用程序图替代深度强化学习，让无人机更安全高效导航
Tangled Program Graphs as an alternative to DRL-based control algorithms for UAVs
- 用简单程序构成图结构处理传感器数据
- 在未知环境中仅靠机载激光雷达导航成功
- 计算量小且决策过程可解释，适合高安全场景
深度强化学习（DRL）是当前自动驾驶控制最主流的AI方法，通过仿真训练的智能体可在真实环境中实现人类级表现。尽管在特定指标上表现优异，该方法存在计算开销大、决策不可解释等显著缺陷，难以应用于安全性要求高的控制任务。为此，本文提出使用纠缠程序图（Tangled Program Graphs, TPGs）作为DRL的替代方案。TPGs通过将输入信号由简单程序处理，并以图结构组合，具有更低的计算需求，且其行为可通过图结构进行解释。本文研究了TPGs在控制任务中的应用，重点关注仅依赖机载激光雷达（LiDAR）在未知环境中导航无人机的问题。实验结果表明，TPGs在控制任务中展现出良好前景。
原文摘要 · Abstract (English)
Deep reinforcement learning (DRL) is currently the most popular AI-based approach to autonomous vehicle control. An agent, trained for this purpose in simulation, can interact with the real environment with a human-level performance. Despite very good results in terms of selected metrics, this approach has some significant drawbacks: high computational requirements and low explainability. Because of that, a DRL-based agent cannot be used in some control tasks, especially when safety is the key issue. Therefore we propose to use Tangled Program Graphs (TPGs) as an alternative for deep reinforcement learning in control-related tasks. In this approach, input signals are processed by simple programs that are combined in a graph structure. As a result, TPGs are less computationally demanding and their actions can be explained based on the graph structure. In this paper, we present our studies on the use of TPGs as an alternative for DRL in control-related tasks. In particular, we consider the problem of navigating an unmanned aerial vehicle (UAV) through the unknown environment based solely on the on-board LiDAR sensor. The results of our work show promising prospects for the use of TPGs in control related-tasks.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。