通信约束下多UUV协同目标感知技术研究

Aimed at improving the autonomous ability, collaborative target perception for multi-UUVs with communication constraints is studied, which lays a technical foundation for UUVs to carry out complex tasks in the marine environment. Firstly, a modeling method for multi-UUV collaborative perception with modular, parametric and hierarchical features is introduced. It satisfies the needs of componentized and generalized model design based on acoustic sensors. Secondly, algorithms of data time registration and multi-feature information fusion are put forward to perceive target information. It contributes to solve the issue of time synchronization in active-passive cooperative detection. At the same time, it can reduce the output data volume, which meets the constraints of communication bandwidth. Thirdly, there exist time delay and communication discontinuity for the interactive information in the multi-UUV system. A distributed data fusion algorithm based on the filtering algorithm, the newest available local estimate (NALE) mechanism and the chaotic ant swam algorithm is recommended to realize continuous and reliable tracking of dynamic targets. Fourthly, the heterogeneous multi-UUV system with master-slave structure has a characteristic of inconsistent situation in case of low communication rate and communication discontinuity. Therefore, a multi-stage cooperative strategy is proposed for an optimal performance of collaborative target perception. Finally, on the basis of the existing distributed simulation environment, simulation experiments are conducted to verify the effectiveness and reliability of the proposed algorithms.

本项目开展通信约束条件下多UUV协同目标感知技术的研究，旨在提升UUV的自主能力，为UUV在海洋环境下执行复杂任务奠定技术基础。第一，提出了一种模块化、参数化、层次化的多UUV水下目标感知建模方法，满足基于声传感器的组件化、通用化模型设计需求。第二，针对主被动协同探测中的时间同步问题，提出了基于数据时间配准与多特征信息融合的目标信息提取算法，降低了平台输出数据量，满足通信带宽的约束。第三，多UUV系统存在交互信息时间延迟和通信断续问题，因此提出了基于滤波算法、最新可利用估计判别机制与混沌蚂蚁算法的分布式数据融合算法，实现对动态目标的持续可靠跟踪。第四，对于主从结构的异构多UUV系统，针对通信速率低与通信断续导致的水下态势不一致性问题，提出了多阶段的UUV协同感知策略，实现联合感知效能最优。最后，依托于现有分布式仿真环境，开展仿真实验，验证所提出算法的有效性和可靠性。

围绕UUV自主能力形成和复杂任务执行的需求，针对通信约束导致的交互数据率低、时间延迟与通信断续等问题，本项目重点研究与突破了基于声传感器的目标感知通用化建模技术、基于分布式融合算法的多UUV联合目标感知技术和多层次协同感知策略等，搭建了多UUV协同目标感知集成验证平台。典型模拟环境下，4个UUV平台协同感知对不少于4个交叉目标具备持续跟踪定位能力；提出了多UUV协同目标感知策略，在典型模拟环境下，具备不少于8个节点的部署策略分析能力。形成了多UUV协同感知技术系列专利，具备向国防应用领域推广应用潜力，预期会应用到无人平台集群协同探测领域，推动多类型无人平台集群自主能力的生成。