Research On Multi-mode Switching Control Of Semi-active Suspension Based On Binocular Distance Recognition

With the development of the economy,drivers have higher requirements for vehicle driving experience,and vehicles with traditional passive suspension can hardly meet this need,making vehicles with active suspension and semi-active suspension appear in turn.Although the active suspension is adjustable in terms of stiffness and damping,it is only used on high-end models due to its high manufacturing cost,high energy consumption and poor maintenance economy.Semi-active suspension combines the characteristics of simple and reliable passive suspension structure and adjustable damping of active suspension,and its application of suitable control methods can make it close to the performance of active suspension.In order to optimize the control effect of semi-active suspension under different working conditions,this paper uses on-board binocular camera to distinguish different road conditions that the vehicle is about to pass through,and designs multi-mode control method switching rules for semi-active suspension based on magnetorheological dampers according to the recognition results of binocular camera to optimize the body attitude under different road conditions and improve ride comfort and handling stability.The main research work of the paper is as follows:Firstly,the dynamic modeling of magnetorheological semi-active suspension system and the construction of four-wheel road input model are completed.Through the damping characteristic test and parameter identification,a dynamic model of semiactive suspension vehicle based on magnetorheological damper is built to realize the real-time adjustment of output current through ideal damping force and suspension dynamic speed.Secondly,the binocular camera perception algorithm design is completed.The principle of camera calibration is introduced and the calibration experiment is completed by using Matlab binocular toolbox to obtain the internal and external parameters of binocular camera.Design a binocular camera vehicle bracket,and then load it to complete the image data acquisition,which provides data support for the design of perception algorithm.In order to distinguish the roads under different working conditions that the vehicle is about to pass,a deep learning target detection method for concave and convex road conditions represented by speed bumps and a traditional lane line detection method aiming at identifying the starting point of the circular curve of the lane line are designed.In order to set more accurate switching rules of multi-mode control methods,the real-time distance of the target is obtained by using the point cloud ranging function of binocular camera,and the real-time efficiency of the method and the accuracy of distance extraction are verified by experiments.Thirdly,the switching rules of multi-mode control methods of semi-active suspension and the specific control methods under each mode are designed.According to the perception results of the binocular camera previewing the driving road surface,the control switching strategy is designed.According to different control objectives,the suspension control is divided into three modes: straight pavement mode,obstacle pavement mode and curved road surface mode.The BP-PID control method is used to optimize the main control indexes in each mode.In order to improve the comprehensive performance of semi-active suspension at the same time,the intelligent group optimization algorithm is used to coordinate the four suspension control forces to achieve the improvement of comprehensive control objectives in each mode.The joint simulation platform of Python perception algorithm and Simulink controller is built.The data collected by binocular camera is used as the input of perception algorithm,and the perception results are output to Matlab.The Simulink control method is controlled by Matlab.The simulation proves the effectiveness of the multi-mode control method and its switching rules of semi-active suspension.Finally,the hardware-in-the-loop experiment of magnetorheological semi-active suspension control algorithm is completed.The hardware-in-the-loop test equipment and test scheme design are introduced.The hardware-in-the-loop test results show that compared with the passive suspension,the multi-mode control method can reduce the root mean square(RMS)value of the vertical acceleration of the vehicle body by 3.82%,the RMS value of the pitch angle acceleration of the vehicle body by 4.47%,and the roll angle acceleration of the vehicle body by 22.29%.The peak-to-peak values of the dynamic deflection of the left front suspension and the right rear suspension decrease by 5.64% and 6.64% respectively,and the RMS values of the dynamic deflection of the left front tire and the right rear tire decrease by 1.85% and 2.88% respectively.In the straight pavement mode,the maximum vertical acceleration of the body decreases by17.85%,the minimum value increases by 25.07%,and the RMS value decreases by22.23%.In the obstacle pavement mode,the maximum value of the body pitch angle acceleration decreases by 14.24%,the minimum value increases by 14.26%,and the RMS value decreases by 8.61%.In the curved road surface mode,the maximum body roll angular acceleration decreases by 32.37%,the minimum value increases by30.83 %,and the RMS value decreases by 32.56%.The effectiveness of the designed method is verified.