R. Bohlin and L. Kavraki. Path planning using lazy prm. In IEEE Int. Conf. Robot. & Autom., 2000.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....paths. The result of this process is a graph (the roadmap) that encodes the topology of CS free . Planning queries are then solved by connecting the start and goal con gurations to the roadmap and searching a path in it. The nal step is the smoothing of this resulting path. Several variants [1, 2, 3, 19, 21] have been proposed to improve the performance of the basic PRM algorithm. For all of them, the generation of random con gurations is a trivial process in the case of open kinematic chains. On the contrary, the diculty to randomly generate con gurations of a mechanism containing closed chains ....

R. Bohlin and L. Kavraki. Path Planning using Lazy PRM. In IEEE International Conference on Robotics and Automation, pp. 521-538, 2000.

....appearing in other sampling techniques could be safely replaced and possibly improved with deterministic sampling; however, case by case comparisons would be necessary, and are beyond the scope of this paper. One PRM variation that considers an idea independent of sampling is the Lazy PRM [6]. In this case, the idea is to rst construct a roadmap that ignores collision constraints, and then perform collision checking to validate edges only while searching for a solution. This results in dramatic reduction in preprocessing time. In the worst case, all edges will be checked, which is ....

....a signi cant factor. IV. Lazy Lattice Roadmap (Lazy LRM) A. The Basic Lazy PRM A recent PRM variant called the Lazy PRM has been proposed for the problem of answering single planning queries eciently, as opposed to building an extensive roadmap prior to consideration of a planning query [6]. The resulting planner is sometimes very ecient in comparison to the original PRM. This represents a shift from the multiple query philosophy of the original PRM [26] and returns to the single query philosophy which was used in earlier planners [4] 15] 33] The key idea in the Lazy PRM is ....

[Article contains additional citation context not shown here]

R. Bohlin and L. Kavraki. Path planning using lazy PRM. In IEEE Int. Conf. Robot. & Autom., 2000.

....problem encountered by the classic methods. Tests of the approach on two important biological problems show that it produces more accurate results and achieves several orders of magnitude reduction in computation time, compared with MC simulation. 1 Introduction Probabilistic roadmap (PRM) [1,8,9,13,16,17,25] planners have been successfully used in recent years to compute collision free paths for robots with many degrees of freedom. A classic PRM planner [16] samples at random a robot s configuration space to construct a network that approximates the connectivity of the free space, and then searches ....

R. Bohlin and L.E. Kavraki. Path planning using lazy PRM. In Proc. IEEE Int. Conf. on Robotics & Automation, pages 521--528, 2000.

....themselves do not collide. Hence, dynamic collision checking remains a major bottleneck in many applications. In particular, probabilistic roadmap (PRM) planners heavily rely on the availability of ecient dynamic checkers to test local paths between randomly sampled con gurations for collision [1, 2,4,11,13,22]. Most such planners use a static BV method to test intermediate con gurations along the local paths at some resolution #. Choosing # involves several trial and error experiments, whichmust be repeated for each new type of robot and environment. Large values of # are acceptable when both robots ....

R. Bohlin and L. Kavraki. Path planning using lazy PRM. In ##### ## ### #### ##### ## ############# ######, pages 521-528, 2000.

....the geometry of the free subset of a configuration space with dimension greater than four or five turns out to have a prohibitively high cost. Random sampling more specifically, PRM methods was introduced to solve (geometric) path planning problems for robots with many dofs [ABD 98, BK00, BKL 97, BL91, BOvdS99, HLM99, HST94, Hsu00, Kav94, K SLO96, Kuf99, LH00, SLL01, Sve97] The costly computation of an explicit representation of the free space is replaced by a collision test on every randomly picked sample and connection between samples. This, of course, can be done with ....

....of precomputing a roadmap can be amortized over a large number of queries. Single query ones are appropriate when the number of queries in a given space is small. Intermediate strategies, which precompute partial roadmaps and complete them to pro4 cess specific queries, have also been proposed [BK00, SMA01] The planner proposed in this paper follows the single query sampling paradigm. Single query strategies often build a new roadmap for each query by growing trees of sampled milestones rooted at the initial and or goal configurations [AG99, HLM97, Hsu00, Kuf99, LK99] but they differ in ....

[Article contains additional citation context not shown here]

R. Bohlin and L.E. Kavraki. Path planning using lazy PRM. In Proc. IEEE Int. Conf. on Robotics and Automation, 2000.

....to the rest of the roadmap. Other strategies generate a greater density of milestones near the boundary of the free space, as the connectivity of narrow regions is more difficult to capture than that of wide open regions [1,4] Delay collisio checks util they are absolutely eeded. The planner in [3] first generates a network by distributing points at random in configuration space. It initially assumes that all points and connections between them are collision free. It then computes the shortest path in this network between two query configurations and tests it for collision. If a collision ....

