49,858 Hits in 4.5 sec

Learning Space Partitions for Path Planning [article]

Kevin Yang, Tianjun Zhang, Chris Cummins, Brandon Cui, Benoit Steiner, Linnan Wang, Joseph E. Gonzalez, Dan Klein, Yuandong Tian
2022 arXiv   pre-print
Recently, LaMCTS empirically learns to partition the search space in a reward-sensitive manner for black-box optimization.  ...  We also propose a new path planning method LaP3 which improves the function value estimation within each sub-region, and uses a latent representation of the search space.  ...  Acknowledgments and Disclosure of Funding We thank the members of the Berkeley NLP group as well as our four anonymous reviewers for their helpful feedback.  ... 
arXiv:2106.10544v4 fatcat:w5pqekogfzc6vniwnnrsqjq2yu

Partition Learning for Multiagent Planning

Jared Wood, J. Karl Hedrick
2012 Journal of Robotics  
The first step is target track estimation and the second step is path planning by optimizing directly over target track estimation. This standard approach works well in many scenarios.  ...  However, an improved approach is needed for the scenario when general, nonparametric estimation is required, and the number of targets is unknown.  ...  However, the average partition size decreases substantially for partition learning classification path planning.  ... 
doi:10.1155/2012/590479 fatcat:pshmruznenffvh2ozvvkwhftse

A Machine Learning Approach for Feature-Sensitive Motion Planning [chapter]

Marco Morales, Lydia Tapia, Roger Pearce, Samuel Rodriguez, Nancy M. Amato
2005 Springer Tracts in Advanced Robotics  
We use a machine learning approach to characterize and partition C-space into (possibly overlapping) regions that are well suited to one of the planners in our library of roadmap-based motion planning  ...  In this paper, we propose an automated framework for feature-sensitive motion planning.  ...  A prototype implementation of our machine learning classification and partitioning approach was shown to achieve results su- In the future, we plan to extend our framework to handle more C-space classes  ... 
doi:10.1007/10991541_25 fatcat:levnv4dy35cjtdqnfcqo2yfcr4

The parti-game algorithm for variable resolution reinforcement learning in multidimensional state-spaces

Andrew W. Moore, Christopher G. Atkeson
1995 Machine Learning  
Parti-game is a new algorithm for learning feasible trajectories to goal regions in high dimensional continuous state-spaces.  ...  It develops a variable resolution partitioning in conjunction with the planning and learning of Algorithm 3.  ...  In practice, parti-game's solution was shorter than the optimal path in the discretized maze.  ... 
doi:10.1007/bf00993591 fatcat:auxbn7yalvcmtpglg52j3bkwnq

Asymptotically Optimal Motion Planning for Learned Tasks Using Time-Dependent Cost Maps

Chris Bowen, Gu Ye, Ron Alterovitz
2015 IEEE Transactions on Automation Science and Engineering  
Index Terms Motion and path planning; assistive robots Contact  ...  We next present an efficient algorithm for planning robot motions to perform the task based on the learned features while avoiding obstacles.  ...  Acknowledgments Alterovitz has co-authored a book on Motion Planning in Medicine, was awarded a patent for a medical device, has received multiple best paper finalist awards at IEEE robotics conferences  ... 
doi:10.1109/tase.2014.2342718 pmid:26279642 pmcid:PMC4535732 fatcat:a3xaxcvvpbgxzbgbstm36uqp5e

Towards Learning Path Planning for Solving Complex Robot Tasks [chapter]

Thomas Frontzek, Thomas Navin Lal, Rolf Eckmiller
2001 Lecture Notes in Computer Science  
For solving complex robot tasks it is necessary to incorporate path planning methods that are able to operate within di erent high-dimensional con guration spaces containing an unknown number of obstacles  ...  For an examplary planning task we show that Adaptive AA* learns movement v ectors which a l l o w larger movements than the initial ones into well-de ned directions of the con guration space.  ...  Donath, Nils Goerke, and Stefan Ho richter for useful comments.  ... 
doi:10.1007/3-540-44668-0_130 fatcat:rxpk56edxvdu5lnmbhhtmqidf4

Learning sensor-based navigation of a real mobile robot in unknown worlds

R. Araujo, A.T. de Almeida
1999 IEEE Transactions on Systems Man and Cybernetics Part B (Cybernetics)  
The parti-game multiresolution learning approach [22] is applied for simultaneous and cooperative construction of a world model, and learning to navigate through an obstacle-free path from a starting position  ...  Index Terms-Learning systems, mobile robot navigation.  ...  Two of the most important tasks for autonomous navigation of a mobile robot are path planning, and building a world model.  ... 
doi:10.1109/3477.752791 pmid:18252290 fatcat:4fzw6aphzrcufjeisjjjds223u

Multi-robot path planning for budgeted active perception with self-organising maps

