Research On Autonomous Vehicle Trajectory Planning Algorithm Considering The Personalized Comfort Of Occupants

With the improvement of the level of automatic driving,the automatic driving system will gradually replace the operation of the driver,and the operating parameters of the system directly determine the driving style of the vehicle and the comfort of the occupants.In order to anthropomorphize the operation of the vehicle and personalize the ride comfort,it is necessary to identify the relationship between the vehicle driving planning parameters that characterize the driving style and the ride comfort of the occupants,and build an automatic driving trajectory planning algorithm that can comprehensively consider the individual ride comfort needs of the occupants.So as to realize the personalized automatic driving considering the comfort of the occupants.By building natural driving data and occupant comfort subjective and objective data collection platform,extract driving characteristic parameters that affect occupant comfort;build neural network model to identify the relationship between driving parameters and occupant’s personalized comfort;plan vehicle trajectory and driving speed according to comfort parameters,And set up a joint simulation platform to conduct closed-loop verification of personalized comfort,and obtain a trajectory that meets the individual comfort of occupants.The main contents of the research are as follows:Analyze the software,hardware,and communication platform architecture of autonomous vehicles,and build a real-vehicle natural driving data collection platform.Determine the planning parameters of the autonomous driving process,and perform time domain and frequency domain analysis on the parameters that affect the comfort of the occupants according to the standard ISO2631,and obtain the subjective and objective comfort parameters of the occupants and the driving parameters of the vehicle.Formulate data collection conditions,conduct real-vehicle data collection experiments,process the collected natural driving data and occupant comfort parameters,and obtain a model identification database.Use factor analysis to extract features of 18 driving parameters that affect occupant comfort,obtain 5 feature quantities through dimensionality reduction analysis,and compare factor weighting,non-linear weighting and neural network weighting algorithms to identify driving parameters and subjective and objective comfort parameters,The identification model of the relationship between planning driving parameters and occupant comfort is obtained.Considering the subjective comfort requirements of different occupants,a Kalman filter algorithm is built to quickly and accurately identify the individual riding needs of occupants.Analyze the driving environment information,safety field and behavior planning of autonomous vehicle following,confluence,lane changing and other working conditions in the highway scene.In the Frenet coordinate system,the horizontal and vertical position and speed are planned based on the fifth degree polynomial to obtain the trajectory The candidate sets are selected,and the trajectory clusters are selected through the safety and ride comfort verification algorithm,and the optimal candidate trajectory considering the comfort of the occupant is obtained.The occupant’s personalized comfort verification model is built in Matlab,and the trajectory planning algorithm considering the occupant’s comfort is closed-loop verified.The model includes upper-level perception,trajectory planning decision-making,trajectory following,vehicle dynamics,occupant comfort verification and other modules.Trajectory following is based on an adaptive model predictive control algorithm,and a Carsim-Simulink dynamic model is built to verify occupant comfort.The real-time observation of the comfort during the simulation shows that the high-speed automatic driving trajectory planning and following algorithm that considers the comfort of the occupants can meet the individual riding needs.