239 / 2023-09-29 20:02:18

Robust Adaptive High-degree CKF for Tightly-coupled GNSS UWB Integration

tightly-coupled GNSS/UWB integration,vehicle cooperative positioning,high-degree CKF,Huber cost function,improved fading factor

       For applications such as intelligent transportation systems and location-based services, maintaining the availability of relative location solution has become a research hotspot of collaborative positioning in vehicle ad hoc networks (VANETs). Currently, the Global Navigation Satellite System (GNSS) has become the preferred navigation mode to realize relative position perception and cooperative positioning due to its global coverage, cost efficiency and simplicity of implementation. However, complicated urban areas pose a great challenge to GNSS, i.e., the limited accuracy and availability of GNSS cannot meet the needs of location-based applications in VANETs. The advances in wireless networks have encouraged the development of cooperative positioning technology in VANETs. Some cooperative positioning techniques, based on vehicle-to-vehicle communication, are proposed to the performance of absolute positioning or/and relative positioning. As widely recognized, UWB (ultra-wide band), a ranging technique that uses the extremely large bandwidths enable high reliability and accurate range estimates, is commonly integrated with GNSS to completely exploit their individual advantages, leading to an optimistic scheme to improve the performance of vehicle cooperative positioning. This paper proposes a tightly-coupled GNSS/UWB integration scheme by using UWB-based range measurements for vehicle cooperative positioning in emerging intelligent transportation systems. In this scheme, the GNSS pseudoranges and Doppler shifts are shared between the communicating vehicles. Then each vehicles fuses the UWB-based range and GNSS measurements to obtain the relative position.

       However, the solution of GNSS/UWB tightly-coupled integration is usually affected by three main issues: model nonlinearity, systematic model error, and uncertainties of GNSS and UWB noises. Although it is possible to use additional high-precision sensors to solve these problems with redundant information, this leads to increased costs and also limits the application areas of integrated systems. In contrast, the data fusion method can provide a cost-effective solution to improve the performance of GNSS/UWB tightly-coupled integration for vehicle positioning.

       This paper presents a robust adaptive high-degree CKF with both robustness and adaptability to address the problems of GNSS/UWB tightly-coupled integration for vehicle positioning. In this strategy, the nonlinear estimation can be optimized by the higher-degree spherical-radial cubature rule. The Huber cost function is used to optimize the measurement information, and the residual covariance matrix is reweighted to reconstruct the measurement noise to suppress the influence of non-Gaussian noise. An exponential weighting method based on fading memory is adopted to establish the improved fading factor to overcome the interference of systematic model error. The experimental and analytical results show that the proposed algorithm can effectively hinder the interferences of both systematic model error and non-Gaussian noise on state estimation, and provide a practical control scheme for vehicle cooperative positioning with tightly-coupled GNSS/UWB integration.