Parallel Path Planning with Multiple Evasion Strategies

Size: px
Start display at page:

Download "Parallel Path Planning with Multiple Evasion Strategies"

Transcription

1 Parallel Path Planning with Multiple Evasion Strategies Stefano Caselli, Monica Reggiani, Roberto Sbravati Dipartimento di Ingegneria dell Informazione Università di Parma Parma - Italy {caselli,reggiani,sbravati}@ce.unipr.it Abstract Probabilistic path planning driven by a potential field is a well established technique and has been successfully exploited to solve complex problems arising in a variety of domains. However, planners implementing this approach are rather inefficient in dealing with certain types of local minima occurring in the potential field, especially those characterized by deep or large attraction basins. In this paper, we present a potential field planner combining smart escape motions from local minima with parallel computation to improve overall performance. The results obtained show significant improvement in planning time, along with remarkable reduction in standard deviation. A performance comparison on a benchmark problem of the potential field planner and an existing, state-of-the-art planner is also included. Our investigation confirms the effectiveness of potential field as heuristic to solve difficult path planning problems. 1 Introduction One of the major approaches toward solving difficult, high dimensional path planning problems involves the definition of a suitable artificial potential field whose local negated gradient helps in guiding the robot configuration toward the goal. Since potential fields for practical problems unavoidably include local minima, the approach must be supplemented with a strategy for dealing with these local minima. Well-established potential field-based planners, such as the Random Path Planner (RPP) algorithm by Barraquand and Latombe [4] which has pioneered the approach, deal with local minima with brownian motion phases. While the effectiveness of potential field planners has been shown by their early application on highly relevant industrial case studies [10, 12, 21, 22], their performance is still inadequate to provide solution of many useful problems within acceptable times. Associated shortcomings include excessive variability in solution times, difficulty in dealing with deep local minima of the potential field, and need to compute the heuristic for each planning environment. Yet, given the computational difficulty of path planning, potential field planners, such as RPP, remain among the algorithms of choice for high dimensional problems. An alternative approach to probabilistic path planning is represented by roadmap methods, initiated by the Probabilistic RoadMap (PRM) algorithm [19, 20, 25]. Roadmap methods randomly sample collision-free configurations and connect them into a graph during a preprocessing phase; this graph is exploited as a roadmap to efficiently answer subsequent path queries [1, 14, 19, 25, 24]. In the last few years, probabilistic roadmap methods have become very popular. However, they suffer a major difficulty in dealing with narrow passages offering minimal visibility between connected components of C-free [16, 15]. It should be noted that both potential field and roadmap methods are prone to inefficient situations owing to the random decisions they make. This issue has been addressed in a few recent extensions of the basic roadmap approach [1, 24, 14], where random sampling of configurations is biased to make it more effective. One of the goals of this paper is to develop and evaluate a similar extension for a potential field planner. Several researchers have proposed to exploit parallel computing techniques to solve motion planning problems [13], and, specifically, in order to overcome shortcomings and extend the range of applicability of potential field planners. First, among others, Gini et al. [9] have developed an OR-parallel formulation of RPP based on random competition parallel search [11]. Qin and Henrich [26] have pursued an AND-parallel approach which generates random subgoals and then tries to connect them in parallel with the initial and final configurations. We have experimentally investigated a parallel RPP implementation, based on message passing, designed according to a random competition scheme, and found it very suitable for PC clusters, thanks to its minimal communication requirements [5]. These parallel planning approaches, along with those relying on parallel probabilistic roadmaps, e.g. [2], have further extended the range of planning problems successfully dealt with [23]. In this paper we describe a parallel potential field planner which aims to overcome some of the limitations of the potential field approach and thus expand its range of applicability. First, we enhance the efficiency of potential field planning by supplementing the random escape motion from local minima with more informed search strategies, while preserving the probabilistic completeness property of the algorithm. Second, we integrate the search strategies of the potential field planner into a parallel computation scheme which proves highly effective for many degree of freedom (d.o.f.) problems. These enhancements are being integrated in a new path planning tool, called parallel Potential Field Planner (ppfp), under development at the University of Parma, which also includes exploitation of past experience [6] and pluggable distance computation functions [27]. The paper is organized as follows. Section 2 reviews the operation of potential field planners, along with the extensions we propose to more efficiently deal with local minima. Section 3 describes a parallel computation framework which allows integration of multiple heuristics for

2 path planning. Section 4 reports experimental results for a set of 11 d.o.f. problems assessing the effectiveness of potential field planning when the proposed local minima heuristics are supported by parallel computation. This section also includes a performance comparison of ppfp with a state-of-the-art probabilistic roadmap planner on a benchmark problem proposed in the literature. A final section summarizes the contributions of the paper. 2 Potential Field Planners Algorithm 1 is a simplified version of a probabilistic complete algorithm exploiting a potential field U as heuristic for the solution of a path planning problem. Algorithm 1 Path planning using a potential field. 1: Execute a gradient motion 2: while current local minimum global minimum do 3: repeat 4: Execute an escape motion 5: until escaped OR max number of tries reached 6: if escaped then 7: Execute a gradient motion 8: else 9: backtracking 10: end if 11: if reached max number of gradient-escape motions then 12: return failure 13: end if 14: end while 15: return success The potential function U is obtained from a potential attracting the robot toward the goal (placed in the global minimum of U) and a repulsive potential around obstacles. Starting from the initial configuration q init of the robot, the algorithm executes a gradient motion, following the negated gradient of the potential U until it finds a local minimum q loc. If q loc is the global minimum q goal (i.e. U = 0), the algorithm terminates with success since a path has been found; otherwise the algorithm tries to escape from q loc executing an escape motion. Once the region of attraction of the local minimum has been successfully escaped, the escape motion is followed by a gradient motion that will find another (possibly new) local minimum. The algorithm therefore alternates gradient motions with escape motions until the goal configuration is reached. To guarantee the probabilistic completeness of the planning algorithm, a random motion is usually executed as escape motion. The idea of using a random motion as a way to escape from local minima has been first proposed in the RPP algorithm that pioneered this stream of research [4]. However, this random escape phase is quite inefficient, owing to its expensive exploration of the configuration space. The chaotic brownian search, indeed, often moves back to previously explored areas. Profiling data from RPP execution on several problems, including the problems described later in this paper, show that the time spent in the aggregated random motion phases is more than three times the time required by the gradient motion phases. Moreover, the average time spent executing random motion with respect to the total computation time increases from 43% for 7 d.o.f. problems to 75% for 11 d.o.f. problems, suggesting that the inefficiency of random escape may aggravate with the complexity of the problem. Barraquand, Langlois, and Latombe [3] have shown that the number of steps t(q l ) required to escape from the attraction domain of a local minimum by a random motion has a quadratic dependence on the average radius of the attraction basin R(q l ), i.e. t(q l ) = E [( R i (q l )] ) 2, i where i is the step size along the i-th axis. When the local minimum has a large attraction basin, i.e. high E [( R i (q l )] ) i, and at the same time a large passage to exit, this motion becomes inefficient. The random motion, indeed, must always execute a number of steps at least equal to the square of the attraction radius, in order to depart enough from the minimum and thus have sufficient chances to escape its attraction basin. In summary, improvements in the escape phase can significantly improve overall planning efficiency, especially in many d.o.f. problems. 2.1 Escape motion for easy local minima We contend that in typical planning problems, most local minima possess broad passages and are more efficiently managed using an optimistic approach. This is the rationale at the base of the approach presented in [7] and summarized in the following. Instead of executing expensive random motions guaranteeing probabilistic completeness, unguaranteed but fast motions are exploited to escape from easy local minima. Two such fast motions are briefly described in the following. A more detailed analysis, along with performance data obtained by sequential computation, is presented in [7]. Algorithm StraightLine (SL) simply selects a random direction from q loc in the n-dimensional C-space, and follows it until it reaches an obstacle or a configuration such that U(q) < U(q loc ). In case a joint limit is reached, a new admissible random direction is chosen leading to a zig-zag path. With either terminating conditions, SL is followed by a gradient motion which eventually leads to a local minimum. If the original local minimum or a minimum with higher potential is reached, the move is discarded and SL tries a new random direction. Up to a maximum number of random directions from q loc are searched until the local minimum is successfully escaped or the algorithm fails. StraightLineSelect (SLS) is a modified straight line escape motion. SLS tries to further optimize the basic idea by early pruning non-promising candidate directions. During each SLS move, the potential U(q) along the straight line or zig-zag path is tracked by a simple state machine. If along the chosen direction the potential increases monotonically, with high chance the ensuing down motion would bring to the same local minimum. These directions are thus discarded without actually performing the down motion. More promising are those directions where U(q) exhibits a non-monotone behavior a clue for a valley in the potential driving to a new local minimum. Only these promising random directions are thus actually followed by the down motion search in SLS. The average cost of an escape motion using the SL or SLS algorithms depends on the average number of attempts required to escape from the local minimum using the cho-

