无人机类脑吸引子神经网络导航技术

当前无人机在非结构化或未知环境下飞行主要采用SLAM进行导航与定位，存在如下突出问题：依赖高精度昂贵激光雷达等环境感知传感器；需要建立准确世界和无人机物理模型；受环境影响较大；自主智能水平较低，无法较好地满足无人机对导航系统的要求，需要发展自主智能的导航方式。基于吸引子神经网络的类脑导航技术，无需训练模型参数，不依赖高精度传感器，无需精确建模，且复杂环境下鲁棒性较强，具有解决上述问题的潜力。简要阐述了动物大脑导航机理，分析了吸引子神经网络和基于吸引子神经网络的类脑导航关键技术，最后讨论了吸引子类脑导航技术在无人机应用中的挑战。

Attractor Neural Network-based Brain-inspired Navigation Technology for UAV

Currently, UAV mainly uses SLAM to navigate and locate itself in unstructured or unknown environment with the following problems: depending on high-precision and expensive sensors such as lidar; using probability method to build accurate physical model of the world and UAV; being greatly affected by environment; poor autonomy and intelligence, which make the SLAM method impractical in UAV. Thus it is urgent to develop a new type of navigation method for UAV. Brain-inspired navigation technology based on attractor neural network (ANN) neither needs training model parameters or accurate modeling, nor relies on high-precision sensors, but has strong robustness in complex environment, which make the technology a good candidate for solving the above problems. This paper briefly describes the mechanism of animal brain navigation, analyses the performance of ANN and ANN-based brain- inspired navigation technology, and finally discusses the challenges of ANN-based brain-inspired navigation technology in the application of UAV.