DWE-Based SRIBO: An Efficient and Resilient Single-Range and Inertial-Based Odometry With Dimension-Reduced Wriggling Estimator

The single-range and inertial-based positioning approach is a low-cost and lightweight solution for multirotor flying robots in challenging scenarios where visual perception degradation occurs. Unfortunately, it may still encounter local observability issues in certain motion patterns. In this study, we address this limitation by proposing a computationally efficient dimension-reduced wriggling estimator, which enhances the positioning performance and robustness in an expanded estimation horizon. This estimator slides the estimation horizon with output matrices approximation based on adjacent historical estimation sequences. It then reduces the computational complexity through dimension reduction using a polynomial approximation approach. The dimension-reduced wriggling estimator enables each estimation to cover a sufficiently long interval with linear programming, thereby enhancing the degree of observability and robustness by leveraging adequate measurement redundancy. Additionally, this estimator can incorporate additional sensors, such as an optical flow sensor. We also prove the theoretical convergence and numerical stability of the position estimator. Indoor and outdoor experiments validate that the proposed single-range and inertial-based system achieves decimeter-level precision at a high frequency of hundreds of Hertz while remaining resilient to sensor failures. Particularly, this positioning system demonstrates excellent robustness in dark, smoky, and blinking scenarios, where commonly used optical sensors fail to provide consistent estimation.