基于地磁多参量多目标搜索的AUV仿生导航研究

Inspired by the geomagnetic navigation behavior of animals, this project is to study the bio-inspired navigation for AUV without prior geomagnetic map, whose realization mechanism and working principle remain unclear at present. In nature, the geomagnetic characteristics utilized by animals mainly depends on their sensitivity to the variation trend of geomagnetism, from the perspective of bionics, it manifests as the convergence of geomagnetic multi-parameter from the present values to objective values, therefore, geomagnetic multi-parameter plays a key role in bio-inspired navigation for AUV. However, without the prior knowledge of the parameter variation principle, it is of great difficulty to study the multi-objective convergence problem with multiple parameters. In this project, dynamic genetic evolution algorithm and multi-objective convergence theory are adopted to study the bio-inspired navigation for AUV based on geomagnetic multi-parameter. The research contents include (1) the action principle of geomagnetic multi-parameter to bio-inspired navigation; (2) bio-inspired navigation algorithm for AUV based on multi-objective search of geomagnetism with multi-parameter; (3) bio-inspired navigation with searching space malposition caused by magnetism abnormality or ocean currents disturbance and its boundary condition. By theoretical analysis, numerical simulation and experimental validation, we attempt to reveal the geomagnetic multi-parameter based bio-inspired navigation mechanism, master the working principle of bio-inspired navigation based on multi-objective search of geomagnetism with multi-parameter, elucidate the boundary condition for bio-inspired navigation caused by magnetism abnormality or ocean currents disturbance and realize the geomagnetism based bio-inspired navigation for AUV, which will be able to provide theoretical foundation for engineering practice.

本项目受生物地磁导航行为的启发，研究无先验地磁图的AUV仿生导航问题，目前对其实现机理及作用规律等尚不清楚。生物对地磁特征的利用主要依赖于其对地磁的趋势性敏感，从仿生角度看，体现为地磁多参量向目标参量的收敛，因此地磁多参量对仿生导航具有关键作用，但在缺乏参量变化规律先验知识的前提下，多参量的多目标收敛具有很大的研究难度。本项目主要采用动态遗传进化与多目标收敛理论，研究基于多参量收敛的AUV仿生导航。包括:(1)地磁多参量对仿生导航的作用机理；(2)基于地磁多参量多目标搜索的AUV仿生导航算法；(3)地磁参量异常、洋流干扰造成搜索空间错位的AUV仿生导航及边界条件等。通过理论分析、数字仿真及实验验证，揭示基于地磁多参量的仿生导航机理；掌握地磁多参量多目标搜索对AUV仿生导航的作用规律；明确地磁异常、洋流影响作用于AUV仿生导航的边界条件；实现基于地磁的AUV仿生导航，为工程实践提供理论依据。

本课题以利用地磁参量特性进行运载体自主导航为研究对象，提出研究不需要先验地磁参量数据库的仿生导航机理与方法。通过对动物迁徙、归巢等行为的分析与借鉴，提出一种不同于传统导航方式的新概念仿生导航技术，其特点是不依赖于先验数据，异于常规的地磁参量匹配导航方法，而是将地磁参量导航归结为对地磁特征的多参量多目标搜索问题，导航行为的依据是使得地磁多参量在同时同地收敛于目标值，从而达到利用地磁参量远程导航的目的。本课题地磁仿生导航研究综合考虑了生物对未知环境搜索的自适应特征，均衡进化策略的探索能力、地磁异常等，系统深入地论证了基于地磁多参量多目标搜索的仿生导航的机理与方法。. 首先，研究了生物地磁导航的行为机理，从仿生角度出发，将生物利用磁趋势的行为归结为参量规律未知情况下的地磁多参量向各自目标值收敛的问题。通过研究生物在地磁作用下的运动行为，明确了生物导航机理，建立了基于地磁多参量搜索的导航模型。.其次，提出利用磁趋势性搜索自主导航的机理，明确地磁多参量对仿生导航的影响规律。分析了地磁场各参量的不同特性，建立了仿生导航中运动路径参量和地磁多参量之间的约束条件，提出了基于时序进化搜索的仿生导航算法。通过对搜索多样性的偏向性分析，讨论了仿生导航算法的稳定性以及收敛特性。. 然后，针对地磁参量异常、洋流干扰等造成搜索空间错位下的导致的局部环路搜索等问题，研究了影响地磁多参量收敛特性的异常因素，明确了地磁参量异常、洋流干扰等对算法参量搜索影响的作用机理，分析并明确干扰因素的边界条件，实现环境异常影响下的地磁仿生导航。. 最后，利用XQ-ROS机器人搭建地磁仿生导航试验平台并进行导航实验，实验结果验证了地磁仿生导航方法的可行性和有效性。