基于动态多源信息的水下移动目标协同监测

The dynamic multi-source information of underwater moving targets, perceived by multi-underwater unmanned vehicles (UUVs) and multi-platforms, can collaboratively realize all-time and all-dimensional surveillance for underwater moving targets by means of effective information fusion and vehicles-platforms collaboration. The underwater acoustic channel is characterized as narrow-bandwidth, long time-delay, and strongly multipath. Meanwhile, the underwater environment is complex and dynamic, as well as UUV is of high degrees of freedom and strongly nonlinear. Constrained by aforementioned features, with the goal of enhancing the collaborative surveillance ability by extremely exploiting the multi-source information, the following three scientific problems are to be solved: the integration mechanism of detection and communication, the fusion mechanism of dynamic multi-source information, and the control theory of multi-UUV adaptive coordination. At signal-level, bandwidth and energy efficiency is improved to avoid interference by designing detection-communication integrated signal. At information-level, based on dynamic multi-source information, multiple targets are cooperatively tracked using target dynamic models and MCMC data association. At control level, dynamic task allocation is accomplished. UUVs cooperatively plan routes and control to track the moving underwater targets. Thus, intermediate-range effective detection and short-range continuous tracking of moving target are achieved, which supplies scientific theoretical foundation and key technologies for dynamic multi-info-based surveillance for underwater moving targets.

由多潜器多平台感知水下移动目标的各类动态多源信息，通过高效的信息融合和多潜器平台间的有机协作，可以实现对移动目标的全天候全方位协同监测。在水声信道窄带宽、长时延、强多径；水下环境复杂、多变；潜器自身高自由度、强非线性等诸多特性限制下，以充分挖掘多源信息提高协同监测能力为目标，围绕“探测通信一体化机理”、“动态多源信息融合机理”和“多潜器自适应协同控制理论”三个科学问题，在信号层面，通过设计探测通信一体化信号，提高带宽能量利用率，避免干扰；在信息层面，基于动态多源信息，通过数据关联，对移动目标动态建模，实现多目标的协同跟踪；在控制层面，进行动态任务分配，多潜器协作航迹规划以及追踪控制。以期实现对移动目标中远程的有效探测和中近程的连续追踪，为基于动态多源信息的水下移动目标协同监测提供科学理论基础和关键技术。

本项目围绕网络协同探测、动态多源信息融合与多潜器协同控制三方面开展研究，在理论方法和关键技术方面取得了如下创新成果：.1.提出了基于多载波的水下探测通信一体化系统架构。一体化系统分为收发合置和收发分置两种类型。该一体化系统可容纳正交频分复用(OFDM)和正交啾分复用(OCDM)两种多载波波形，为应对各种水下环境提供了灵活性。更进一步，项目组针对目标的网络协同探测，考虑了声线弯曲、时钟偏差、锚节点漂移等多种实际的水下不确定因素，提出了一系列的目标探测定位跟踪算法，提升了目标网络协同探测的精度。.2.对多潜器多平台收集的多源信息进行融合，提出多种数据关联融合方案。针对复杂环境会使得感知信息易受噪声和干扰的影响，设计了基于改进马尔可夫链蒙特卡洛(MCMC)的多目标数据关联算法。针对设备功能受限场景，提出了距离估计和位置求解的联合定位算法，开发了异步传输协议消除异步时钟的影响。利用洋流信息辅助的定位与跟踪框架对目标轨迹进行预测，设计基于置信度传播间歇式定位跟踪策略，实现多潜器分布式协同定位与目标跟踪，克服了目标高机动性和高通信能耗的影响，提高定位和轨迹跟踪精度。.3.提出了基于资源模型的任务分配方法，结合任务需求确定最佳的分配方案；通过研究潜器在网络信道中时延情况下的跟踪问题，设计了无模型的控制器以实现准确的轨迹跟踪。提出了协同包围策略，以实现对非合作目标的协同追踪围捕。.4.开发了整合嵌入式水声通信系统的水下潜器和仿虎鲸的背腹式推进水下潜器，并进行了实地实验。.项目研究成果包括英文专著2部，SCI论文39篇，EI收录文章34篇；申请发明专利2项，已授权国内发明专利11项，授权美国专利1项。获得2017年OI中国水下机器人大赛一等奖、2018上海市“临港杯”水下智能机器人大赛最佳创新奖。指导与本课题相关的博士生9名，硕士生15名。主办国际会议(WiOpt'18, ISAS'19)2次，发起水下网络相关主题会议NFC3(2017, 2021)2次。