[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85463-en":3,"doc-seo-85463-105":28,"detail-sidebar-cat-0-en-105":90},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":11,"language":21,"language_code":22,"site_id":23,"html_lang":22,"table_of_contents":24,"faqs":25,"seo_title":13,"seo_description":14,"update_tm":26,"read_time":27},85463,549758146520,"Patrick","https://ap-avatar.wpscdn.com/avatar/80002397d8c0411e94?_k=1775819394049821470",8,"Research & Report","CoRL-MPPI: 用可学习行为增强 MPPI 的高效且可证明安全的多机器人避碰","Decentralized collision avoidance is a central challenge in scalable multi-robot systems, and MPPI is an attractive model predictive approach with strong theoretical guarantees and support for complex motion models. However, standard MPPI can yield suboptimal trajectories because performance depends on uninformed random sampling. CoRL-MPPI integrates cooperative reinforcement learning with MPPI by training a neural action policy in simulation and embedding it to bias sampling toward cooperative, collision-free actions. The method preserves regular MPPI guarantees and improves navigation efficiency and safety in dense dynamic scenarios.","CoRL-MPPI: Enhancing MPPI With Learnable Behaviours For Efficient And Provably-Safe Multi-Robot Collision Avoidance  \nStepan Dergachev, Artem Pshenitsyn, Aleksandr Panov, Alexey Skrynnik, Konstantin Yakovlev  \narXiv :2511 .09331v3 [ cs .RO] 11 Jul 2026  \nAbstract—Decentralized collision avoidance is a core challenge for scalable multi-robot systems. A promising approach to this problem is Model Predictive Path Integral (MPPI) control – a framework that naturally handles arbitrary motion models and provides strong theoretical guarantees. Still, in practice an MPPI-based controller may produce suboptimal trajectories because its performance relies heavily on uninformed random sampling. We introduce CoRL-MPPI, a fusion of Cooperative Reinforcement Learning and MPPI that addresses this limitation. We train an action policy, approximated by a deep neural network, in simulation to learn local cooperative collisionavoidance behaviors. This learned policy is then embedded into the MPPI framework to guide its sampling distribution, biasing it toward more intelligent and cooperative actions in scenarios that may differ substantially from those used during training. Moreover, CoRL-MPPI preserves the theoretical guarantees of regular MPPI. We evaluate our approach in dense, dynamic setups against classical and learning-based state-ofthe-art baselines. Our results demonstrate that CoRL-MPPI outperforms competing methods and significantly improves navigation efficiency, measured by success rate and delay, as well as safety, enabling agile and robust multi-robot navigation.  \nI. INTRODUCTION  \nThe deployment of multi-robot systems promises a significant boost in efficiency in warehouse logistics, search-andrescue, disaster management, etc. A fundamental problem in a multi-robot system is decentralized collision avoidance: each robot must navigate to its goal while proactively avoiding conflicts with the others. This problem is inherently challenging due to the non-linear nature of robot interactions, the curse of dimensionality as the number of agents grows, and the necessity for real-time computation under uncertainty.  \nThe known approaches to this problem are mostly reactive. Typical examples include Velocity Obstacles and Optimal Reciprocal Collision Avoidance (ORCA) [1], which compute collision-free velocities based on the current states of neighboring robots. While highly computationally efficient, reactive methods are inherently myopic. They operate on a one-step time horizon, which can lead to oscillatory behavior, deadlocks, and a general lack of cooperation.  \nConversely, methods based on receding-horizon optimal control, most notably the Model Predictive Control (MPC) framework, explicitly optimize a trajectory over multiple steps while accounting for predicted future states. The Model Predictive Path Integral (MPPI) [2], a sampling-based variant of MPC, has gained significant attention for its ability to handle non-linear dynamics and complex cost functions without the need for gradient computation. MPPI allows flexible formulation of both motion models and cost functions and is widely used in mobile robotics.  \nFig. 1: Left: the standard MPPI controller samples random rollouts (grey), which can lead to collisions and suboptimal control (red trajectory) . Right: the proposed fusion of RL and MPPI uses learned policy rollouts (blue) to bias sampling toward cooperative, collision-free behavior.  \nStill, the performance of MPPI-based methods is critically dependent on the quality of its sampled trajectories. In its standard formulation, control sequences are drawn from a Gaussian distribution centered around a prior (often the previous solution) . Such sampling may be very inefficient in complex multi-agent settings, where the vast majority of sampled trajectories may lead to uncooperative behavior. Consequently, even when the number of samples is high, the resultant trajectories may be overly egoistic, leading to an overall degrada","cbCaistsHxWJapjY","https://ap.wps.com/l/cbCaistsHxWJapjY","pdf",1499934,1,"English","en",105,"# Introduction\n## Decentralized collision avoidance challenges\n## Reactive versus predictive control approaches\n## MPPI sampling limitations\n## Proposed CoRL-MPPI contributions\n# Related Works","[{\"question\":\"为什么标准 MPPI 在多机器人避碰中可能表现不够理想？\",\"answer\":\"标准 MPPI 的采样质量高度依赖于以高斯分布为中心的随机采样，而在复杂多智能体场景中，大多数采样轨迹可能会导致不合作行为，从而使整体性能下降。\"},{\"question\":\"CoRL-MPPI 如何提升 MPPI 的采样效率与合作性？\",\"answer\":\"CoRL-MPPI 先通过强化学习在仿真中训练一个动作策略，用该策略生成的控制分布来引导 MPPI 的采样，从而偏向更聪明且更具合作性的动作与轨迹。\"},{\"question\":\"CoRL-MPPI 是否会破坏 MPPI 的理论安全保证？\",\"answer\":\"文中指出 CoRL-MPPI 保留正则 MPPI 的理论保证，并且在执行噪声条件下仍能保持安全性。\"}]",1784203745,20,{"code":4,"msg":29,"data":30},"ok",{"site_id":23,"language":22,"slug":31,"title":13,"keywords":32,"description":14,"schema_data":33,"social_meta":85,"head_meta":87,"extra_data":89,"updated_unix":26},"corl-mppi-enhancing-mppi-with-learnable-behaviours-for-efficient-and-provably-safe-multi-robot-collision-avoidance","",{"@graph":34,"@context":84},[35,52,67],{"@type":36,"itemListElement":37},"BreadcrumbList",[38,42,46,49],{"item":39,"name":40,"@type":41,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":43,"name":44,"@type":41,"position":45},"https://docshare.wps.com/document/","Document",2,{"item":47,"name":12,"@type":41,"position":48},"https://docshare.wps.com/document/research-report/",3,{"item":50,"name":13,"@type":41,"position":51},"https://docshare.wps.com/document/corl-mppi-enhancing-mppi-with-learnable-behaviours-for-efficient-and-provably-safe-multi-robot-collision-avoidance/85463/",4,{"url":50,"name":13,"@type":53,"author":54,"headline":13,"publisher":56,"fileFormat":59,"inLanguage":22,"description":14,"dateModified":60,"datePublished":61,"encodingFormat":59,"isAccessibleForFree":62,"interactionStatistic":63},"DigitalDocument",{"name":9,"@type":55},"Person",{"url":39,"name":57,"@type":58},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",true,{"@type":64,"interactionType":65,"userInteractionCount":20},"InteractionCounter",{"@type":66},"ViewAction",{"@type":68,"mainEntity":69},"FAQPage",[70,76,80],{"name":71,"@type":72,"acceptedAnswer":73},"为什么标准 MPPI 在多机器人避碰中可能表现不够理想？","Question",{"text":74,"@type":75},"标准 MPPI 的采样质量高度依赖于以高斯分布为中心的随机采样，而在复杂多智能体场景中，大多数采样轨迹可能会导致不合作行为，从而使整体性能下降。","Answer",{"name":77,"@type":72,"acceptedAnswer":78},"CoRL-MPPI 如何提升 MPPI 的采样效率与合作性？",{"text":79,"@type":75},"CoRL-MPPI 先通过强化学习在仿真中训练一个动作策略，用该策略生成的控制分布来引导 MPPI 的采样，从而偏向更聪明且更具合作性的动作与轨迹。",{"name":81,"@type":72,"acceptedAnswer":82},"CoRL-MPPI 是否会破坏 MPPI 的理论安全保证？",{"text":83,"@type":75},"文中指出 CoRL-MPPI 保留正则 MPPI 的理论保证，并且在执行噪声条件下仍能保持安全性。","https://schema.org",{"og:url":50,"og:type":86,"og:title":13,"og:site_name":57,"og:description":14},"article",{"robots":88,"canonical":50},"index,follow",{"doc_id":7,"site_id":23},{"code":4,"msg":5,"data":91},[92,96,100,104,109,114,119,122,126,129,133],{"id":20,"doc_module":4,"doc_module_name":44,"category_name":93,"show_sort_weight":94,"slug":95},"Story & Novel",90,"story-novel",{"id":45,"doc_module":4,"doc_module_name":44,"category_name":97,"show_sort_weight":98,"slug":99},"Literature",80,"literature",{"id":51,"doc_module":4,"doc_module_name":44,"category_name":101,"show_sort_weight":102,"slug":103},"Exam",70,"exam",{"id":105,"doc_module":4,"doc_module_name":44,"category_name":106,"show_sort_weight":107,"slug":108},5,"Comic",60,"comic",{"id":110,"doc_module":4,"doc_module_name":44,"category_name":111,"show_sort_weight":112,"slug":113},6,"Technology",50,"technology",{"id":115,"doc_module":4,"doc_module_name":44,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":44,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":44,"category_name":124,"show_sort_weight":27,"slug":125},9,"Religion & Spirituality","religion-spirituality",{"id":27,"doc_module":4,"doc_module_name":44,"category_name":127,"show_sort_weight":27,"slug":128},"World Cup","world-cup",{"id":130,"doc_module":4,"doc_module_name":44,"category_name":131,"show_sort_weight":130,"slug":132},10,"Lifestyle","lifestyle",{"id":134,"doc_module":4,"doc_module_name":44,"category_name":135,"show_sort_weight":105,"slug":136},19,"General","general"]