....and tests it for collision. If a collision is detected, the node and or segment where it occurs are erased, and a new shortest path is computed and tested; and so on. We think that delaying collision tests is a promising approach, but its potential has only been partially exploited in [3]. One must decide in advance how large the network should be. If it is too coarse, it may fail to contain a solution path. But, if it is too dense, time will be wasted checking similar paths for collision. The focus on shortest paths may be costly when obstacles force the robot to take long ....

[Article contains additional citation context not shown here]

Bohlin R., Kavraki L.E. (2000) Path Planning Using Lazy PRM. Proc. IEEE Int. Conf. Robotics & Autore., San Francisco, CA.

....workspaces. Preliminary experiments show that the hybrid sampling strategy enables relatively small roadmaps to reliably capture the connectivity of configuration spaces with difficult narrow passages. 1 Introduction During the past decade, probabilistic roadmap (PRM) planners [ABD 98, BK00, BOvdS99, HLM99, K SLO96, NSL99, LK01] have emerged as a powerful framework for path planning of robots with many degrees of freedom (dofs) A classic PRM planner [K SLO96] samples at random a robot s configuration space to construct a network, called a roadmap, that approximates the connectivity ....

R. Bohlin and L.E. Kavraki. Path planning using lazy PRM. In Proc. IEEE Int. Conf. on Robotics & Automation, pages 521--528, 2000.

....them sensitive to the particular subdivision of a scene into individual obstacles. Recently some new global techniques have been described. Nissoux et al. 16] describe a technique called visibility based PRM that tries to keep the roadmap small by only adding useful nodes. Bohlin and Kavraki [5] describe a lazy evaluation technique that only checks those parts of the roadmap that are relevant for a query. Both of these techniques are complimentary to our approach and can be used in combination with it. Because the problems are caused by uniform sampling, the obvious idea is to try to ....

R. Bohlin, L.E. Kavraki, Path planning using lazy PRM, Proc. IEEE Int. Conf. Robotics and Automation, 2000, pp. 521-528.

....It can be shown that the approach converges to a roadmap that covers the entire free space. The number of nodes tends to remain very small, unless the free space has a very complicated structure. Another idea is not to test whether the paths are collision free unless they are really needed[6]. Such a lazy approach only checks whether the nodes are collision free and when nodes are close to each other they are connected with an edge. Only when an actual motion planning query must be solved we test whether the edges on the shortest path in the graph are collision free. If not we try ....

R. Bohlin, L.E. Kavraki, Path planning using lazy PRM, Proc. IEEE Int. Conf. on Robotics and Automation, 2000, pp. 521-528.

.... generate a greater density of milestones near the boundary of the free space, as the connectivity of narrow regions is more difficult to capture than that of wide open regions [11, 12] # Postpone collision tests until they are absolutely needed (lazy collision checking) The planner in [13] distributes points uniformly at random in configuration space. It initially assumes that all points and connections between them are collision free. It computes the shortest path in this network between two query configurations and tests it for collision. If a collision is detected, the node ....

....and tests it for collision. If a collision is detected, the node and or segment where it occurs are erased, and a new shortest path is computed and tested; and so on. We think that lazy collision checking is a promising approach, but that its potential has only been partially exploited in [13]. The network built is reminiscent of a roadmap pre computed by a multi query planner [2, 3] One must decide in advance how large it should be. If it is too coarse, it may fail to contain a solution path. But, if it is too dense, time will be wasted checking similar paths for collision. The focus ....

[Article contains additional citation context not shown here]

R. Bohlin and L. E. Kavraki. Path planning using lazy PRM. In Proc. IEEE Int. Conf. Rob. & Autom., pages 521--528, 2000.

....la simple d etection de collision) pour calculer dans un premier temps un graphe de configurations autorisant de faibles p en etrations avec les obstacles; le graphe est ensuite modifi e en repoussant progressivement les configurations en collision dans l espace libre. D autres variantes (e.g. [7]) visent a limiter le nombre d appels au d etecteur de collision en introduisant des m ecanismes d evaluation paresseuse lors de la v erification de la validit e des chemins calcul es par la m ethode locale. La compl etude probabiliste , etablie dans un premier temps [15] pour des syst emes ....

R. Bohlin and L. Kavraki. Path planning using lazy PRM. In IEEE International Conference on Robotics and Automation, San Francisco (USA), 2000.