3 sen algorithm, E[X S], and the average cost of every attempt, R(q l )C cc + E[C pd ]q, where C cc is the cost of a collision detection operation, E[C pd ] the average cost of the down motion, and q the probability to execute the down motion. In SL, q = 1 because the down motion is always executed, whereas in SLS q depends on the probability that the SLS motion crosses a potential valley. The main computational advantage of SL and SLS with respect to brownian search stems from the linear rather than quadratic number of collision check operations they perform in escaping the local minimum. Additionally, SLS only performs the down path computation for a fraction q of the explored directions. However, for complex local minima these factors could be offset by an increase in the number of attempts E[X S]. The SL and SLS heuristics turn out to be effective in many easy local minima, with high probability to escape from the local minimum with a straight line motion and low E[X S], that are dealt with quite inefficiently by brownian search. These easy local minima are likely to occur rather frequently in many planning problems. For difficult local minima, however, the probability to escape with straight line motion decreases, possibly reaching zero, and SL and SLS could become more expensive than a random motion. Hence, we have modified SL and SLS, limiting the number of attempts they perform to a value empirically chosen based on statistical data. The planner, once reached this limit, autonomously classifies the problem as too difficult for the optimistic escape motion. It then reverts to random motion, thereby guaranteeing also the probabilistic completeness of the planning algorithm. 2.2 Comparing escape techniques We have evaluated the StraightLine and StraightLine- Select, along with the random motion in RPP (labeled as R in the following tables), on the four problems shown in Figures 1, all referring to a 7 d.o.f. robot consisting of 2 links with a free base. These planning problems have been designed to cover the workspace classification proposed by Hwang and Ahuja [17]. Problem 1 (Figure 1(a)) belongs to the set of easy problems, and fits the first class in Hwang and Ahuja s classification. The space among obstacles is large compared to robot dimensions and no narrow passage is present. Problems 2 and 3 (Figures 1(b) and 1(c)) belong to the second class, i.e. intermediate difficulty problems. The planner must find a path for the robot when its movements are restricted in a narrow space among close obstacles. Problem 4 (Figure 1(d)) belongs to the final class, comprising the most difficult problems. The robot is required to enter a narrow slot, where the goal is located. If the entering orientation is incorrect, the robot might be forced to exit the slot in order to re-orient itself, a situation likely involving a very deep local minimum. Table 1 reports average and standard deviation of the planning times achieved by the modified heuristics for the four problems, where each problem has been solved 100 times. Performance results in this section have been obtained with a 450 MHz Pentium-II PC with 256 MB main memory, under the Linux operating system. Table 1 shows the improvement attained by the SL and SLS heuristics over the standard random motion. The improvement is about a factor across the span of problems, and tends to be larger for the most difficult ones, problem R SL SLS 1 avg std avg std avg std avg std Table 1: Sequential planning times for 7 d.o.f. problems. (Times in seconds; 100 independent solutions per problem). where RPP random search indeed requires an increased fraction of time. 3 Parallel motion planning Given the complexity of motion planning, the goal of a parallel approach is both to decrease the time needed for the solution of a given problem and to improve the quality of the solution. The fundamental phase of the algorithms exploiting a potential field heuristic is the construction and search in the graph of the local minima. We have chosen not to pursue an AND-parallel approach because of the difficulty to attain a satisfactory subdvision of configuration space exploration among processes without incurring in high communication overhead. Instead, we have chosen to investigate an OR-parallel random competition strategy [11], with concurrent processes exploiting different heuristics in their search for a solution [8, 6]. Briefly, random competition exploits a set of independent parallel processes executing different algorithms or the same algorithm with different initial search seeds. The execution of the planning algorithm is coordinated by a master process. Each processing node executes one worker process. Thus, parallel path planning consists of the following phases: 1. startup phase: The master process transmits the entire task to all the worker processes and sets up the algorithms or the random seeds of the algorithm in order to differentiate the search paths. 2. working phase: Each worker attempts to solve the entire problem according to the master s instructions. A worker process interrupts its search in case of success or when it receives a message from the master. 3. termination: When a worker identifies a solution of the problem or fails (i.e. it does not find a solution within the predefined time), it communicates to the master the result of its own execution. After an arbitration among received messages, the master interrupts all workers and returns the final result. This approach always improves the solution time of the problem when random algorithms are used, or when different algorithms are available with problem-dependent relative performance. In these situations, execution times are variable and it is impossible to establish in advance which

4 (a) (b) (c) (d) Figure 1: A set of 7 d.o.f. problems: (a) easy; (b) and (c) intermediate difficulty; (d) difficult. algorithm or which initialization seed will yield the best performance. Moreover, since every process can solve the problem, even if one or more processes fail the remaining ones can still find the solution. Random competition is also flexible and simple to implement, since the addition of one processor does not imply changes to the code. As in random competition communication among processors is limited to the startup and termination phases, the approach introduces a minimal communication overhead and hence is well suited even for PC clusters. Algorithm 2 Parallel path planning in a potential field with Random competition. 1: if master then 2: for all workers do 3: Send initialization seed or choose the algorithm for the worker 4: end for 5: repeat 6: Receive a message from a worker 7: until winner s message OR all workers failed 8: if winner then 9: for all workers do 10: Stop the worker 11: end for 12: Output solution 13: else 14: Output failure message 15: end if 16: else 17: Execute a gradient motion 18: while current local minimum global minimum do 19: Check for a winner message from master 20: repeat 21: Execute an escape motion 22: until escaped OR max number of tries reached 23: Check for a winner message from master 24: if escaped then 25: Execute a gradient motion 26: else 27: backtracking 28: end if 29: if reached max number of gradient-escape motions then 30: Send a failure message to master 31: end if 32: end while 33: Send success message with solution to master 34: end if The random competition paradygm can be adapted in two different ways in order to introduce concurrency in algorithm 1. In the global random competition approach, all worker processes compete in the exploration of the whole search space without any cooperation until one of them finds a solution (algorithm 2). In the initialization phase the master starts a different algorithm or sends a different seed to each worker process (line 3). In the working phase each worker executes the path planning algorithm (lines 17-33). Finally, in the termination phase a winner worker sends the solution to the master (line 33), and the master terminates the other workers. In the local random competition approach, all worker processes compete in every escape phase from a local minimum. A competition round ends when one of the workers finds a new local minimum with a lower potential value. A new competition is then started from the new local minimum. While appealing in principle, local random competition requires expensive startup and termination phases for each local minimum, thus increasing the cost for collaboration among workers, which reduces the benefit of a parallel implementation. Furthermore, this approach tends to develop paths connecting many close local minima, thereby decreasing the effectiveness of the new SL and SLS heuristics. We have tested a POSIX thread-based implementation of local random competition on an SMP shared memory computer and found it rather inefficient indeed. The experimental results of the global approach are presented in the next section. 4 Experimental results This section reports the results obtained combining the smart techniques to escape from local minima described in Section 2 with the parallel computation approach described in Section 3. The data in Tables 2 and 3 refer to experiments involving an 11 d.o.f. robot in an environment similar to Problem 2 (Figure 1(b)) solved using three workers. Analogous results have been obtained for robots with different number of d.o.f. s and using more workers. The results relying on parallel computation reported in this section have been obtained on a low-cost parallel architecture consisting of a NOW platform of 450 MHz Pentium-II PCs, with 256 MB main memory per node. Four sets of experiments have been carried out on ten