Graeme Best, Jan Faigl, Robert Fitch
2016 2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)  
We propose a self-organising map (SOM) algorithm as a solution to a new multi-goal path planning problem for active perception and data collection tasks.  ...  We optimise paths for a multi-robot team that aims to maximally observe a set of nodes in the environment.  ...  This initialisation is reasonable since the paths quickly spread over the input space during the first learning epochs. B. Viewpoint Rewards Each node has an associated reward for being visited.  ... 
doi:10.1109/iros.2016.7759489 dblp:conf/iros/BestFF16 fatcat:mpfhbq6jhrfx7bjt6565vm6q74

Hierarchical Learning in Stochastic Domains: Preliminary Results [chapter]

Leslie Pack Kaelbling
1993 Machine Learning Proceedings 1993  
This paper presents the HDG learning algorithm, which uses a hierarchical decomposition of the state space to make learning to achieve goals more efficient with a small penalty in path quality.  ...  Special care must be taken when performing hierarchical planning and learning in stochastic domains, because macro-operators cannot be executed ballistically.  ...  LANDMARK NETWORKS When we plan a route for driving a long distance, we don't even attempt to find the optimal path; rather, we take advantage of an existing path hierarchy.  ... 
doi:10.1016/b978-1-55860-307-3.50028-9 dblp:conf/icml/Kaelbling93 fatcat:jhxt6fv4tzfrbkwsbkwvhcvmqi

Taking Learning Out of Real-Time Heuristic Search for Video-Game Pathfinding [chapter]

Ramon Lawrence, Vadim Bulitko
2010 Lecture Notes in Computer Science  
The new approach has a fast move time and eliminates learning and "scrubbing".  ...  However, a common issue is that the pre-computation time can be large, and there is no guarantee that the precomputed data adequately covers the search space.  ...  The edge costs are 1 for cardinal moves and 1.4 for diagonal moves. An agent plans its next action by considering states in a local search space surrounding its current position.  ... 
doi:10.1007/978-3-642-17432-2_41 fatcat:sdydtuk7yngedmhefa552lxmva

Automatic Curriculum Learning For Deep RL: A Short Survey

Rémy Portelas, Cédric Colas, Lilian Weng, Katja Hofmann, Pierre-Yves Oudeyer
2020 Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence  
Automatic Curriculum Learning (ACL) has become a cornerstone of recent successes in Deep Reinforcement Learning (DRL).  ...  They can optimize domain randomization for Sim2Real transfer, organize task presentations in multi-task robotic settings, order sequences of opponents in multi-agent scenarios, etc.  ...  Initial discussions started at the Dagstuhl seminar 18071 on Planning and Operations Research.  ... 
doi:10.24963/ijcai.2020/663 dblp:conf/ijcai/PommereningRHCR20 fatcat:uazsfxe2lne4bljtm42moao6jq

A Non-uniform Sampling Approach for Fast and Efficient Path Planning [article]

James P. Wilson and Zongyuan Shen and Shalabh Gupta
2021 arXiv   pre-print
In this paper, we develop a non-uniform sampling approach for fast and efficient path planning of autonomous vehicles.  ...  This sampling framework is incorporated into the RRT* path planner.  ...  Section II formulates the path planning problem for AUVs. Section III presents the smart sampling procedure and the path planning algorithm.  ... 
arXiv:2108.01291v1 fatcat:vfwow7eeffblzoplhqimng7q3u

Learning Abstract and Transferable Representations for Planning [article]

Steven James, Benjamin Rosman, George Konidaris
2022 arXiv   pre-print
We restrict our focus to learning representations for long-term planning, a class of problems that state-of-the-art learning methods are unable to solve.  ...  We propose a framework for autonomously learning state abstractions of an agent's environment, given a set of skills.  ...  We intend to learn an abstract representation suitable for planning.  ... 
arXiv:2205.02092v1 fatcat:cbu5i3j7gbdbpiidircx5zdsu4

Learning Search Space Partition for Black-box Optimization using Monte Carlo Tree Search [article]

Linnan Wang, Rodrigo Fonseca, Yuandong Tian
2022 arXiv   pre-print
Unlike previous approaches, LA-MCTS learns the partition of the search space using a few samples and their function values in an online fashion.  ...  ., BO, TuRBO) as its local models, achieving strong performance in general black-box optimization and reinforcement learning benchmarks, in particular for high-dimensional problems.  ...  Following the path of learning, one under-explored direction is to learn the space partition.  ... 
arXiv:2007.00708v2 fatcat:l7wgsqtzxff6ngd4jhomov7tve

PPVF: A Novel Framework for Supporting Path Planning Over Carpooling

Bin Wang, Rui Zhu, Siting Zhang, Zheng Zhao, Xiaochun Yang, Guoren Wang
2019 IEEE Access  
In this paper, we propose an efficient framework, named PPVF(short for path prediction and verification-based framework) for path planning over the road network.  ...  Furthermore, we propose a searching and verification-based algorithm for further improving the path planning quality.  ...  For the path partition, we partition the elements in KP into a few sub-paths, use a group of MBRs with relatively small area to bound these paths.  ... 
doi:10.1109/access.2019.2891570 fatcat:j7ydj34bn5birooihutp3qhjly
« Previous Showing results 1 — 15 out of 49,858 results