.... but only in the regions determined promising by the initial coarse processing, which can save a signi cant amount of processing time. A similar philosophy 2 (a) b) Figure 1: a) control roadmap, and (b) the resulting approximate roadmap for an environment. is proposed in the Lazy PRM [6] and Fuzzy PRM [16] methods, which advocate an even coarser validation than c prm either no validation at all [6] i.e. accepting all roadmap nodes and edges in the construction phase) or validating nodes but not edges [16] Our experience is that some, even very limited, validation of both ....

....amount of processing time. A similar philosophy 2 (a) b) Figure 1: a) control roadmap, and (b) the resulting approximate roadmap for an environment. is proposed in the Lazy PRM [6] and Fuzzy PRM [16] methods, which advocate an even coarser validation than c prm either no validation at all [6] (i.e. accepting all roadmap nodes and edges in the construction phase) or validating nodes but not edges [16] Our experience is that some, even very limited, validation of both nodes and edges is essential to provide an estimate of the free space to guide path selection, which in turn can signi ....

R. Bohlin and L. E. Kavraki. Path planning using lazy prm. In Proc. IEEE Int. Conf. Robot. Autom. (ICRA), pages 521-528, 2000.

....further improved using smoothing techniques. PRM s have been applied for many different types of motion planning (free flying, carlike, robot arms, flexible, etc. and many techniques have been suggested to improve the performance (e.g. visibility roadmaps, Gaussian sampling, and lazy PRM s) See [1, 2, 3, 4, 8, 7, 10, 13] for some of the many important results obtained recently. Besides hard constraints on the resulting paths (no intersection with the obstacles and feasible for the moving body) in many applications there are also soft constraints, e.g. to preferably stay away from certain obstacles and to avoid ....

R. Bohlin, L.E. Kavraki, Path Planning Using Lazy PRM, in: Proceedings of the IEEE International Conference on Robotics and Automation, 2000, pp. 521-528.

....narrow passages as long as those passages are created by two distinct obstacles. Triples are generated rather than pairs: if two nodes are in collision with two distinct obstacles and the third is in C free, all no more than d away from each other, the node is added to the graph. 2. Lazy PRM In [15], Lazy PRM is the usual PRM built with all collision checking postponed until a query is presented. At that time a path through the roadmap is found and only then is collision checking performed from the outside nodes (start and goal) towards the middle at increasingly ner resolution. After ....

....sequential one. By iteratively incrementing the sampling resolution when validating an edge, we increase the chance, on average, of detecting collisions sooner. There is a slight overhead associated with calculating the recursive midpoints to check. As noted previously (Chapter III D 2) Lazy PRM [15] uses an increasing resolution strategy on entire paths (multiple edges) at a time. We are aware of no further documentation of this heuristic although it is likely that other researchers with mature development platforms have also considered it. 4. Line Segment Approximation No matter how the ....

R. Bohlin and L. E. Kavraki, \Path planning using lazy PRM," in Proc. IEEE Int. Conf. Robot. Autom. (ICRA), pp. 521-528, 2000.

.... For example, problems where the solution path must traverse narrow passages in C space [12] As a result, a number of prm variants speci cally targeted at this problem have been proposed, e.g. 1, 6, 12] Also, some novel approaches for improving prm eciency have shown promising results [5, 20] (these methods do not address the narrow passage problem) Unfortunately, however, prm solutions to many problems still take prohibitively long. Another shortcoming of prms is that while they are very good at nding a path, they do not support applications which might impose particular, ....

....O(nkd) validity checks the majority of which are not needed for any particular solution path. Our strategy of postponing validation leads to faster roadmap construction, while incurring only slightly higher query times. As discussed below, similar strategies have been used by other researchers [5, 20]. 1 An important bene t of our approach is that it gives one the ability to customize the roadmap in accordance with any speci ed, variable, query requirements. For example, it might be speci ed that the solution path should maintain a certain clearance from the obstacles, that the robot s ....

[Article contains additional citation context not shown here]

R. Bohlin and L. E. Kavraki. Path planning using lazy prm. In Proc. IEEE Int. Conf. Robot. Autom. (ICRA), pages 521-528, 2000.

....questions by illustrating some of the advantages of quasi random sampling in the context of the probabilistic roadmap (PRM) framework for path planning [1, 16] a glimpse is given in Figure 1. We present implemented, quasi random variants of both the classical PRM [16] and the recent Lazy PRM [4], and indicate some advantages over their randomized counterparts. At rst glance, the progression from deterministic to randomized, and then back to deterministic might appear absurd; thus, some explanation is required. There appear to be two prevailing reasons for the preference of randomized ....

