仿猛禽视顶盖信息中转整合的加油目标跟踪

PDF下载HTML

Title:

Aerial refueling target tracking using a falcon visual tectum information integrating like method

作者:

李晗¹, 段海滨¹, 李淑宇², 丁希仑³

1. 北京航空航天大学 自动化科学与电气工程学院, 北京 100083;
2. 北京航空航天大学 生物与医学工程学院, 北京 100083;
3. 北京航空航天大学 机械工程及自动化学院, 北京 100083

Author(s):

LI Han¹, DUAN Haibin¹, LI Shuyu², DING Xilun³

1. School of Automation Science and Electrical Engineering, Beihang University, Beijing 100083, China;
2. School of Biological Science and Medical Engineering, Beihang University, Beijing 100083, China;
3. School of Mechanical Engineering and Automation, Beihang University, Beijing 100083, China

关键词:

自主空中加油; 猛禽视觉; 视顶盖; 双凹结构; 目标跟踪; 神经网络

Keywords:

autonomous aerial refueling; falcon visual system; tectum; bifovea; target tracking; neural net work

分类号:

V249.32

DOI:

10.11992/tis.201909005

摘要:

无人机自主空中加油是当前作战模式下非常重要的军事能力之一。空中加油对接阶段的视觉辅助导引是对接阶段导航的研究重点，而加油目标跟踪则是其中的重要一环。本文通过对猛禽优异的视觉系统与视觉导航能力的研究，发现其视觉通路中视顶盖的信息中转整合能力对于解决跟踪问题具有一定指导作用。通过对这一信息处理模式的模拟，研究了一种仿猛禽交叉通路的神经网络，并针对空中加油视觉任务中可能出现的目标丢失与再入问题，引入猛禽双中央凹扩大搜索策略，设计了一种完整的模拟猛禽视顶盖信息中转整合的目标跟踪网络。仿真实验验证了所提方法的可行性和有效性。

Abstract:

Autonomous aerial refueling is one of the most important military capabilities under current combat situation In docking phase navigation, visual-aid guidance during the docking phase is a main research focus, of which the tracking of refueling target is an important part and challenge. During researching on the excellent visual system and visual navigation ability of falcon, it was found that the information transfer and integration ability of the visual tectum could help to solve the tracking problem. By simulating information processing mode, constructing a neural network with cross-pathway, and introducing the target out of view and reentry problem that may occur in the aerial refueling visual task, the raptor bifovea based searching strategy is introduced to construct a complete tracking method. Simulation and experiments are done to verify the feasibility and effectiveness of the proposed method.