5 Table 2: Average execution time for a set of ten 11 d.o.f. problems solved by random competition among three workers. (Times in seconds; 20 independent solutions per problem). goal 3R 3SL 3SLS R+SL+SLS Execution time (sec.) R+R+R SL+SL+SL SLS+SLS+SLS R+SL+SLS Figure 2: Average execution time for the ten problems with an 11 d.o.f. robot (20 independent solutions per problem). Table 3: Standard deviation from the mean for a set of ten 11 d.o.f. problems solved by random competition among three workers. (Times in seconds; 20 independent solutions per problem.) goal 3R 3SL 3SLS R+SL+SLS SLS SL R problems, all sharing the same environment of Figure 1(b) and the 11 d.o.f. robot but differing in the goal position (generated randomly). The first three sets of experiments solve the problems with a random competition among three workers, all using the same escape technique (random motion, StraightLine, and StraightLineSelect, respectively) but differing in the initialization seed. In the fourth experiment, instead, each worker uses a different escape technique. The ten problem routine is identical in the four experiments and each experiment comprises 20 independent executions with a fresh set of initial random seeds. Tables 2 and 3 report the average execution time and the standard deviation from the mean resulting from the experiments. These experimental results show the remarkable improvement in execution time and standard deviation obtained by the proposed escape techniques. The average planning time reduces, across the set of 11 d.o.f. problems investigated, of a factor between 10 (for SL) and 13 (for SLS) compared to the planning time of three workers exploiting the standard random motion heuristic. The additional gains of reduced variance in planning time and improved path quality are also essential for practical path planning applications. Random competition, indeed, increases the likelihood that at least one of the workers can Figure 3: Percentage of problem instances solved by Random motion, SL, and SLS, on the set of ten 11 d.o.f. problems (20 solutions per problem). reach the goal going through a reduced number of difficult local minima. This situation vastly reduces execution time, as random motions now become the exception rather than the rule. While both SL and SLS show large improvement over random motion (see Figure 2, showing graphically the same data of Table 2), the marginal advantage of SLS over SL is inconclusive until supported by additional evidence. Figure 3 refers to the final set of experiments previously discussed, i.e. to R+SL+SLS concurrent searches. The winner process has been traced during the 20 solutions of the ten problems. The percentage of success of the different techniques in finding the path for each problem is shown in Figure 3. Usually, the worker process exploiting the random motion during its escape phase is slower than the other workers. Only for a few problems, such as the eighth, random motion is the winner for a significant number of solutions. This analysis confirms the effectiveness of the SL and SLS heuristics.

6 Table 4: Average execution time and standard deviation for 30 independent sequential solutions of the alpha1.5 benchmark. (Times in seconds; 30 independent solutions). solver avg std PRM PFP Figure 4: The alpha puzzle. 4.1 Comparing performance against alternative approaches A hovering question is whether a potential field planner such as ppfp is indeed a competitive planner, considering the availability of advanced path planners based on alternative approaches. This section reports a performance comparison of PFP, the sequential build of ppfp, with a state-of-the-art probabilistic roadmap planner on a benchmark problem proposed in the literature. The comparison is admittedly very limited (based on a single problem, a single alternative package, and only for sequential solution), owing to the many idiosyncrasies and input format incompatibilities of the different planners. The chosen reference planner is the wellknown Probabilistic RoadMap planner (PRM) for rigid objects in 3D [18]. The PRM package has been made available by prof. L. Kavraki of Rice University at: We have compared the performance of PFP and PRM in solving the alpha puzzle problem shown in Figure 4. This problem has been proposed by prof. Amato of (Texas A&M University) as a standard benchmark [28] for comparing path planning performance and is available at benchmarks/. One of the alpha rings must be moved with respect to the other until they are separated. The original problem, even though only 6 d.o.f., turns out to be very difficult, owing to the very small clearance between the rings. Scaled problem versions are obtained by proportionally increasing the gap between the two wings of a ring. Performance results in Table 4 refer to the scaled 1.5 version. PRM has been executed with its default parameter settings. Apparently, PRM solution time exhibits very high variance on this problem. Indeed, in a few instances PRM has been able to solve the problem in a short time with the first set of random nodes an no enhanced nodes, whereas in other cases many additional random and enhanced nodes are required, leading to very high computation times. We should also emphasize that PRM is designed to efficiently solve multiple query problems, rather than a single problem in a given workspace. We reiterate that results in Table 4 should be looked at with some caution, as they refer to a single problem; however, they suggest that advanced planners based on potential field should at least be regarded as a credible alternative to the prevailing roadmap planners for single shot problems. 5 Conclusions and future work The paper has presented a potential field planner combining parallel computation and smart escape techniques from local minima in order to attain good performance in solving complex path planning problems. The results reported in the paper confirm the effectiveness of potential field as heuristic for the solution of path planning problems. Moreover, they show that significant improvement in performance can be obtained without compromising completeness. In the heuristics proposed in this paper, random motion is only resorted to when the minimum is classified as a difficult one, thereby guaranteeing the probabilistic completeness of the proposed algorithm. As far as the future developments are concerned, we plan to extend the parallel implementation of our path planner to allow different levels of collaboration among the processes. We conjecture that the optimal competition level among workers should be somewhere between the two alternatives investigated, i.e. global path search and competition at each local minimum. Acknowledgments We thank prof. Lydia Kavraki (Rice University) for making available to us her roadmap planner, and prof. Nancy Amato (Texas A&M University) for making available to the research community the alpha puzzle benchmark. Work on this paper has been supported in part by MURST, Italy (Project ISIDE, Sviluppo di sistemi di elaborazione reattivi ed affidabili per applicazioni industriali) and by a research project supported by ENEA on parallel visualization techniques for robot systems. References [1] N. Amato and Y. Wu. A randomized roadmap method for path and manipulation planning. In IEEE International Conference on Robotics and Automation, pages , Minneapolis, MN, [2] N. M. Amato and L. K. Dale. Probabilistic roadmap methods are embarrassingly parallel. In IEEE International

7 Conference on Robotics and Automation, pages , Detroit, MI, [3] J. Barraquand, B. Langlois, and J.-C. Latombe. Numerical potential field techniques for robot path planning. IEEE Transactions on Systems, Man, and Cybernetics, 22(2): , [4] J. Barraquand and J.-C. Latombe. Robot motion planning: a distributed representation approach. Internation Journal of Robotics Research, 10(6): , [5] S. Caselli and M. Reggiani. Randomized motion planning on parallel and distributed architectures. In Euromicro Workshop on Parallel and Distributed Processing, pages , Funchal, Portugal, [6] S. Caselli and M. Reggiani. ERPP: an Experience-based Randomized Path Planner. In IEEE International Conference on Robotics and Automation, pages , San Francisco, CA, [7] S. Caselli, M. Reggiani, and R. Rocchi. Heuristic methods for randomized path planning in potential fields. In IEEE International Symposium on Computational Intelligence in Robotics and Automation, Banff, Canada, [8] D. J. Challou, D. Boley, M. Gini, and V. Kumar. A parallel formulation of informed randomized search for robot motion planning problems. In IEEE International Conference on Robotics and Automation, pages , Nagoya, Japan, [9] D. J. Challou, M. Gini, and V. Kumar. Parallel search algorithms for robot motion planning. In IEEE International Conference on Robotics and Automation, pages 46 51, Atlanta, GA, [10] H. Chang and T.-J. Li. Assembly maintainability study with motion planning. In IEEE International Conference on Robotics and Automation, pages , Nagoya, Japan, [11] W. Ertel. Random competition: a simple, but efficient method for parallelizing inference systems. Technical Report FKI , Technische Universitat at Munchen, [12] L. Graux, P. Millies, P. L. Kociemba, and B. Langlois. Integration of a path generations algorithm into off-line programming of AIRBUS panels. In Aerospace Automated Fastening Conference and Exposition, SAE Technical Paper [13] D. Henrich. Fast motion planning by parallel processing a review. Journal of Intelligent and Robotic Systems, 20(1):45 69, [14] D. Hsu. Randomized Single-Query Motion Planning in Expansive Spaces. PhD thesis, Dept. of Computer Science, Stanford University, Stanford, CA, [15] D. Hsu, L. E. Kavraki, J.-C. Latombe, R. Motwani, and S. Sorkin. On finding narrow passages with probabilistic roadmap planners. In Workshop on the Algorithmic Foundations of Robotics, Houston, TX, [16] D. Hsu, J.-C. Latombe, and R. Motwani. Path planning in expansive configuration spaces. In IEEE International Conference on Robotics and Automation, pages , Albuquerque, NM, [17] Y. K. Hwang and N. Ahuja. Gross motion planning - a survey. ACM Computing Surveys, 24(3): , [18] L. Kavraki and J.-C. Latombe. Probabilistic roadmaps for robot path planning. In K. Gupta and P. del Pobil, editors, Pratical Motion Planning in Robotics:Current Approaches and Future Directions, pages John Wiley, [19] L. E. Kavraki and J.-C. Latombe. Randomized preprocessing of configuration space for fast path planning. In IEEE International Conference on Robotics and Automation, pages , San Diego, CA, [20] L. E. Kavraki, P. Švestka, J.-C. Latombe, and M. H. Overmars. Probabilistic roadmaps for path planning in highdimensional configuration spaces. IEEE Transaction on Robotics and Automation, 12(4): , [21] Y. Koga, K. Kondo, J. J. Kuffner, and J.-C. Latombe. Planning motions with intentions. In SIGGRAPH 94, pages , Orlando, FL, [22] Y. Koga and J.-C. Latombe. On multi-arm manipulation planning. In IEEE International Conference on Robotics and Automation, pages , S. Diego, CA, [23] J.-C. Latombe. Motion planning: A journey of robots, molecules, digital actors, and other artifacts. Internation Journal of Robotics Research, 18(11): , [24] S. M. LaValle and J. J. Kuffner. Rapidly-exploring random trees: Progress and prospects. In Fourth Workshop on the Algorithmic Foundations of Robotics, Hanover, NH, [25] M. H. Overmars and P. Švestka. A probabilistic learning approach to motion planning. Technical Report UU-CS , Department of Computer Science, University of Utrecht, [26] C. Qin and D. Henrich. Randomized parallel motion planning for robot manipulation. Technical Report Computer Science 5/96, University of Karlsruhe, Germany, [27] M. Reggiani, M. Mazzoli, and S. Caselli. An experimental evaluation of collision detection packages for robot motion planning. Submitted to IEEE Int. Conf. on Robotics and Automation, [28] D. Vallejo, I. Remmler, and N. M. Amato. An adaptive framework for single shot motion planning: A self-tuning system for rigid and articulated robots. In IEEE International Conference on Robotics and Automation, pages 21 26, Seoul, Korea, 2001.

