Showing papers by "Nancy M. Amato published in 2004"

Approximate convex decomposition of polygons

Abstract: We propose a strategy to decompose a polygon, containing zero or more holes, into ``approximately convex'' pieces. For many applications, the approximately convex components of this decomposition provide similar benefits as convex components, while the resulting decomposition is significantly smaller and can be computed more efficiently. Moreover, our approximate convex decomposition (ACD) provides a mechanism to focus on key structural features and ignoreless significant artifacts such as wrinkles and surface texture a user specified tolerance determines allowable concavity. We propose a simple algorithm that computes an ACD of a polygon by iteratively removing (resolving) the most significant non-convex feature (notch). As a by product, it produces an elegant hierarchical representation that provides a series of `increasingly convex' decompositions. Our algorithm computes an ACD of a simple polygon with n verticesand r notches in O(nr) time. In contrast, exact convex decomposition is NP-hard or,if the polygon has no holes, takes O(nr2) time.

Shepherding behaviors

Abstract: Shepherding behaviors are a type of flocking behavior in which outside agents guide or control members of a flock. Shepherding behaviors can be found in various forms in nature. For example, herding, covering, patrolling and collecting are common types of shepherding behaviors. In this work, we investigate ways to simulate these types of behaviors. A shepherd uses roadmaps to steer the flock and to re-group separated flock members. This paper focuses on improving the shepherd's movements to gain better control of the flock's motion and use this improved control to demonstrate a wider variety of shepherding behaviors.

A Machine Learning Approach for Feature-Sensitive Motion Planning.

Abstract: Although there are many motion planning techniques, there is no method that outperforms all others for all problem instances. Rather, each technique has different strengths and weaknesses which makes it best-suited for certain types of problems. Moreover, since an environment can contain vastly different regions, there may not be a single planner that will perform well in all its regions. Ideally, one would use a suite of planners in concert and would solve the problem by applying the best-suited planner in each region. In this paper, we propose an automated framework for feature-sensitive motion planning. We use a machine learning approach to characterize and partition C-space into regions that are well suited to one of the methods in our library of roadmapbased motion planners. After the best-suited method is applied in each region, the resulting region roadmaps are combined to form a roadmap of the entire planning space. Over a range of problems, we demonstrate that our simple prototype system reliably outperforms any of the planners on their own.

Distributed reconfiguration of metamorphic robot chains

Abstract: The problem we address is the distributed reconfiguration of a planar metamorphic robotic system composed of any number of hexagonal modules. After presenting a framework for classifying motion planning algorithms for metamorphic robotic systems, we describe distributed algorithms for reconfiguring a straight chain of hexagonal modules to any intersecting straight chain configuration. We prove our algorithms are correct, and show that they are either optimal or asymptotically optimal in the number of moves and asymptotically optimal in the time required for parallel reconfiguration.

Approximate convex decomposition of polyhedra

Abstract: Decomposition is a technique commonly used to partition complex models into simpler components. While decomposition into convex components results in pieces that are easy to process, such decompositions can be costly to construct and can result in representations with an unmanageable number of components. In this paper we explore an alternative partitioning strategy that decomposes a given model into “approximately convex” pieces that may provide similar benefits as convex components, while the resulting decomposition is both significantly smaller (typically by orders of magnitude) and can be computed more efficiently. Indeed, for many applications, an approximate convex decomposition (ACD) can more accurately represent the important structural features of the model by providing a mechanism for ignoring less significant features, such as surface texture. We describe a technique for computing ACDs of threedimensional polyhedral solids and surfaces of arbitrary genus. We provide results illustrating that our approach results in high quality decompositions with very few components and applications showing that comparable or better results can be obtained using ACD decompositions in place of exact convex decompositions (ECD) that are several orders of magnitude larger. CR Categories: I.3.5 [COMPUTER GRAPHICS]: Computational Geometry and Object Modeling—Geometric algorithms, languages, and systems

Approximate convex decomposition

Abstract: Decomposition is a technique commonly used to break complex models into sub-models that are easier to handle. Convex decomposition, which partitions the model into convex components, is interesting because many algorithms perform more efficiently on convex objects than on non-convex objects. One issue with convex decompositions, however, is that they can be costly to construct and can result in representations with an unmanageable number of components. In many applications, the detailed features of the model are not crucial and in fact considering them only serves to obscure important structural features and adds to the processing cost. In such cases, an approximate representation of the model that captures the key structural features would be preferable.

