In recent times, localization and positioning techniques have rapidly developed with the increasing demand for unmanned vehicles. Most positioning systems for land vehicles based on GPS-IMU, use a non-holonomic constraint to determine misalignment between sensor and vehicle body frame; however, misalignment estimation depending on non-holonomic constraint has limitations in high speed environments and there is a lack of observability for roll misalignment. This paper suggests an online misalignment estimation method under dynamic conditions that violates the non-holonomic constraint. It provides roll, pitch and yaw misalignment angles of IMU mounted on a vehicle, and corresponding sideslip angle of the vehicle at the position of IMU. The misalignment estimator is designed as a linear error state Kalman filter, which takes the results of a strapdown inertial navigation working simultaneously. Computer simulations and real environment experiments with consumer grade GPS and MEMS IMU are performed to demonstrate the performance and reliability of the given method.

近年来，随着对无人驾驶车辆的需求不断增加，定位和定位技术得到了迅速发展。大多数基于 GPS-IMU 的陆地车辆定位系统使用非完整约束来确定传感器和车身框架之间的错位；然而，依赖于非完整约束的错位估计在高速环境中存在局限性，并且缺乏对滚动错位的可观察性。本文提出了一种违反非完整约束的动态条件下的在线错位估计方法。它提供安装在车辆上的IMU的横滚、俯仰和偏航错位角，以及车辆在IMU位置处的相应侧滑角。未对准估计器被设计为线性误差状态卡尔曼滤波器，这需要同时工作的捷联惯性导航的结果。使用消费级 GPS 和 MEMS IMU 进行计算机模拟和真实环境实验，以证明给定方法的性能和可靠性。