Providing Haptic Hints to Automatic Motion Planners

Providing Haptic Hints to Automatic Motion Planners Providing Haptic Hints to Automatic Motion Planners O. Burchan Bayazit Guang Song Nancy M. Amato fburchanb,gsong,amatog@cs.tamu.edu Department of Computer Science, Texas A&M University College Station,

More information

Exploiting collision information in probabilistic roadmap planning

Exploiting collision information in probabilistic roadmap planning Exploiting collision information in probabilistic roadmap planning Serene W. H. Wong and Michael Jenkin Department of Computer Science and Engineering York University, 4700 Keele Street Toronto, Ontario,

More information

INCREASING THE CONNECTIVITY OF PROBABILISTIC ROADMAPS VIA GENETIC POST-PROCESSING. Giuseppe Oriolo Stefano Panzieri Andrea Turli

INCREASING THE CONNECTIVITY OF PROBABILISTIC ROADMAPS VIA GENETIC POST-PROCESSING. Giuseppe Oriolo Stefano Panzieri Andrea Turli INCREASING THE CONNECTIVITY OF PROBABILISTIC ROADMAPS VIA GENETIC POST-PROCESSING Giuseppe Oriolo Stefano Panzieri Andrea Turli Dip. di Informatica e Sistemistica, Università di Roma La Sapienza, Via Eudossiana

More information

II. RELATED WORK. A. Probabilistic roadmap path planner

II. RELATED WORK. A. Probabilistic roadmap path planner Gaussian PRM Samplers for Dynamic Configuration Spaces Yu-Te Lin and Shih-Chia Cheng Computer Science Department Stanford University Stanford, CA 94305, USA {yutelin, sccheng}@cs.stanford.edu SUID: 05371954,

More information

CS 4649/7649 Robot Intelligence: Planning

CS 4649/7649 Robot Intelligence: Planning CS 4649/7649 Robot Intelligence: Planning Probabilistic Roadmaps Sungmoon Joo School of Interactive Computing College of Computing Georgia Institute of Technology S. Joo (sungmoon.joo@cc.gatech.edu) 1

More information

T S. Configuration Space S1 LP S0 ITM. S3 Start cfg

T S. Configuration Space S1 LP S0 ITM. S3 Start cfg An Adaptive Framework for `Single Shot' Motion Planning Λ Daniel R. Vallejo Christopher Jones Nancy M. Amato Texas A&M University Sandia National Laboratories Texas A&M University dvallejo@cs.tamu.edu

More information

Workspace Skeleton Tools for Motion Planning

Workspace Skeleton Tools for Motion Planning Workspace Skeleton Tools for Motion Planning Kiffany Lyons, Diego Ruvalcaba, Mukulika Ghosh, Shawna Thomas, and Nancy Amato Abstract Motion planning is the ability to find a valid path from a start to

More information

A Two-level Search Algorithm for Motion Planning

A Two-level Search Algorithm for Motion Planning Proceedings of the 1997 IEEE International Conference on Robotics and Automation, IEEE Press, 2025-2031. A Two-level Search Algorithm for Motion Planning Pekka Isto Laboratory of Information Processing

More information

Workspace Skeleton Tools for Motion Planning

Workspace Skeleton Tools for Motion Planning Workspace Skeleton Tools for Motion Planning Diego Ruvalcaba, Kiffany Lyons, Mukulika Ghosh, Shawna Thomas, and Nancy Amato Abstract Motion planning is the ability to find a valid path from a start to

More information

Geometric Path Planning McGill COMP 765 Oct 12 th, 2017

Geometric Path Planning McGill COMP 765 Oct 12 th, 2017 Geometric Path Planning McGill COMP 765 Oct 12 th, 2017 The Motion Planning Problem Intuition: Find a safe path/trajectory from start to goal More precisely: A path is a series of robot configurations

More information

vizmo++: a Visualization, Authoring, and Educational Tool for Motion Planning

vizmo++: a Visualization, Authoring, and Educational Tool for Motion Planning vizmo++: a Visualization, Authoring, and Educational Tool for Motion Planning Aimée Vargas E. Jyh-Ming Lien Nancy M. Amato. aimee@cs.tamu.edu neilien@cs.tamu.edu amato@cs.tamu.edu Technical Report TR05-011

More information

Nonholonomic motion planning for car-like robots

Nonholonomic motion planning for car-like robots Nonholonomic motion planning for car-like robots A. Sánchez L. 2, J. Abraham Arenas B. 1, and René Zapata. 2 1 Computer Science Dept., BUAP Puebla, Pue., México {aarenas}@cs.buap.mx 2 LIRMM, UMR5506 CNRS,

More information

Clearance Based Path Optimization for Motion Planning

Clearance Based Path Optimization for Motion Planning Clearance Based Path Optimization for Motion Planning Roland Geraerts Mark Overmars institute of information and computing sciences, utrecht university technical report UU-CS-2003-039 www.cs.uu.nl Clearance

More information

6.141: Robotics systems and science Lecture 9: Configuration Space and Motion Planning

6.141: Robotics systems and science Lecture 9: Configuration Space and Motion Planning 6.141: Robotics systems and science Lecture 9: Configuration Space and Motion Planning Lecture Notes Prepared by Daniela Rus EECS/MIT Spring 2012 Figures by Nancy Amato, Rodney Brooks, Vijay Kumar Reading:

More information

Balancing Exploration and Exploitation in Motion Planning

Balancing Exploration and Exploitation in Motion Planning 2008 IEEE International Conference on Robotics and Automation Pasadena, CA, USA, May 19-23, 2008 Balancing Exploration and Exploitation in Motion Planning Markus Rickert Oliver Brock Alois Knoll Robotics

More information

Providing Haptic Hints to Automatic Motion Planners

Providing Haptic Hints to Automatic Motion Planners Providing Haptic Hints to Automatic Motion Planners Nancy M. Amato O. Burchan Bayazit Kyunghwan Kim amato@cs.tamu.edu burchanb@cs.tamu.edu kimk@cs.tamu.edu Wookho Son wooks@cs.tamu.edu Guang Song gsong@cs.tamu.edu

