高动态环境下的多普勒测速脉冲星导航方法研究

In the encounter period of deep space exploration which is a high dynamic environment, the autonomous navigation is the key to a successful mission. In the encounter period, as the acceleration of spacecraft varies fast, the orbit dynamic model has high dynamic characters including nonlinear and fast time-varying. The velocity error calculated by the orbit dynamic model is very large, which affects the whole navigation system performance. Besides the pulse time-of-arrival, the newly-developed Doppler velocimetry pulsar navigation method can also provide the Doppler velocity directly, and applies to the encounter period. We research the Doppler velocimetry pulsar navigation from the aspects of the prediction model, the observation information and the filter. .① The prediction model integrated by the high dynamic orbit dynamic model has low accuracy. To solve this problem, the efficient integral error compensation method is adopted to construct the real-time and highly precise prediction model. .②The raw data amount of pulsar signals is very large. The computational load of the Doppler velocimetry, which is proportional to the raw data amount, is very large. To obtain the observation real-time, the pulse time-of-arrival and Doppler velocity joint estimation method based on compressive sensing is proposed. .③Subjected to the nonlinear prediction model, the nonlinear prediction bias caused by the state estimation error is large. The relationship between the prediction bias and the state estimation error is analyzed. The high dynamic navigation filter with the relationship between the two is proposed. .The investigated achievements will provide the spacecraft with the time-real and accurate autonomous navigation information in the encounter period.

在深空探测捕获段这一高动态环境下，自主导航极为关键。在捕获段，航天器加速度变化快，轨道动力学模型呈强非线性、快时变等高动态特点。结合该模型计算出的速度误差较大，这将影响导航系统性能。除脉冲到达时间外，新兴的多普勒测速脉冲星导航还可直接提供多普勒速度这一观测量，适用于捕获段。本项目面向捕获段，从预测模型、观测量和滤波器等方面开展多普勒测速脉冲星导航研究：①由高动态轨道动力学模型积分而成的预测模型精度低。为解决该问题，拟通过高效的积分误差补偿方法构造实时高精度的预测模型。②脉冲星信号原始数据量大，与其成正比的多普勒测速计算量极大。为实时获取观测量，设计基于压缩感知的多普勒速度和脉冲到达时间联合估计方法。③由于预测模型强非线性，状态估计误差引起的非线性预测偏差较大，通过研究预测偏差与状态估计误差的耦合关系，提出二者相关条件下的高动态导航滤波器。以上成果将在捕获段为航天器提供实时高精度的自主导航。

在深空探测捕获段这一高动态环境下，自主导航极为关键。在捕获段，航天器加速度变化快，轨道动力学模型呈强非线性、快时变等高动态特点。结合该模型计算出的速度误差较大，这将影响导航系统性能。除脉冲到达时间外，新兴的多普勒测速脉冲星导航还可直接提供多普勒速度这一观测量，适用于捕获段。本项目开展多普勒测速脉冲星导航研究,取得了一些有意义的结果：.(1) 基于星光多普勒的状态预测模型。在该方法中，星光多普勒测量值不再作为用于更新的测量值，而是取代轨道动力学模型直接提供航天器速度信息。在高动态环境下，仍能提供高精度的预测状态。.(2) 基于压缩感知的脉冲星多普勒测速和到达时间估计方法。该方法利用了脉冲相位偏移而非轮廓畸变估计频移，实现了快速脉冲星多普勒测速和到达时间联合估计，将计算量降低了四个数量级。.(3) 面向深空捕获段的天文测角/测速/测距深度组合导航系统及可观测性分析方法。”深度”二字体现在测角和测速信息抑制了脉冲星信号的多普勒效应。提出了分数阶可观测性矩阵构建方法和结构性条件数，能准确反映天文测角组合导航性能。为了优化多普勒导航恒星与脉冲星的组合，提出了冗余导航系统中的多普勒导航恒星选择策略，实现了优化组合，在仅增加一颗多普勒导航恒星的条件下，提高了组合导航精度。. (4) 脉冲星到达时间快速估计方法。利用三阶累积量、萤火虫群优化算法、两级压缩感知、小波变换等数学工具，优化脉冲星到达时间估计方法，降低了计算量。. (5) 太阳耀斑到达时间差分量测方法。利用太阳直射光与行星反射光的强度相关性，提出了一种太阳耀斑到达时间差分导航方法。传统天文自主导航方法无法直接提供深空探测器与目标天体之间的距离这一关键信息，而该方法实现了这一点。.以上成果构成了一套具有自主知识产权的高动态环境下的多普勒测速脉冲星自主导航解决方案，预计将在捕获段为航天器提供实时高精度的自主导航。