Robot navigation with deep reinforcement learning (RL) achieves higher performance and performs well under complex environment. Meanwhile, the interpretation of the decision-making of deep RL models becomes a critical problem for more safety and reliability of autonomous robots. In this paper, we propose a visual explanation method based on an attention branch for deep RL models. We connect attention branch with pre-trained deep RL model and the attention branch is trained by using the selected action by the trained deep RL model as a correct label in a supervised learning manner. Because the attention branch is trained to output the same result as the deep RL model, the obtained attention maps are corresponding to the agent action with higher interpretability. Experimental results with robot navigation task show that the proposed method can generate interpretable attention maps for a visual explanation.
"Why This Avoidance Maneuver?" Contrastive Explanations in Human-Supervised Maritime Autonomous Navigation
Department of ICT and Natural Sciences, Norwegian University of Science and Technology (NTNU), Ålesund, Norway(挪威科技大学(NTNU)奥勒松校区信息通信技术与自然科学系);Department of Ocean Operations and Civil Engineering, Norwegian University of Science and Technology (NTNU), Ålesund, Norway(挪威科技大学(NTNU)奥勒松校区海洋运营与土木工程系);Department of Science and Mathematics, Volda University College (HVO), Volda, Norway(沃尔达大学学院(HVO)科学与数学系);Department of Engineering Cybernetics, Norwegian University of Science and Technology (NTNU), Trondheim, Norway(挪威科技大学(NTNU)特隆赫姆校区工程控制论系)
Immersive Explainability: Visualizing Robot Navigation Decisions through XAI Semantic Scene Projections in Virtual Reality
University of Bonn(波恩大学);Lamarr Institute for Machine Learning and Artificial Intelligence(拉马尔人工智能与机器学习研究所);Center for Robotics(机器人中心);University of Mainz(美因茨大学);Fraunhofer Institute for Intelligent Analysis and Information Systems IAIS(弗劳恩霍夫智能分析与信息系统研究所)
Deep Learning-based Robust Autonomous Navigation of Aerial Robots in Dense Forests
Department of Remote Sensing and Photogrammetry, Finnish Geospatial Research Institute FGI, The National Land Survey of Finland(遥感与摄影测量系,芬兰地理空间研究 institute FGI,芬兰国家土地测绘局)
A 2D Image Feedback-Based Autonomous Endovascular Navigation Method: A Perception and Control Framework
Beijing Institute of Technology(北京理工大学);Advanced Institute for Ocean Research, Southern University of Science and Technology(南方科技大学海洋科学与工程研究院);Shenzhen Institute of Advanced Biomedical Robot Company Limited(深圳先进生物医学机器人有限公司);Southern University of Science and Technology(南方科技大学);School of Mechanical, Electronic and Control Engineering, Beijing Jiaotong University(北京交通大学机械与电子控制工程学院);Aerospace Center Hospital(航天中心医院)
CLASH: Collaborative Large-Small Hierarchical Framework for Continuous Vision-and-Language Navigation
Tongji University(同济大学);Key Laboratory of Smart Manufacturing in Energy Chemical Process, Ministry of Education, East China University of Science and Technology(能源化工过程智能制造重点实验室,教育部,东华大学)
Evaluating Explanation Methods for Vision-and-Language Navigation
The University of Hong Kong(香港大学);Centre for Transformative Garment Production (TransGP)(转型服装生产中心(TransGP));Southern University of Science and Technology(南方科技大学)
Towards Probabilistic Clearance, Explanation and Optimization
TU Darmstadt(达姆施塔特工业大学);Honda Research Institute Europe GmbH(本田欧洲研究院);TU Eindhoven(埃因霍温理工大学);Hessian AI(黑森人工智能机构);Centre for Cognitive Science(认知科学中心);German Center for Artificial Intelligence (DFKI)(德国人工智能研究中心)
Enhancing Robot Explanation Capabilities through Vision-Language Models: a Preliminary Study by Interpreting Visual Inputs for Improved Human-Robot Interaction