This work presents a contracting hierarchical observer that fuses position and orientation measurements with an IMU to generate smooth position, linear velocity, orientation, and IMU bias estimates that are guaranteed to converge to their true values. The proposed approach is composed of two contracting observers. The first is a quaternion-based orientation observer that also estimates gyroscope bias. The output of the orientation observer serves as an input for another contracting observer that estimates position, linear velocity, and accelerometer bias thus forming a hierarchy. We show that the proposed observer guarantees all state estimates converge to their true values. Simulation results confirm the theoretical performance guarantees.

翻译：本文提出了一种收缩式分层观测器，该观测器将位置与姿态测量值与惯性测量单元（IMU）数据融合，以生成平滑的位置、线速度、姿态及IMU偏置估计值，并保证这些估计值收敛至真实值。所提方法由两个收缩式观测器组成：第一个是基于四元数的姿态观测器，同时估计陀螺仪偏置；姿态观测器的输出作为另一个收缩式观测器的输入，该观测器估计位置、线速度及加速度计偏置，从而形成分层结构。理论分析表明，该观测器能保证所有状态估计值收敛至真实值。仿真结果验证了理论性能保证。