....of the number of nodes required to generate successful plans. 4 Quasi random Lazy PRM A recent PRM variant called the Lazy PRM has been proposed for the problem of answering single planning queries eciently, as opposed to building an extensive roadmap prior to consideration of a planning query [4]. The resulting planner is sometimes very ecient in comparison to the original PRM. The primary novelty of the Lazy PRM is that the roadmap is initially constructed without the use of a 3 2 1 0 1 2 3 3 2 1 0 1 2 3 Workspace of a 2 Link Planar Manipulator Figure 3: A two link ....

[Article contains additional citation context not shown here]

R. Bohlin and L. Kavraki. Path planning using lazy PRM. In IEEE Int. Conf. Robot. & Autom., 2000.

.... based on rapidly exploring random trees (RRTs) in the configuration space[7] This algorithm was selected for its speed in solving single query path planning problems, particularly in character animation[8] However, other successful single query path planning techniques could potentially be used[2, 5]. 3. Experiments Figure 2 shows some example manipulation tasks solved by the planner. All motions were computed in 1 to 3 seconds on average on a 270 MHz SGI O2 running Irix 6.5 (see table below) The human arm is modeled as a 7 DOF kinematic chain, and each scene contains over 10,000 triangle ....

R. Bohlin and L. Kavraki. Path planning using lazy PRM. In In Proc. IEEE Int'l Conf. on Robotics and Automation (ICRA'2000), Apr. 2000.

No context found.

R. Bohlin and L. Kavraki. Path planning using lazy prm. In IEEE Int. Conf. Robot. & Autom., 2000.

....have resulted in the successful application of these techniques to diverse domains, such as assembly planning, virtual prototyping, drug design, and computer animation. Much of the progress can be attributed to the introduction of probabilistic roadmap techniques [10] and their various extensions [1, 2, 8, 9]. Despite these advances, however, some areas of application have still remained out of reach for automated planning algorithms. Applications requiring robots with many degrees of freedom to operate in highly dynamic and unpredictably changing environments fall into that category. To operate ....

R. Bohlin and L. E. Kavraki. Path planning using lazy PRM. In Proc. Intl. Conf. on Robotics and Automation, volume 1, pages 521-528, 2000.

.... of the basic path planning problem have been solved with PRM based methods (e.g. 23] The experimental success of the planner has motivated many researchers to seek a theoretical basis for explaining its performance and relative successes in this direction have been reported, among others, in [5, 6, 16, 15, 28, 13, 26]. This paper presents a further extension in this direction by using the mechanism of measure theory [4] Typically, there are two related questions that one asks about PRM operation. How quickly can we find a path between some x and y How quickly can we find most of the paths we are interested ....

R. Bohlin and L. Kavraki. Path planning using lazy prm. In IEEE Int. Conf. Robot. & Autom., 2000.

....for its solution. Section 3 establishes the terminology that will be used in the rest of the paper. Section 4 gives a description of the basic PRM, a number of variations of the algorithm, and other closely related algorithms. In Section 5 we describe Lazy PRM. We draw our discussion from [6, 7, 8]. Related ideas about lazy evaluation have been developed concurrently and independently in [35] An analysis of the planner is given in Section 6 while experimental results are given in Section 7. We conclude in Section 8 with a discussion of the capabilities and limitations of Lazy PRM. Some of ....

R. Bohlin and L.E. Kavraki. Path planning using lazy PRM. In Proc. IEEE Int. Conf. on Rob. 4 Aut., 2000.

....have resulted in the successful application of these techniques to diverse domains, such as assembly planning, virtual prototyping, drug design, and computer animation. Much of the progress can be attributed to the introduction of probabilistic roadmap techniques [9] and their various extensions [1, 2, 7, 8, 15]. Despite these advances, however, some areas of application have still remained out of reach for automated planning algorithms. Applications requiring robots with many degrees of freedom to operate in highly dynamic and unpredictably changing environments fall into that category. To operate ....

R. Bohlin and L. E. Kavraki. Path planning using lazy PRM. In Proc. Intl. Conf. on Robotics and Automation, volume 1, pages 521-528, 2000.

....have resulted in the successful application of these techniques to diverse domains, such as assembly planning, virtual prototyping, drug design, and computer animation. Much of the progress can be attributed to the introduction of probabilistic roadmap techniques [8] and their various extentions [1, 2, 6, 7, 11, 13]. Despite these advances, however, some areas of application have still remained out of reach for automated planning algorithms. Applications requiring robots with many degrees of freedom to operate in highly dynamic and unpredictably changing environments fall into that category. To operate ....

