2021/10/12 by Chao Yu, Xinyi Yang, Yu, Chao +9 · 2 citations
Computer Science · Engineering · #Artificial intelligence #Code (set theory) #Computer Vision and Pattern Recognition (cs.CV) #Computer science #Distributed computing #FOS: Computer and information sciences #Generalization #Inference #Machine Learning (cs.LG) #Machine learning #Multiagent Systems (cs.MA) #Overhead (engineering) #Planner #Reinforcement Learning in Robotics #Reinforcement learning #Representation (politics) #Robotic Path Planning Algorithms #Robotics and Sensor-Based Localization #Set (abstract data type) #cs.CV #cs.LG #cs.MA
paper · pdf · doi:10.48550/arxiv.2110.05734
published in arXiv (Cornell University) (Cornell University) · First three authors share equal contribution. This paper has been accepted by ECCV (https://link.springer.com/chapter/10.1007/978-3-031-19842-7_29)
openalex publication_date 2021/10/12 · arxiv created 2022/11/22 · arxiv updated 2022/11/23 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
We tackle the problem of cooperative visual exploration where multiple agents need to jointly explore unseen regions as fast as possible based on visual signals. Classical planning-based methods often suffer from expensive computation overhead at each step and a limited expressiveness of complex cooperation strategy. By contrast, reinforcement learning (RL) has recently become a popular paradigm for tackling this challenge due to its modeling capability of arbitrarily complex strategies and minimal inference overhead. In this paper, we extend the state-of-the-art single-agent visual navigation method, Active Neural SLAM (ANS), to the multi-agent setting by introducing a novel RL-based planning module, Multi-agent Spatial Planner (MSP).MSP leverages a transformer-based architecture, Spatial-TeamFormer, which effectively captures spatial relations and intra-agent interactions via hierarchical spatial self-attentions. In addition, we also implement a few multi-agent enhancements to process local information from each agent for an aligned spatial representation and more precise planning. Finally, we perform policy distillation to extract a meta policy to significantly improve the generalization capability of final policy. We call this overall solution, Multi-Agent Active Neural SLAM (MAANS). MAANS substantially outperforms classical planning-based baselines for the first time in a photo-realistic 3D simulator, Habitat. Code and videos can be found at https://sites.google.com/view/maans.