High-precision and Real-time Localization Algorithm for Automatic Driving Vehicles

For the problem of vehicle positioning in Vehicular Ad-hoc NETworks (VANETs), in order to improve the positioning accuracy and real-time performance, a high-precision and real-time localization algorithm for automatic driving vehicles is proposed, including two technologies based on Matrix Pencil (MP) and Non-Linear Fitting (NLF), and visual perception. The MP-NLF technology uses joint TOA/AOA estimation to locate vehicles with a single station, and introduces high resolution estimation technology to improve the estimation accuracy. The visual perception based technology completes the localization by extracting the feature information of visual perceptual images in positioning area, carries on the unscented Kalman filter combined with the inertial sensor information to further improve the positioning accuracy. The simulation results show that, compared with the traditional multipath fingerprinting algorithm, the proposed algorithm has better performance even in the case of low Signal-to-Noise Ratio (SNR).