More information

cbook 26/2/ :49 PAGE PROOFS for John Wiley & Sons Ltd (jwcbook.sty v5.0, )

cbook 26/2/ :49 PAGE PROOFS for John Wiley & Sons Ltd (jwcbook.sty v5.0, ) Lydia E. Kavraki Jean-Claude Latombe! Department of Computer Scince, Rice University, Houston, TX 77005! Department of Computer Science, Stanford University, Stanford, CA 94305 Abstract The Probabilistic

More information

The Corridor Map Method: Real-Time High-Quality Path Planning

The Corridor Map Method: Real-Time High-Quality Path Planning 27 IEEE International Conference on Robotics and Automation Roma, Italy, 1-14 April 27 WeC11.3 The Corridor Map Method: Real-Time High-Quality Path Planning Roland Geraerts and Mark H. Overmars Abstract

More information

Probabilistic Roadmap Planner with Adaptive Sampling Based on Clustering

Probabilistic Roadmap Planner with Adaptive Sampling Based on Clustering Proceedings of the 2nd International Conference of Control, Dynamic Systems, and Robotics Ottawa, Ontario, Canada, May 7-8 2015 Paper No. 173 Probabilistic Roadmap Planner with Adaptive Sampling Based

More information

Geometric Path Planning for General Robot Manipulators

Geometric Path Planning for General Robot Manipulators Proceedings of the World Congress on Engineering and Computer Science 29 Vol II WCECS 29, October 2-22, 29, San Francisco, USA Geometric Path Planning for General Robot Manipulators Ziyad Aljarboua Abstract

More information

Probabilistic Motion Planning: Algorithms and Applications

Probabilistic Motion Planning: Algorithms and Applications Probabilistic Motion Planning: Algorithms and Applications Jyh-Ming Lien Department of Computer Science George Mason University Motion Planning in continuous spaces (Basic) Motion Planning (in a nutshell):

More information

Workspace-Guided Rapidly-Exploring Random Tree Method for a Robot Arm

Workspace-Guided Rapidly-Exploring Random Tree Method for a Robot Arm WorkspaceGuided RapidlyExploring Random Tree Method for a Robot Arm Jaesik Choi choi31@cs.uiuc.edu August 12, 2007 Abstract Motion planning for robotic arms is important for real, physical world applications.

More information

Variable-resolution Velocity Roadmap Generation Considering Safety Constraints for Mobile Robots

Variable-resolution Velocity Roadmap Generation Considering Safety Constraints for Mobile Robots Variable-resolution Velocity Roadmap Generation Considering Safety Constraints for Mobile Robots Jingyu Xiang, Yuichi Tazaki, Tatsuya Suzuki and B. Levedahl Abstract This research develops a new roadmap

More information

crrt : Planning loosely-coupled motions for multiple mobile robots

crrt : Planning loosely-coupled motions for multiple mobile robots crrt : Planning loosely-coupled motions for multiple mobile robots Jan Rosell and Raúl Suárez Institute of Industrial and Control Engineering (IOC) Universitat Politècnica de Catalunya (UPC) Barcelona

More information

biasing the sampling based on the geometric characteristics of configurations known to be reachable from the assembled configuration (the start). In p

biasing the sampling based on the geometric characteristics of configurations known to be reachable from the assembled configuration (the start). In p Disassembly Sequencing Using a Motion Planning Approach Λ Sujay Sundaram Department of Mechanical Engineering Texas A&M University College Station, TX 77843 sps7711@cs.tamu.edu Ian Remmler Nancy M. Amato

More information

Optimizing Schedules for Prioritized Path Planning of Multi-Robot Systems

Optimizing Schedules for Prioritized Path Planning of Multi-Robot Systems Proceedings of the 20 IEEE International Conference on Robotics & Automation Seoul, Korea May 21-26, 20 Optimizing Schedules for Prioritized Path Planning of Multi-Robot Systems Maren Bennewitz y Wolfram

More information

CHAPTER SIX. the RRM creates small covering roadmaps for all tested environments.

CHAPTER SIX. the RRM creates small covering roadmaps for all tested environments. CHAPTER SIX CREATING SMALL ROADMAPS Many algorithms have been proposed that create a roadmap from which a path for a moving object can be extracted. These algorithms generally do not give guarantees on

More information

Introduction to Robotics

Introduction to Robotics Jianwei Zhang zhang@informatik.uni-hamburg.de Universität Hamburg Fakultät für Mathematik, Informatik und Naturwissenschaften Technische Aspekte Multimodaler Systeme 05. July 2013 J. Zhang 1 Task-level

More information

Can work in a group of at most 3 students.! Can work individually! If you work in a group of 2 or 3 students,!

Can work in a group of at most 3 students.! Can work individually! If you work in a group of 2 or 3 students,! Assignment 1 is out! Due: 26 Aug 23:59! Submit in turnitin! Code + report! Can work in a group of at most 3 students.! Can work individually! If you work in a group of 2 or 3 students,! Each member must

More information

Figure 1: A typical industrial scene with over 4000 obstacles. not cause collisions) into a number of cells. Motion is than planned through these cell

Figure 1: A typical industrial scene with over 4000 obstacles. not cause collisions) into a number of cells. Motion is than planned through these cell Recent Developments in Motion Planning Λ Mark H. Overmars Institute of Information and Computing Sciences, Utrecht University, P.O. Box 80.089, 3508 TB Utrecht, the Netherlands. Email: markov@cs.uu.nl.

More information

Mobile Robot Path Planning in Static Environments using Particle Swarm Optimization

Mobile Robot Path Planning in Static Environments using Particle Swarm Optimization Mobile Robot Path Planning in Static Environments using Particle Swarm Optimization M. Shahab Alam, M. Usman Rafique, and M. Umer Khan Abstract Motion planning is a key element of robotics since it empowers

More information

Sampling-Based Robot Motion Planning. Lydia Kavraki Department of Computer Science Rice University Houston, TX USA

Sampling-Based Robot Motion Planning. Lydia Kavraki Department of Computer Science Rice University Houston, TX USA Sampling-Based Robot Motion Planning Lydia Kavraki Department of Computer Science Rice University Houston, TX USA Motion planning: classical setting Go from Start to Goal without collisions and while respecting

More information

RRT-Connect: An Efficient Approach to Single-Query Path Planning

RRT-Connect: An Efficient Approach to Single-Query Path Planning In Proc. 2000 IEEE Int l Conf. on Robotics and Automation (ICRA 2000) RRT-Connect: An Efficient Approach to Single-Query Path Planning James J. Kuffner, Jr. Computer Science Dept. Stanford University Stanford,

More information

Decomposition-based Motion Planning: Towards Real-time Planning for Robots with Many Degrees of Freedom. Abstract

Decomposition-based Motion Planning: Towards Real-time Planning for Robots with Many Degrees of Freedom. Abstract Decomposition-based Motion Planning: Towards Real-time Planning for Robots with Many Degrees of Freedom Oliver Brock Lydia E. Kavraki Department of Computer Science Rice University, Houston, Texas 77005

More information

Kinodynamic Motion Planning by Interior-Exterior Cell Exploration

Kinodynamic Motion Planning by Interior-Exterior Cell Exploration Kinodynamic Motion Planning by Interior-Exterior Cell Exploration Ioan A. Şucan 1 and Lydia E. Kavraki 1 Department of Computer Science, Rice University, {isucan, kavraki}@rice.edu Abstract: This paper

More information

Mobile Robots Path Planning using Genetic Algorithms

Mobile Robots Path Planning using Genetic Algorithms Mobile Robots Path Planning using Genetic Algorithms Nouara Achour LRPE Laboratory, Department of Automation University of USTHB Algiers, Algeria nachour@usthb.dz Mohamed Chaalal LRPE Laboratory, Department

More information

Configuration Space. Ioannis Rekleitis

Configuration Space. Ioannis Rekleitis Configuration Space Ioannis Rekleitis Configuration Space Configuration Space Definition A robot configuration is a specification of the positions of all robot points relative to a fixed coordinate system

More information

UOBPRM: A Uniformly Distributed Obstacle-Based PRM

UOBPRM: A Uniformly Distributed Obstacle-Based PRM UOBPRM: A Uniformly Distributed Obstacle-Based PRM Hsin-Yi (Cindy) Yeh 1, Shawna Thomas 1, David Eppstein 2 and Nancy M. Amato 1 Abstract This paper presents a new sampling method for motion planning that

More information

6.141: Robotics systems and science Lecture 9: Configuration Space and Motion Planning

6.141: Robotics systems and science Lecture 9: Configuration Space and Motion Planning 6.141: Robotics systems and science Lecture 9: Configuration Space and Motion Planning Lecture Notes Prepared by Daniela Rus EECS/MIT Spring 2011 Figures by Nancy Amato, Rodney Brooks, Vijay Kumar Reading:

More information

A Hybrid Approach for Complete Motion Planning

A Hybrid Approach for Complete Motion Planning A Hybrid Approach for Complete Motion Planning Liangjun Zhang 1 Young J. Kim 2 Dinesh Manocha 1 1 Dept. of Computer Science, University of North Carolina at Chapel Hill, USA, {zlj,dm}@cs.unc.edu 2 Dept.