Robert Bohlin and Lydia E. Kavraki. Path planning using lazy PRM. In Proc. of the Intl. Conf. on Robotics and Automation, volume 1, pages 521-528, 2000.

No context found.

R. Bohlin and L. Kavraki. Path planning using lazy prm. In IEEE Int. Conf. Robot. & Autom., 2000.

No context found.

R. Bohlin and L. Kavraki. Path planning using lazy PRM. In IEEE Int. Conf. Robot. & Autom., 2000.

No context found.

R. Bohlin and L. Kavraki. Path planning using Lazy PRM. In IEEE Int. Conf. Robot. & Autom., 2000.

No context found.

R. Bohlin and L. Kavraki. Path planning using lazy prm. In IEEE Int. Conf. Robot. & Autom., 2000.

No context found.

R. Bohlin and L. Kavraki. Path planning using Lazy PRM. In IEEE Int. Conf. Robot. & Autom., 2000.

No context found.

R. Bohlin and L. Kavraki. Path planning using lazy prm. In IEEE Int. Conf. Robot. & Autom., 2000. 7

No context found.

R. Bohlin and L. Kavraki, "Path planning using lazy PRM," in IEEE International Conference on Robotics and Automation, 2000, pp. 521-- 528.

No context found.

R. Bohlin and L. E. Kavraki, "Path planning using lazy PRM," in Proceedings of the 2000.

No context found.

Bohlin, R., L.E. Kavraki. Path Planning Using Lazy PRM. In Proc. IEEE Int. Conf. on Robotics & Automation, pp.521-528, 2000.

No context found.

R. Bohlin, L. Kavraki. "Path Planning Using Lazy PRM". In Proc. of the Int. Conf. on Robotics and Automation, 2000.

No context found.

R. Bohlin and L.E. Kavraki. Path planning using lazy prm. Proc. IEEE Int. Conf. on Robotics and Automation, pages 521--528, 2000.

No context found.

R. Bohlin and L. Kavraki. Path planning using Lazy PRM. In IEEE Int. Conf. Robot. & Autom., 2000.

No context found.

R. Bohlin and L.E. Kavraki, "Path planning using lazy prm," in Proceedings of the IEEE International Conference on Robotics and Automation, Seoul, May 2000, pp. 1469--1474.

No context found.

R. Bohlin and L. Kavraki. Path planning using Lazy PRM. In IEEE Int. Conf. Robot. & Autom., 2000.

No context found.

R. Bohlin and L. Kavraki. Path planning using Lazy PRM. In Proc. IEEE Int. Conf. Robot. & Autom. (ICRA), April 2000.

No context found.

R. Bohlin and L. Kavraki. Path planning using Lazy PRM. In Proc. IEEE Int. Conf. Robot. & Autom. (ICRA), April 2000.

No context found.

R. Bohlin and L. Kavraki. Path planning using Lazy PRM. In IEEE Int. Conf. Robot. & Autom., 2000.

No context found.

R. Bohlin and L. E. Kavraki. Path planning using Lazy PRM. In Proc. IEEE Int. Conf. Robot. Autom. (ICRA), pages 521-- 528, 2000.

No context found.

Robert Bohlin and Lydia E. Kavraki. Path planning using lazy PRM. In Proceedings of the International Conference on Robotics and Automation, volume 1, pages 521-528, San Francisco, USA, 2000.

No context found.

R. Bohlin, L.E. Kavraki, Path planning using lazy PRM, Proc. IEEE Int. Conf. on Robotics and Automation, 2000, pp. 521--528.

No context found.

R. Bohlin and L. Kavraki. Path planning using lazy prm. In IEEE Int. Conf. Robot. & Autom., 2000. 7

No context found.

R. Bohlin and L. Kavraki, "Path planning using lazy PRM", Proc. of the International Conference on Robotics and Automation (ICRA), San Francisco, USA, 2000, pp. 521-528.

No context found.

R. Bohlin and L. E. Kavraki, Path Planning Using Lazy PRM, Proceedings of the 2000.

No context found.

R. Bohlin and L. Kavraki. Path planning using lazy prm. In IEEE Int. Conf. Robot. & Autom., 2000.

No context found.

R. Bohlin and L. Kavraki. Path Planning using Lazy PRM. In IEEE Int. Conf. on Robotics and Automation, 2000.

No context found.

R. Bohlin and L. Kavraki. Path Planning using Lazy PRM. In Proc. of IEEE Int. Conference on Robotics and Automation, ICRA, 2000.

No context found.

R. Bohlin and L. Kavraki. Path planning using lazy prm. In International ConferenceonRobotics and Automation, San Francisco, CA, 2000.

First 50 documents  Next 50

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC