arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2009.11212cs.ROcs.LG
收藏

基于鲁棒强化学习的仿真与现实自主驾驶Agent

Robust Reinforcement Learning-based Autonomous Driving Agent for Simulation and Real World

Budapest University of Technology and Economics(布达佩斯技术与经济大学)

Péter Almási, Róbert Moni, Bálint Gyires-Tóth

摘要:

深度强化学习(Deep Reinforcement Learning,DRL)近年来已成功应用于解决复杂棋盘游戏和计算机游戏等挑战。然而,用DRL解决现实世界的机器人任务似乎是一个更困难的挑战。理想的方法是在模拟器中训练代理,然后将其转移到现实世界中。尽管如此,由于这些差异,在模拟器中训练的模型在现实环境中往往表现不佳。在本文中,我们提出了一种基于DRL的算法,能够使用深度Q网络(DQN)进行自主机器人控制。在我们的方法中,代理在模拟环境中进行训练,并且能够在模拟和真实世界环境中导航。该方法在Duckietown环境中进行评估,其中代理必须根据单目摄像机输入来跟踪车道。经过训练的代理能够在有限的硬件资源上运行,其性能与最先进的方法相当。

英文摘要:

Deep Reinforcement Learning (DRL) has been successfully used to solve different challenges, e.g. complex board and computer games, recently. However, solving real-world robotics tasks with DRL seems to be a more difficult challenge. The desired approach would be to train the agent in a simulator and transfer it to the real world. Still, models trained in a simulator tend to perform poorly in real-world environments due to the differences. In this paper, we present a DRL-based algorithm that is capable of performing autonomous robot control using Deep Q-Networks (DQN). In our approach, the agent is trained in a simulated environment and it is able to navigate both in a simulated and real-world environment. The method is evaluated in the Duckietown environment, where the agent has to follow the lane based on a monocular camera input. The trained agent is able to run on limited hardware resources and its performance is comparable to state-of-the-art approaches.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