Position Determination Algorithm For Intelligent Unmanned Vehicles Based On LightGBM Model Assistance

Navigation and positioning are among the most widely used technologies in intelligent transportation systems,advanced vehicle control,and vehicle safety systems.The main idea behind its algorithm is to use satellite data and vehicle dynamics data to calculate the current specific position of the vehicle.The algorithm’s robustness determines the accuracy of the information output from the vehicle navigation system and its ability to adapt to the environment.To solve the problem of data accuracy degradation of vehicle GNSS/INS integrated navigation systems when the GNSS signal is unavailable or there is a GNSS outage,this dissertation proposes a position determination algorithm for intelligent unmanned vehicles based on Light GBM model assistance,and analyses and verifies the performance and accuracy of the algorithm through road tests.The main research contents of this dissertation are as follows:(1)This dissertation collates knowledge about ensemble learning,GBDT,and analyses the basic principles and optimisation methods of Light GBM.The specific training process of the Light GBM regression task is described.(2)This dissertation introduces the INS Mechanization,and provides a detailed overview and analysis of the GNSS/INS Loosely coupled integration navigation algorithm.Accordingly,an intelligent unmanned vehicle integration navigation platform is designed and implemented.The platform consists of a self-developed GNSS/INS integrated navigation system,an intelligent unmanned vehicle,and upper computer software.The platform can acquire and process the navigation data of the intelligent unmanned vehicle.The hardware device of the integrated navigation system includes a power supply board,a data fusion board and a data acquisition board.The upper computer software includes a serial port class,a serial port connection module,a data display and data persistence module,and a data visualisation and trajectory display module.The dissertation outlines the development process of the hardware device,the process consists of driver development,the implementation of the logical functions;and the design ideas and development process of the upper computer software.The dissertation gives an overview of the practical problems encountered in the development process and the corresponding solutions.In addition,field tests are conducted by using intelligent unmanned vehicle.The field tests proves that the platform can complete the acquisition of IMU raw data and GNSS raw data.Besides,the platform can carry out stable,accurate solutions for integration navigation and visualize the data of integration navigation.(3)This dissertation improves the existing GNSS/INS integration methodology for land vehicle navigation based on the AI method.Firstly,a GNSS/INS integration methodology for land vehicle navigation based on Position Update Architecture(PUA)using Light GBM regression for predicting the position of a vehicle during a GNSS outage is presented.It uses Light GBM to model the relationship between INS data and vehicle position changes.On-board INS and GNSS data are collected when the GNSS signal is available and used to train the PUA-Light GBM model;in the event of a GNSS outage,INS data are used as input to the PUA-Light GBM to predict the change in vehicle position.Secondly,a vehicle navigation data acquisition system was designed for model validation.This included a self-developed GNSS/INS integrated navigation system and a Novatel pwrpak7-e1 GNSS/INS integrated navigation system for data acquisition on six road segments,respectively.Finally,the collected data were used for machine learning training of the PUA-Light GBM model and the existing PUA-Random Forest model.As a result,the PUA-Light GBM predicts the vehicle position with less error in the event of a GNSS outage and takes less time to train.It was also demonstrated that by allowing the model to be dynamically trained or updated while the vehicle is moving,the PUA-Light GBM could adapt perfectly to the predictions of vehicle position changes in different complex road segments.