Automatic motion planning has applications ranging from traditional robotics to computer-aided design to computational biology and chemistry. Randomized planners, such as probabilistic roadmap methods (prms), have been highly successful in solving these high degree of freedom problems. However, the traditional prm framework fails to address several practical issues. One of the most important issues is the difficulty of deciding what size roadmap is required to solve a given problem efficiently. prms do not provide an automated way to determine appropriate roadmap size. In this paper, we propose a new prm-based framework called Incremental Map Generation (img) to address this problem. Our strategy is to break the map generation into independent processes. Each process generates an independent roadmap component. img proceeds by adding independent roadmap components to an existing roadmap until some user defined criteria are satisfied. In addition to addressing the roadmap size problem, this framework supports roadmap reproducibility in that any of the roadmap increments can be reproduced by using the same set of seeds. Finally, these independent processes Automatic motion planning has applications in many areas such as robotics [26], computer animation, computeraided design/virtual prototyping, and computational biology and chemistry. Although many deterministic motion

We show that paths for dexterous robots can be generated in a few seconds or less using parallel informed randomized search on multicomputers. The experimental results we present have been obtained for a simulated 7-jointed arm operating in realistic 3D workspaces. We also present a new method for predicting the solution times that our parallel formulation can deliver on increasing numbers of processors. 1 Introduction Motion planning is the process of computing paths that will allow a robot to move to different positions in its environment without hitting obstacles. Many motion planning algorithms have been developed [13], but most are never used in practice because of their computational complexity [8]. The intent of this paper is to show that plans for multi-jointed dexterous robot arms which operate in realistic environments can be synthesized very quickly by parallel algorithms, and to show how to estimate the number of processors necessary to obtain results in an acceptable ti...

Recently, a new class of randomized path planning methods, known as Probabilistic Roadmap Methods (PRMs) have shown great potential for solving complicated high-dimensional problems. PRMs use randomization (usually during preprocessing) to construct a graph (a roadmap) of representative paths in the robot's con guration space. Vertices correspond to collision-free con gurations of the robot. An edge exists between two vertices if a path between the two corresponding con gurations can be found by a local planning method. PRMs solve many high degree of freedom (dof) motion planning problems. Unfortunately, for some problems running times may still be unacceptably large and solutions sub-optimal. We provide speed and quality optimization strategies applicable in cluttered 3-dimensional workspaces....

Abstract We present parallel algorithms to accelerate collision queries for samplebased motion planning. Our approach is designed for current many-core GPUs and exploits the data-parallelism and multi-threaded capabilities. In order to take advantage of high number of cores, we present a clustering scheme and collision-packet traversal to perform efficient collision queries on multiple configurations simultaneously. Furthermore, we present a hierarchical traversal scheme that performs workload balancing for high parallel efficiency. We have implemented our algorithms on commodity NVIDIA GPUs using CUDA and can perform 500,000 collision queries/second on our benchmarks, which is 10X faster than prior GPU-based techniques. Moreover, we can compute collision-free paths for rigid and articulated models in less than 100 milliseconds for many benchmarks, almost 50-100X faster than current CPU-based planners. 1

We show how paths for dexterous robots can be generated in a few seconds or less using parallel informed randomized search on multicomputers. The experimental results we present have been obtained for a variety of robots with six or more joints operating in realistic 3D workspaces.

In a constantly changing and partially unpredictable environment, robot motion planning must be on-line. The planner then receives a continuous #ow of information about occurring events and generates motoin plans. While planned motions are being executed, new plans are generated in response to incoming events. This thesis describes an on-line planner for two cooperating arms whose task is to grab parts of various types on a conveyor belt and transfer them to their respective goals while avoiding collision with obstacles. Parts arrive on the belt in random order, at any time. Both goals and obstacles may be dynamically changed. This scenario is typical of manufacturing cells serving machine-tools, assembling products, or packaging objects. The proposed approach breaks the overall planning problem into subproblems, eachinvolving a low-dimension con#guration or con#guration#time space, and orchestrates very fast primitives solving these subproblems. This thesis describes this approach and...

In this paper we present an analysis that indentifies opportunities of introducing parallelism in the calculation of the configuration space. When the Convolution Theorem is used for this evaluation, two levels have appeared within the basic operation of the algorithms and another one for the employed computational tool (Fast Fourier Transform, FFT).

the study of robotic tracking and catching operations, ” in Proc. ASME Compuf. in Eng., 1992, vol. 1, pp. 141-146. [ 1 I] _ _,“Tracking a moving object with a 6-DOF manipulator, ” in Proc.