More information

Planning Movement of a Robotic Arm for Assembly of Products

Planning Movement of a Robotic Arm for Assembly of Products Journal of Mechanics Engineering and Automation 5 (2015) 257-262 doi: 10.17265/2159-5275/2015.04.008 D DAVID PUBLISHING Planning Movement of a Robotic Arm for Assembly of Products Jose Ismael Ojeda Campaña

More information

Enhancing Randomized Motion Planners: Exploring with Haptic Hints

Enhancing Randomized Motion Planners: Exploring with Haptic Hints Enhancing Randomized Motion Planners: Exploring with Haptic Hints O. Burchan Bayazit Guang Song Nancy M. Amato burchanb@cs.tamu.edu gsong@cs.tamu.edu amato@cs.tamu.edu Technical Report 99-021 Department

More information

Learning to Guide Random Tree Planners in High Dimensional Spaces

Learning to Guide Random Tree Planners in High Dimensional Spaces Learning to Guide Random Tree Planners in High Dimensional Spaces Jörg Röwekämper Gian Diego Tipaldi Wolfram Burgard Fig. 1. Example paths for a mobile manipulation platform computed with RRT-Connect [13]

More information

Efficient Motion Planning for Humanoid Robots using Lazy Collision Checking and Enlarged Robot Models

Efficient Motion Planning for Humanoid Robots using Lazy Collision Checking and Enlarged Robot Models Efficient Motion Planning for Humanoid Robots using Lazy Collision Checking and Enlarged Robot Models Nikolaus Vahrenkamp, Tamim Asfour and Rüdiger Dillmann Institute of Computer Science and Engineering

More information

Choosing Good Distance Metrics and Local Planners for Probabilistic Roadmap Methods

Choosing Good Distance Metrics and Local Planners for Probabilistic Roadmap Methods Choosing Good and Local Planners for Probabilistic Roadmap Methods Nancy M. Amato O. Burchan Bayazit Lucia K. Dale Christopher Jones Daniel Vallejo Technical Report 98- Department of Computer Science Texas

More information

Probabilistic roadmaps for efficient path planning

Probabilistic roadmaps for efficient path planning Probabilistic roadmaps for efficient path planning Dan A. Alcantara March 25, 2007 1 Introduction The problem of finding a collision-free path between points in space has applications across many different

More information

Single-Query Motion Planning with Utility-Guided Random Trees

Single-Query Motion Planning with Utility-Guided Random Trees 27 IEEE International Conference on Robotics and Automation Roma, Italy, 1-14 April 27 FrA6.2 Single-Query Motion Planning with Utility-Guided Random Trees Brendan Burns Oliver Brock Department of Computer

More information

A Slicing Connection Strategy for Constructing PRMs in High-Dimensional Cspaces

A Slicing Connection Strategy for Constructing PRMs in High-Dimensional Cspaces A Slicing Connection Strategy for Constructing PRMs in High-Dimensional Cspaces Abstract - This paper presents a connection strategy for PRM-based motion planning in high-dimensional configuration spaces.

More information

1 Introduction Automatic motion planning deals with finding a feasible sequence of motions to take some moving object (the `robot') from a given initi

1 Introduction Automatic motion planning deals with finding a feasible sequence of motions to take some moving object (the `robot') from a given initi Customizing PRM Roadmaps at Query Time Λ Guang Song Shawna Miller Nancy M. Amato Dept. of Computer Science Dept. of Computer Science Dept. of Computer Science gsong@cs.tamu.edu slm5563@cs.tamu.edu amato@cs.tamu.edu

More information

Incrementally Reducing Dispersion by Increasing Voronoi Bias in RRTs

Incrementally Reducing Dispersion by Increasing Voronoi Bias in RRTs Incrementally Reducing Dispersion by Increasing Voronoi Bias in RRTs Stephen R. Lindemann Steven M. LaValle Dept. of Computer Science University of Illinois Urbana, IL 61801 USA {slindema, lavalle}@uiuc.edu

More information

Path Planning among Movable Obstacles: a Probabilistically Complete Approach

Path Planning among Movable Obstacles: a Probabilistically Complete Approach Path Planning among Movable Obstacles: a Probabilistically Complete Approach Jur van den Berg 1, Mike Stilman 2, James Kuffner 3, Ming Lin 1, and Dinesh Manocha 1 1 Department of Computer Science, University

More information

Planning. Daniel Vallejo Christopher Jones Nancy M. Amato. Texas A&M University. Abstract

Planning. Daniel Vallejo Christopher Jones Nancy M. Amato. Texas A&M University. Abstract An Adaptive Framework for `Single Shot' Motion Planning Daniel Vallejo Christopher Jones Nancy M. Amato dvallejo@cs.tamu.edu cvj3341@cs.tamu.edu amato@cs.tamu.edu Technical Report 99-024 Department of

More information

Sampling-based Planning 2

Sampling-based Planning 2 RBE MOTION PLANNING Sampling-based Planning 2 Jane Li Assistant Professor Mechanical Engineering & Robotics Engineering http://users.wpi.edu/~zli11 Problem with KD-tree RBE MOTION PLANNING Curse of dimension

More information

Parallel Search Algorithms for Robot Motion Planning *

Parallel Search Algorithms for Robot Motion Planning * From: AAAI Technical Report SS-93-04. Compilation copyright 1993, AAAI (www.aaai.org). All rights reserved. Parallel Search Algorithms for Robot Motion Planning * Daniel J. Challou Maria Gini Department

More information

Planning with Reachable Distances: Fast Enforcement of Closure Constraints

Planning with Reachable Distances: Fast Enforcement of Closure Constraints Planning with Reachable Distances: Fast Enforcement of Closure Constraints Xinyu Tang Shawna Thomas Nancy M. Amato xinyut@cs.tamu.edu sthomas@cs.tamu.edu amato@cs.tamu.edu Technical Report TR6-8 Parasol

More information

Incremental Map Generation (IMG)

Incremental Map Generation (IMG) Incremental Map Generation (IMG) Dawen Xie, Marco A. Morales A., Roger Pearce, Shawna Thomas, Jyh-Ming Lien, Nancy M. Amato {dawenx,marcom,rap2317,sthomas,neilien,amato}@cs.tamu.edu Technical Report TR06-005

More information

The Toggle Local Planner for Sampling-Based Motion Planning

The Toggle Local Planner for Sampling-Based Motion Planning The Toggle Local Planner for Sampling-Based Motion Planning Jory Denny and Nancy M. Amato Abstract Sampling-based solutions to the motion planning problem, such as the probabilistic roadmap method (PRM),

More information

Steps Toward Derandomizing RRTs

Steps Toward Derandomizing RRTs Steps Toward Derandomizing RRTs Stephen R. Lindemann Steven M. LaValle Dept. of Computer Science University of Illinois Urbana, IL 61801 USA {slindema, lavalle}@uiuc.edu Abstract We present two motion

More information

RESAMPL: A Region-Sensitive Adaptive Motion Planner

RESAMPL: A Region-Sensitive Adaptive Motion Planner REAMPL: A Region-ensitive Adaptive Motion Planner amuel Rodriguez, hawna Thomas, Roger Pearce, and Nancy M. Amato Parasol Lab, Department of Computer cience, Texas A&M University, College tation, TX UA

