and environment perception subsystem are analyzed.The robotic arm adopts a 7-DOF scheme.The mechanical leg that controls the brake pedal and the accelerator pedal adopts a 6-DOF scheme.The joint subsystem adopts the design idea of a modular joint,and the motor and reducer are directly connected to form a preliminary calculating output indicators,corresponding to hierarchical control.Using single-joint position control based on BP neural network PID,the pose description methods and D-H parameters for multiple degree of freedom control is described.Then,the autopilot joint function was analyzed,and the joint motor and reducer were designed.The brushless DC motor was selected as the joint motor,and the joint motor was checked and analyzed using magnetic circuit method and finite element method to obtain the internal characteristic parameters,external characteristic curve,magnetic field distribution and temperature rise of the motor.After several optimization analyses,the motor assembly is realized;the reducer selects the small tooth number difference with a large speed ratio,the gear ratio of the internal gear is calculated after the reducer speed ratio is determined.The gear strength is verified by the finite element method and reducer assembly is realized.Modular joint integration is completed after the brushless motor and reducer are determined.Meanwhile,on the basis of modular joints,joint links are designed,lightweight optimization is achieved by means of shape optimization,and the strength,stiffness,and stability of the links are checked,and then the integration of the autopilot base subsystem is completed.Based on the control research of the base subsystem,the single joint is deduced based on the BP neural network PID,which corresponds to hierarchical control;the forward kinematics solution and the inverse kinematics solution are solved for the multi-freedom manipulator,corresponding to the middle control.Finally,simulation of single-joint control and multiple-freedom robotic arms and mechanical legs.Through a single joint control simulation,a PID controller based on BP neural network was verified;simulation of a 7-DOF autopilot arm and a 6-DOF autopilot leg was obtained,and a modular joint was obtained at steering wheel control,shifting,and stepping on brake and throttle.The output torque and speed conditions verified that the joint's output requirements were permit.