基于环境感知的抗差自适应无源探测关键技术研究

With increasing complexity and volatility of the environment, the previous single-observer passive location and tracking technology is facing challenges such as vulnerability to outliers and inelasticity to the environment. This project introduces the concept of environmental awareness and deals with the key problems such as the adaptively robust algorithm for single-observer bearings-only location and tracking, and environment self-adaptation match of the algorithms on the basis of the exploration of the relationship between outliers and environment. Firstly, the typical measurement environment is built by designing the platform of semi-physical simulation. The environment parameters can be adjusted adaptively, which reveal the influence mechanism of the outliers to the environment essentially. Secondly, robust estimation is introduced to the bearings-only location problem, and the generalized M-estimation (GME) robust criterion is proposed so as to combine the validity and reliability of algorithm. Moreover, considering the possible abnormality in both state model and measurement model of moving target, the outlier discrimination static is constructed and the robustness and adaptiation of the algorithm can be improved in complex environment. Finally, the adaptive factor is created based on the environmental awareness information, which makes the algorithms changes with its surrounding correspondingly. The project's development provides new ideas and new ways for the passive location and tracking technology and is of great significance in theory and practice for the improvement of national space detection ability and realizing the intellectualization of non-cooperative passive location and tracking.

测量环境的复杂多变，使得传统单站纯方位无源定位跟踪面临易受异常误差干扰、适应环境能力弱等问题的挑战。为此，本项目引入环境感知理念，在探求异常误差与测量环境之间关系的基础上，针对单站纯方位无源探测，围绕抗差定位和跟踪算法，以及算法环境自适应匹配等关键问题展开研究。首先，构建半实物仿真平台模拟典型观测环境，实现参数和状态的自适应调节，从本质上揭示观测环境与定位跟踪异常误差的关系；然后，引入抗差估计思想，构建广义M估计抗差定位准则，实现定位算法有效性和可靠性的统一；第三，针对运动目标状态模型和观测模型异常带来的影响，重构判别统计量，提高复杂环境下无源定位跟踪算法的抗差能力和适应能力；最后，利用环境感知信息构建自适应因子，使算法能够根据环境变化进行自适应切换。本项目的开展将为无源定位跟踪的研究提供新思路和新方法，对于提升我国空间探测能力，实现非合作无源定位跟踪的智能化，有重要的理论意义和实用价值。

项目组按照研究计划实现了研究目标，针对单站纯方位探测面临的测量和模型的高不确定性，环境适应能力弱等问题，深入研究了基于环境感知的抗差自适应无源定位与跟踪方法。具体创新点如下：（1）基于环境感知的噪声模型自适应设计；（2）广义M估计准则下的抗差结构总体最小二乘定位算法；（3）基于误差判别基础的运动目标无源跟踪抗差算法；（4）基于判决延迟策略的抗差自适应滤波方法等；通过上述研究成果，有效提升了无源定位跟踪算法的测量精度与性能。