More information

Specialized PRM Trajectory Planning For Hyper-Redundant Robot Manipulators

Specialized PRM Trajectory Planning For Hyper-Redundant Robot Manipulators Specialized PRM Trajectory Planning For Hyper-Redundant Robot Manipulators MAHDI F. GHAJARI and RENE V. MAYORGA Department of Industrial Systems Engineering University of Regina 3737 Wascana Parkway, Regina,

More information

Merging the adaptive random walks planner with the randomized potential field planner

Merging the adaptive random walks planner with the randomized potential field planner Merging the adaptive random walks planner with the randomized potential field planner Stefano Carpin School of Engineering and Science International University Bremen Germany Gianluigi Pillonetto Department

More information

OOPS for Motion Planning: An Online, Open-source, Programming System

OOPS for Motion Planning: An Online, Open-source, Programming System IEEE Inter Conf on Robotics and Automation (ICRA), Rome, Italy, 2007, pp. 3711 3716 OOPS for Motion Planning: An Online, Open-source, Programming System Erion Plaku Kostas E. Bekris Lydia E. Kavraki Abstract

More information

D-Plan: Efficient Collision-Free Path Computation for Part Removal and Disassembly

D-Plan: Efficient Collision-Free Path Computation for Part Removal and Disassembly 774 Computer-Aided Design and Applications 2008 CAD Solutions, LLC http://www.cadanda.com D-Plan: Efficient Collision-Free Path Computation for Part Removal and Disassembly Liangjun Zhang 1, Xin Huang

More information

Useful Cycles in Probabilistic Roadmap Graphs

Useful Cycles in Probabilistic Roadmap Graphs Useful Cycles in Probabilistic Roadmap Graphs Dennis Nieuwenhuisen Mark H. Overmars institute of information and computing sciences, utrecht university technical report UU-CS-24-64 www.cs.uu.nl Useful

More information

Probabilistic Roadmaps of Trees for Parallel Computation of Multiple Query Roadmaps

Probabilistic Roadmaps of Trees for Parallel Computation of Multiple Query Roadmaps Probabilistic Roadmaps of Trees for Parallel Computation of Multiple Query Roadmaps Mert Akinc, Kostas E. Bekris, Brian Y. Chen, Andrew M. Ladd, Erion Plaku, and Lydia E. Kavraki Rice University Department

More information

for Motion Planning RSS Lecture 10 Prof. Seth Teller

for Motion Planning RSS Lecture 10 Prof. Seth Teller Configuration Space for Motion Planning RSS Lecture 10 Monday, 8 March 2010 Prof. Seth Teller Siegwart & Nourbahksh S 6.2 (Thanks to Nancy Amato, Rod Brooks, Vijay Kumar, and Daniela Rus for some of the

More information

Lucia K. Dale Guang Song Nancy M. Amato. Texas A&M University. January 16, Abstract

Lucia K. Dale Guang Song Nancy M. Amato. Texas A&M University. January 16, Abstract Faster, More Eective Connection for Probabilistic Roadmaps Lucia K. Dale Guang Song Nancy M. Amato dalel@cs.tamu.edu gsong@cs.tamu.edu amato@cs.tamu.edu Technical Report 00-005 Department of Computer Science

More information

Customizing PRM Roadmaps at Query Time Λ

Customizing PRM Roadmaps at Query Time Λ Customizing PRM Roadmaps at Query Time Λ Guang Song Shawna Miller Nancy M. Amato Department of Computer Science Texas A&M University College Station, TX 77843-3112 fgsong,slm5563,amatog@cs.tamu.edu Abstract

More information

cbook 26/2/ :49 PAGE PROOFS for John Wiley & Sons Ltd (jwcbook.sty v5.0, )

cbook 26/2/ :49 PAGE PROOFS for John Wiley & Sons Ltd (jwcbook.sty v5.0, ) Sample Contributed Book ii Sample Contributed Book Edited by J. SMITH JOHN WILEY & SONS Chichester. New York. Brisbane. Toronto. Singapore iv Contents vi CONTENTS 4 Probabilistic Roadmaps for Robot Path

More information

Artificial Potential Biased Probabilistic Roadmap Method

Artificial Potential Biased Probabilistic Roadmap Method Artificial Potential Biased Probabilistic Roadmap Method Daniel Aarno, Danica Kragic and Henrik I Christensen Centre for Autonomous Systems Royal Institute of Technology Stockholm, Sweden Email:bishop@kth.se,

More information

Motion Planning of Multiple Mobile Robots for Cooperative Manipulation and Transportation

Motion Planning of Multiple Mobile Robots for Cooperative Manipulation and Transportation IEEE TRANSACTIONS ON ROBOTICS AND AUTOMATION, VOL. 19, NO. 2, APRIL 2003 223 Motion Planning of Multiple Mobile Robots for Cooperative Manipulation and Transportation Atsushi Yamashita, Member, IEEE, Tamio

More information

Model-Based Motion Planning

Model-Based Motion Planning University of Massachusetts Amherst ScholarWorks@UMass Amherst Computer Science Department Faculty Publication Series Computer Science 2004 Model-Based Motion Planning Brendan Burns University of Massachusetts

More information

Full paper. Visibility-based probabilistic roadmaps for motion planning. T. SIMÉON, J.-P. LAUMOND and C. NISSOUX

Full paper. Visibility-based probabilistic roadmaps for motion planning. T. SIMÉON, J.-P. LAUMOND and C. NISSOUX Advanced Robotics, Vol. 14, No. 6, pp. 477 493 (2000) Ó VSP and Robotics Society of Japan 2000. Full paper Visibility-based probabilistic roadmaps for motion planning T. SIMÉON, J.-P. LAUMOND and C. NISSOUX

More information

Evaluation of the K-closest Neighbor Selection Strategy for PRM Construction

Evaluation of the K-closest Neighbor Selection Strategy for PRM Construction Evaluation of the K-closest Neighbor Selection Strategy for PRM Construction Troy McMahon, Sam Jacobs, Bryan Boyd, Lydia Tapia, Nancy M. Amato Parasol Laboratory, Dept. of Computer Science and Engineering,

More information

Path Planning. Jacky Baltes Dept. of Computer Science University of Manitoba 11/21/10

Path Planning. Jacky Baltes Dept. of Computer Science University of Manitoba   11/21/10 Path Planning Jacky Baltes Autonomous Agents Lab Department of Computer Science University of Manitoba Email: jacky@cs.umanitoba.ca http://www.cs.umanitoba.ca/~jacky Path Planning Jacky Baltes Dept. of

More information

Autonomous and Mobile Robotics Prof. Giuseppe Oriolo. Motion Planning 1 Retraction and Cell Decomposition

Autonomous and Mobile Robotics Prof. Giuseppe Oriolo. Motion Planning 1 Retraction and Cell Decomposition Autonomous and Mobile Robotics Prof. Giuseppe Oriolo Motion Planning 1 Retraction and Cell Decomposition motivation robots are expected to perform tasks in workspaces populated by obstacles autonomy requires

More information

Path Planning for a Robot Manipulator based on Probabilistic Roadmap and Reinforcement Learning

Path Planning for a Robot Manipulator based on Probabilistic Roadmap and Reinforcement Learning 674 International Journal Jung-Jun of Control, Park, Automation, Ji-Hun Kim, and and Systems, Jae-Bok vol. Song 5, no. 6, pp. 674-680, December 2007 Path Planning for a Robot Manipulator based on Probabilistic

More information

correspond to collision-free congurations of the robot. roadmap, and then nding a path in the roadmap between (repeat as desired) II.

correspond to collision-free congurations of the robot. roadmap, and then nding a path in the roadmap between (repeat as desired) II. Nancy M. Amato, Texas A&M University, College Station, TX, USA O. Burchan Bayazit, Texas A&M University, College Station, TX, USA Lucia K. Dale, Texas A&M University, College Station, TX, USA Christopher

More information

Path Planning in Repetitive Environments

Path Planning in Repetitive Environments MMAR 2006 12th IEEE International Conference on Methods and Models in Automation and Robotics 28-31 August 2006 Międzyzdroje, Poland Path Planning in Repetitive Environments Jur van den Berg Mark Overmars

More information

Introduction to State-of-the-art Motion Planning Algorithms. Presented by Konstantinos Tsianos

Introduction to State-of-the-art Motion Planning Algorithms. Presented by Konstantinos Tsianos Introduction to State-of-the-art Motion Planning Algorithms Presented by Konstantinos Tsianos Robots need to move! Motion Robot motion must be continuous Geometric constraints Dynamic constraints Safety

More information

A Hybrid Approach for Complete Motion Planning

A Hybrid Approach for Complete Motion Planning A Hybrid Approach for Complete Motion Planning Liangjun Zhang 1 Young J. Kim 2 Dinesh Manocha 1 1 Dept. of Computer Science, University of North Carolina at Chapel Hill, USA, {zlj,dm}@cs.unc.edu 2 Dept.

