اشترك بالحزمة الذهبية واحصل على وصول غير محدود شمرا أكاديميا
تسجيل مستخدم جديدPRIMAL2: Pathfinding via Reinforcement and Imitation Multi-Agent Learning -- Lifelong
471
0
0.0
(
0
)
نشر من قبل
Guillaume Sartoretti
تاريخ النشر
2020
مجال البحث
الهندسة المعلوماتية
والبحث باللغة
English
اسأل ChatGPT حول البحث
ﻻ يوجد ملخص باللغة العربية
Multi-agent path finding (MAPF) is an indispensable component of large-scale robot deployments in numerous domains ranging from airport management to warehouse automation. In particular, this work addresses lifelong MAPF (LMAPF) - an online variant of the problem where agents are immediately assigned a new goal upon reaching their current one - in dense and highly structured environments, typical of real-world warehouse operations. Effectively solving LMAPF in such environments requires expensive coordination between agents as well as frequent replanning abilities, a daunting task for existing coupled and decoupled approaches alike. With the purpose of achieving considerable agent coordination without any compromise on reactivity and scalability, we introduce PRIMAL2, a distributed reinforcement learning framework for LMAPF where agents learn fully decentralized policies to reactively plan paths online in a partially observable world. We extend our previous work, which was effective in low-density sparsely occupied worlds, to highly structured and constrained worlds by identifying behaviors and conventions which improve implicit agent coordination, and enable their learning through the construction of a novel local agent observation and various training aids. We present extensive results of PRIMAL2 in both MAPF and LMAPF environments and compare its performance to state-of-the-art planners in terms of makespan and throughput. We show that PRIMAL2 significantly surpasses our previous work and performs comparably to these baselines, while allowing real-time re-planning and scaling up to 2048 agents.
قيم البحث
اقرأ أيضاً
Multi-agent path finding (MAPF) is an essential component of many large-scale, real-world robot deployments, from aerial swarms to warehouse automation. However, despite the communitys continued efforts, most state-of-the-art MAPF planners still rely
on centralized planning and scale poorly past a few hundred agents. Such planning approaches are maladapted to real-world deployments, where noise and uncertainty often require paths be recomputed online, which is impossible when planning times are in seconds to minutes. We present PRIMAL, a novel framework for MAPF that combines reinforcement and imitation learning to teach fully-decentralized policies, where agents reactively plan paths online in a partially-observable world while exhibiting implicit coordination. This framework extends our previous work on distributed learning of collaborative policies by introducing demonstrations of an expert MAPF planner during training, as well as careful reward shaping and environment sampling. Once learned, the resulting policy can be copied onto any number of agents and naturally scales to different team sizes and world dimensions. We present results on randomized worlds with up to 1024 agents and compare success rates against state-of-the-art MAPF planners. Finally, we experimentally validate the learned policies in a hybrid simulation of a factory mockup, involving both real-world and simulated robots.
Recent work on the multi-agent pathfinding problem (MAPF) has begun to study agents with motion that is more complex, for example, with non-unit action durations and kinematic constraints. An important aspect of MAPF is collision detection. Many coll
ision detection approaches exist, but often suffer from issues such as high computational cost or causing false negative or false positive detections. In practice, these issues can result in problems that range from inefficiency and annoyance to catastrophic. The main contribution of this technical report is to provide a high-level overview of major categories of collision detection, along with methods of collision detection and anticipatory collision avoidance for agents that are both computationally efficient and highly accurate.
The Persistent Monitoring (PM) problem seeks to find a set of trajectories (or controllers) for robots to persistently monitor a changing environment. Each robot has a limited field-of-view and may need to coordinate with others to ensure no point in
the environment is left unmonitored for long periods of time. We model the problem such that there is a penalty that accrues every time step if a point is left unmonitored. However, the dynamics of the penalty are unknown to us. We present a Multi-Agent Reinforcement Learning (MARL) algorithm for the persistent monitoring problem. Specifically, we present a Multi-Agent Graph Attention Proximal Policy Optimization (MA-G-PPO) algorithm that takes as input the local observations of all agents combined with a low resolution global map to learn a policy for each agent. The graph attention allows agents to share their information with others leading to an effective joint policy. Our main focus is to understand how effective MARL is for the PM problem. We investigate five research questions with this broader goal. We find that MA-G-PPO is able to learn a better policy than the non-RL baseline in most cases, the effectiveness depends on agents sharing information with each other, and the policy learnt shows emergent behavior for the agents.
Multi-agent collision-free trajectory planning and control subject to different goal requirements and system dynamics has been extensively studied, and is gaining recent attention in the realm of machine and reinforcement learning. However, in partic
ular when using a large number of agents, constructing a least-restrictive collision avoidance policy is of utmost importance for both classical and learning-based methods. In this paper, we propose a Least-Restrictive Collision Avoidance Module (LR-CAM) that evaluates the safety of multi-agent systems and takes over control only when needed to prevent collisions. The LR-CAM is a single policy that can be wrapped around policies of all agents in a multi-agent system. It allows each agent to pursue any objective as long as it is safe to do so. The benefit of the proposed least-restrictive policy is to only interrupt and overrule the default controller in case of an upcoming inevitable danger. We use a Long Short-Term Memory (LSTM) based Variational Auto-Encoder (VAE) to enable the LR-CAM to account for a varying number of agents in the environment. Moreover, we propose an off-policy meta-reinforcement learning framework with a novel reward function based on a Hamilton-Jacobi value function to train the LR-CAM. The proposed method is fully meta-trained through a ROS based simulation and tested on real multi-agent system. Our results show that LR-CAM outperforms the classical least-restrictive baseline by 30 percent. In addition, we show that even if a subset of agents in a multi-agent system use LR-CAM, the success rate of all agents will increase significantly.
Social learning is a key component of human and animal intelligence. By taking cues from the behavior of experts in their environment, social learners can acquire sophisticated behavior and rapidly adapt to new circumstances. This paper investigates
whether independent reinforcement learning (RL) agents in a multi-agent environment can learn to use social learning to improve their performance. We find that in most circumstances, vanilla model-free RL agents do not use social learning. We analyze the reasons for this deficiency, and show that by imposing constraints on the training environment and introducing a model-based auxiliary loss we are able to obtain generalized social learning policies which enable agents to: i) discover complex skills that are not learned from single-agent training, and ii) adapt online to novel environments by taking cues from experts present in the new environment. In contrast, agents trained with model-free RL or imitation learning generalize poorly and do not succeed in the transfer tasks. By mixing multi-agent and solo training, we can obtain agents that use social learning to gain skills that they can deploy when alone, even out-performing agents trained alone from the start.
سجل دخول لتتمكن من نشر تعليقات
التعليقات
جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
