Micro-electro-mechanical system (MEMS)-based inertial measurement unit (IMU) together with single-frequency Global Navigation Satellite System (GNSS) receiver provides a cost-efficient solution for outdoor navigation. An accurate initial alignment is necessary to the GNSS/IMU integrated navigation system to guarantee a good performance. Existing in-motion alignment methods mostly rely on the velocity derived from the GNSS receiver, which is hard to acquire when the receiver suffers from temporary signal blockage or deliberate interference. To mitigate the problem, we propose a carrier Doppler-based initial alignment method, which extends the availability of attitude estimation to cases where only three satellites are available. The proposed method is composed of two steps: the first step estimates a coarse initial attitude without any prior knowledge, and the second step dynamically refines the attitude estimation. Field experiment results show that the proposed method can provide attitude estimation with enough accuracy when only three satellites are available and has equivalent performance compared with the traditional optimization-based-alignment (OBA) method under general cases. The implementation of the proposed method is available on https://github.com/Wiggrim/D2R.

基于微机电系统（MEMS）的惯性测量单元（IMU）与单频全球导航卫星系统（GNSS）接收器一起为户外导航提供了一种经济高效的解决方案。GNSS / IMU集成导航系统必须进行准确的初始对准，才能保证良好的性能。现有的运动对准方法主要依赖于GNSS接收器得出的速度，当接收器遭受临时信号阻塞或故意干扰时，很难获得该速度。为了缓解该问题，我们提出了一种基于载波多普勒的初始对准方法，该方法将姿态估计的可用性扩展到只有三颗卫星可用的情况。所提出的方法包括两个步骤：第一步在没有任何先验知识的情况下估计粗略的初始姿态，第二步动态地改进姿态估计。现场实验结果表明，与一般的基于优化的对准方法相比，该方法在只有三颗卫星可用的情况下可以提供足够的姿态估计，并且具有同等的性能。提议的方法的实现可在https://github.com/Wiggrim/D2R上获得。