More information

Kinodynamic Motion Planning with Space-Time Exploration Guided Heuristic Search for Car-Like Robots in Dynamic Environments

Kinodynamic Motion Planning with Space-Time Exploration Guided Heuristic Search for Car-Like Robots in Dynamic Environments Kinodynamic Motion Planning with Space-Time Exploration Guided Heuristic Search for Car-Like Robots in Dynamic Environments Chao Chen 1 and Markus Rickert 1 and Alois Knoll 2 Abstract The Space Exploration

More information

Collided Path Replanning in Dynamic Environments Using RRT and Cell Decomposition Algorithms

Collided Path Replanning in Dynamic Environments Using RRT and Cell Decomposition Algorithms Collided Path Replanning in Dynamic Environments Using RRT and Cell Decomposition Algorithms Ahmad Abbadi ( ) and Vaclav Prenosil Department of Information Technologies, Faculty of Informatics, Masaryk

More information

Algorithms for Sensor-Based Robotics: Sampling-Based Motion Planning

Algorithms for Sensor-Based Robotics: Sampling-Based Motion Planning Algorithms for Sensor-Based Robotics: Sampling-Based Motion Planning Computer Science 336 http://www.cs.jhu.edu/~hager/teaching/cs336 Professor Hager http://www.cs.jhu.edu/~hager Recall Earlier Methods

More information

OPTIMIZATION TECHNIQUES FOR PROBABILISTIC ROADMAPS A Dissertation by LUCIA KATHRYN DALE Submitted to Texas A&M University in partial fulfillment of th

OPTIMIZATION TECHNIQUES FOR PROBABILISTIC ROADMAPS A Dissertation by LUCIA KATHRYN DALE Submitted to Texas A&M University in partial fulfillment of th OPTIMIZATION TECHNIQUES FOR PROBABILISTIC ROADMAPS A Dissertation by LUCIA KATHRYN DALE Submitted to the Office of Graduate Studies of Texas A&M University in partial fulfillment of the requirements for

More information

CS Path Planning

CS Path Planning Why Path Planning? CS 603 - Path Planning Roderic A. Grupen 4/13/15 Robotics 1 4/13/15 Robotics 2 Why Motion Planning? Origins of Motion Planning Virtual Prototyping! Character Animation! Structural Molecular

More information

Real-time Reach Planning for Animated Characters Using Hardware Acceleration

Real-time Reach Planning for Animated Characters Using Hardware Acceleration Real-time Reach Planning for Animated Characters Using Hardware Acceleration Ying Liu Norman I. Badler Center for Human Modeling and Simulation Department of Computer and Information Science University

More information

A Machine Learning Approach for Feature-Sensitive Motion Planning

A Machine Learning Approach for Feature-Sensitive Motion Planning A Machine Learning Approach for Feature-Sensitive Motion Planning Marco Morales, Lydia Tapia, Roger Pearce, Samuel Rodriguez, and Nancy M. Amato Parasol Laboratory, Dept. of Computer Science, Texas A&M

More information

Instant Prediction for Reactive Motions with Planning

Instant Prediction for Reactive Motions with Planning The 2009 IEEE/RSJ International Conference on Intelligent Robots and Systems October 11-15, 2009 St. Louis, USA Instant Prediction for Reactive Motions with Planning Hisashi Sugiura, Herbert Janßen, and

More information

1 Introduction. 2 Iterative-Deepening A* 3 Test domains

1 Introduction. 2 Iterative-Deepening A* 3 Test domains From: AAAI Technical Report SS-93-04. Compilation copyright 1993, AAAI (www.aaai.org). All rights reserved. Fast Information Distribution for Massively Parallel IDA* Search Diane J. Cook Department of

More information

video 1 video 2 Motion Planning (It s all in the discretization) Digital Actors Basic problem Basic problem Two Possible Discretizations

video 1 video 2 Motion Planning (It s all in the discretization) Digital Actors Basic problem Basic problem Two Possible Discretizations Motion Planning (It s all in the discretization) Motion planning is the ability for an agent to compute its own motions in order to achieve certain goals. ll autonomous robots and digital actors should

More information

Workspace Importance Sampling for Probabilistic Roadmap Planning

Workspace Importance Sampling for Probabilistic Roadmap Planning Workspace Importance Sampling for Probabilistic Roadmap Planning Hanna Kurniawati David Hsu Department of Computer Science National University of Singapore Singapore, Republic of Singapore {hannakur, dyhsu}@comp.nus.edu.sg

More information

Deformable Robot Motion Planning in a Reduced-Dimension Configuration Space

Deformable Robot Motion Planning in a Reduced-Dimension Configuration Space 2010 IEEE International Conference on Robotics and Automation Anchorage Convention District May 3-8, 2010, Anchorage, Alaska, USA Deformable Robot Motion Planning in a Reduced-Dimension Configuration Space

More information

Algorithms for Sensor-Based Robotics: Sampling-Based Motion Planning

Algorithms for Sensor-Based Robotics: Sampling-Based Motion Planning Algorithms for Sensor-Based Robotics: Sampling-Based Motion Planning Computer Science 336 http://www.cs.jhu.edu/~hager/teaching/cs336 Professor Hager http://www.cs.jhu.edu/~hager Recall Earlier Methods

More information

Path-Planning for Multiple Generic-Shaped Mobile Robots with MCA

Path-Planning for Multiple Generic-Shaped Mobile Robots with MCA Path-Planning for Multiple Generic-Shaped Mobile Robots with MCA Fabio M. Marchese and Marco Dal Negro Dipartimento di Informatica, Sistemistica e Comunicazione Università degli Studi di Milano - Bicocca

More information

Open Access Narrow Passage Watcher for Safe Motion Planning by Using Motion Trend Analysis of C-Obstacles

Open Access Narrow Passage Watcher for Safe Motion Planning by Using Motion Trend Analysis of C-Obstacles Send Orders for Reprints to reprints@benthamscience.ae 106 The Open Automation and Control Systems Journal, 2015, 7, 106-113 Open Access Narrow Passage Watcher for Safe Motion Planning by Using Motion

More information

Free-Form Shape Optimization using CAD Models

Free-Form Shape Optimization using CAD Models Free-Form Shape Optimization using CAD Models D. Baumgärtner 1, M. Breitenberger 1, K.-U. Bletzinger 1 1 Lehrstuhl für Statik, Technische Universität München (TUM), Arcisstraße 21, D-80333 München 1 Motivation

More information

EFFECT OF COOPERATIVE WORK IN OBJECT TRANSPORTATION BY MULTY-AGENT SYSTEMS IN KNOWN ENVIRONMENTS

EFFECT OF COOPERATIVE WORK IN OBJECT TRANSPORTATION BY MULTY-AGENT SYSTEMS IN KNOWN ENVIRONMENTS Proceedings of MUSME 2011, the International Symposium on Multibody Systems and Mechatronics Valencia, Spain, 25-28 October 2011 EFFECT OF COOPERATIVE WORK IN OBJECT TRANSPORTATION BY MULTY-AGENT SYSTEMS

More information

Sampling Techniques for Probabilistic Roadmap Planners

Sampling Techniques for Probabilistic Roadmap Planners Sampling Techniques for Probabilistic Roadmap Planners Roland Geraerts Mark H. Overmars Institute of Information and Computing Sciences Utrecht University, the Netherlands Email: {roland,markov}@cs.uu.nl.

More information

Learning Humanoid Reaching Tasks in Dynamic Environments

Learning Humanoid Reaching Tasks in Dynamic Environments Proceedings of the 2007 IEEE/RSJ International Conference on Intelligent Robots and Systems San Diego, CA, USA, Oct 29 - Nov 2, 2007 TuD3.5 Learning Humanoid Reaching Tasks in Dynamic Environments Xiaoxi

More information