基于仿人发育的混合空域智能无人机 防碰撞策略研究

Anti-collision of Intelligent UAVs in Mixed Airspace Based on Humanlike Development

摘要:

随着低空开放政策的实施，无人机将会脱离隔离空域，进入低空与各式各样飞行器混合进行飞行。而防碰撞能力作为无人机的关键技术，在混合空域中保持无人机安全飞行起到了至关重要的作用。通过对人的发育式学习防碰撞机理进行分析，建立了面向混合空域中密集障碍的实体模型和发育模型，最终构建了基于仿人发育的无人机防碰撞方法。仿真实验表明，提出的方法不仅使无人机可以通过学习与发育不断提高飞行安全性，而且还可以随着环境熟悉度的提升，加快无人机防碰撞行为，提高了无人机在混合空域密集障碍中的安全性与智能性。

Abstract:

Along with the strategies of opening low-altitude airspace, the UAVs will get out of the isolation of the airspace, and fly with diverse aerial vehicles. As the key technology of UAVs, autonomous anti-collision plays a important role in safe flying, especially in mixed airspace. By analyzing humans’ developmental anti-collision mechanism, the specific models and developmental models have been proposed in this paper, in order to construct UAVs’ anti-collision method based on humanlike development. The simulations illustrate that our method can not only improve the safety of UAVs in mixed airspace via learning, but also develop itself to avoid collision faster along with the familiarly of environment, which has improved the safety and intelligence of UAVs in mixed airspace with crowded threats.