Adaptive workspace biasing for sampling-based planners

Zucker, Matt; Kuffner, James; Bagnell, J. Andrew

doi:10.1109/robot.2008.4543787

Cited by 73 publications

(47 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In sampling-based planning, one approach biases sampling during a planning query based on previous experience [4]. However, this approach requires identifying features that capture the quality of a sample, which may be difficult to do in a given problem domain.…”

Section: Related Workmentioning

confidence: 99%

A robot path planning framework that learns from experience

Berenson

Abbeel

Goldberg

2012

2012 IEEE International Conference on Robotics and Automation

193

135

View full text Add to dashboard Cite

Abstract-We propose a framework, called Lightning, for planning paths in high-dimensional spaces that is able to learn from experience, with the aim of reducing computation time. This framework is intended for manipulation tasks that arise in applications ranging from domestic assistance to robot-assisted surgery. Our framework consists of two main modules, which run in parallel: a planning-from-scratch module, and a module that retrieves and repairs paths stored in a path library. After a path is generated for a new query, a library manager decides whether to store the path based on computation time and the generated path's similarity to the retrieved path. To retrieve an appropriate path from the library we use two heuristics that exploit two key aspects of the problem: (i) A correlation between the amount a path violates constraints and the amount of time needed to repair that path, and (ii) the implicit division of constraints into those that vary across environments in which the robot operates and those that do not.We evaluated an implementation of the framework on several tasks for the PR2 mobile manipulator and a minimally-invasive surgery robot in simulation. We found that the retrieve-andrepair module produced paths faster than planning-fromscratch in over 90% of test cases for the PR2 and in 58% of test cases for the minimally-invasive surgery robot.

show abstract

Section: Related Workmentioning

confidence: 99%

A robot path planning framework that learns from experience

Berenson

Abbeel

Goldberg

2012

2012 IEEE International Conference on Robotics and Automation

193

135

View full text Add to dashboard Cite

show abstract

“…Tuning planners (specifically, Rapidly-exploring Randomized Trees (RRTs)) using Policy Search Reinforcement Learning has previously been explored by [35] as a principled approach to navigating the high-dimensional space of parameters associated with complex heuristics. Here we emphasize the theoretical connection between this approach and Structured Prediction forms of IOC, as well as the broader scope of the framework as discussed in Section II-B.…”

Section: Related Workmentioning

confidence: 99%

Direct Loss Minimization Inverse Optimal Control

Doerr

Ratliff

Bohg

et al. 2015

Robotics: Science and Systems XI

View full text Add to dashboard Cite

Abstract-Inverse Optimal Control (IOC) has strongly impacted the systems engineering process, enabling automated planner tuning through straightforward and intuitive demonstration. The most successful and established applications, though, have been in lower dimensional problems such as navigation planning where exact optimal planning or control is feasible. In higher dimensional systems, such as humanoid robots, research has made substantial progress toward generalizing the ideas to model free or locally optimal settings, but these systems are complicated to the point where demonstration itself can be difficult. Typically, real-world applications are restricted to at best noisy or even partial or incomplete demonstrations that prove cumbersome in existing frameworks. This work derives a very flexible method of IOC based on a form of Structured Prediction known as Direct Loss Minimization. The resulting algorithm is essentially Policy Search on a reward function that rewards similarity to demonstrated behavior (using Covariance Matrix Adaptation (CMA) in our experiments). Our framework blurs the distinction between IOC, other forms of Imitation Learning, and Reinforcement Learning, enabling us to derive simple, versatile, and practical algorithms that blend imitation and reinforcement signals into a unified framework. Our experiments analyze various aspects of its performance and demonstrate its efficacy on conveying preferences for motion shaping and combined reach and grasp quality optimization.

show abstract

“…Among the most relevant probabilistic approaches are the Rapidly-exploring Random Trees planners (RRT) [6] and Probabistic Road Map planners (PRM) [13]. These original probabilistic approaches have some problems when there are narrow passages and in order to overcome them several variations were developed, like a multi-resolution PRM planner [14], a dynamics domain RRTs [15], a retraction base RRTs [16], a adaptive workspace biasing [17], and a sampling method based on Principal Component Analysis [18]. In order to speed up the query path planning, some variants of PRM planners build a roadmap without checking for collisions, then, once a potential solution path was found the existence of collisions is verified and if they occur the corresponding nodes and edges are removed from the roadmap and a new search is started; the process is repeated until a collision free path is found (e.g.…”

Section: Related Workmentioning

confidence: 99%

Planning manipulation movements of a dual-arm system considering obstacle removing

Rodríguez

Montano

Suárez

2014

Robotics and Autonomous Systems

View full text Add to dashboard Cite

The paper deals with the problem of planning movements of two hand-arm robotic systems, considering the possibility of using the robot hands to remove potential obstacles in order to obtain a free access to grasp a desired object. The approach is based on a variation of a Probabilistic Road Map that does not rule out the samples implying collisions with removable objects but instead classifies them according to the collided obstacle(s), and allows the search of free paths with the indication of which objects must be removed from the work-space to make the path actually valid, we call it Probabilistic Road Map with Obstacles (PRMwO). The proposed system includes a task assignment system that distributes the task among the robots, using for that purpose a precedence graph built from the results of the PRMwO. The approach has been implemented for a real dualarm robotic system, and some simulated and real running examples are presented in the paper.

show abstract

Adaptive workspace biasing for sampling-based planners

Cited by 73 publications

References 22 publications

A robot path planning framework that learns from experience

A robot path planning framework that learns from experience

Direct Loss Minimization Inverse Optimal Control

Planning manipulation movements of a dual-arm system considering obstacle removing

Contact Info

Product

Resources

About