A kinematics-based probabilistic roadmap method for high DOF closed chain systems

Abstract: We consider the motion planning problem for arbitrary articulated structures with one or more closed kinematic chains in a workspace with obstacles. This is an important class of problems and there are applications in many areas such as robotics, closed molecular chains, graphical animation, and reconfigurable robots. We use the kinematics-based probabilistic roadmap (KBPRM) strategy proposed in [Han, L and Amato, NM (March 2000)] that conceptually partitions the linkage into a set of open chains and applies random generation methods to some of the chains and traditional inverse kinematics methods to the others. The efficiency of the method depends critically on how the linkage is partitioned into open chains. The original method assumed the partition was provided as input to the problem. We propose a fully automated method for partitioning an arbitrary linkage into open chains and for determining which should be positioned using the inverse kinematic solver. Even so, the size (number of links) of the closed loops that can be handled by this method is limited because the inverse solver can only be applied to small chains. To handle high dof closed loops, we show how we can use the iterative relaxation of constraints (IRC) strategy proposed by Bayazit to efficiently handle large loops while still only using inverse kinematics for small chains. Our results in 3-dimensional workspaces both for planar and spatial linkages show that our framework performs well for general linkages. We also use our planner to simulate an adjustable lamp called Luxo. Using IRC, our planner can handle a single loop of up to 98 links.

Using motion planning to study RNA folding kinetics

Abstract: We propose a novel, motion planning based approach to approximately map the energy landscape of an RNA molecule. Our method is based on the successful probabilistic roadmap motion planners that we have previously successfully applied to protein folding. The key advantage of our method is that it provides a sparse map that captures the main features of the landscape and which can be analyzed to compute folding kinetics. In this paper, we provide evidence that this approach is also well suited to RNA. We compute population kinetics and transition rates on our roadmaps using the master equation for a few moderately sized RNA and show that our results compare favorably with results of other existing methods.

Better Group Behaviors Using Rule-Based Roadmaps

Abstract: While techniques exist for simulating group behaviors, these methods usually only provide simplistic navigation and planning capabilities. In this work, we explore the benefits of integrating roadmap-based path planning methods with flocking techniques. We show how group behaviors such as exploring can be facilitated by using dynamic roadmaps (e.g., modifying edge weights) as an implicit means of communication between flock members. Extending ideas from cognitive modeling, we embed behavior rules in individual flock members and in the roadmap. These behavior rules enable the flock members to modify their actions based on their current location and state. We propose new techniques for three distinct group behaviors: homing, exploring (covering and goal searching) and passing through narrow areas. Animations of these behaviors can be viewed at http://parasol.tamu.edu/dsmft.

Swarming behavior using probabilistic roadmap techniques

Abstract: While techniques exist for simulating swarming behaviors, these methods usually provide only simplistic navigation and planning capabilities. In this review, we explore the benefits of integrating roadmap-based path planning methods with flocking techniques to achieve different behaviors. We show how group behaviors such as exploring can be facilitated by using dynamic roadmaps (e.g., modifying edge weights) as an implicit means of communication between flock members. Extending ideas from cognitive modeling, we embed behavior rules in individual flock members and in the roadmap. These behavior rules enable the flock members to modify their actions based on their current location and state. We propose new techniques for several distinct group behaviors: homing, exploring (covering and goal searching), passing through narrow areas and shepherding. We present results that show that our methods provide significant improvement over methods that utilize purely local knowledge and moreover, that we achieve performance approaching that which could be obtained by an ideal method that has complete global knowledge. Animations of these behaviors can be viewed on our webpages.

A generalized framework for interactive dynamic simulation for multirigid bodies

Abstract: This paper presents a generalized framework for dynamic simulation realized in a prototype simulator called the Interactive Generalized Motion Simulator (I-GMS), which can simulate motions of multirigid-body systems with contact interaction in virtual environments. I-GMS is designed to meet two important goals: generality and interactivity. By generality, we mean a dynamic simulator which can easily support various systems of rigid bodies, ranging from a single free-flying rigid object to complex linkages such as those needed for robotic systems or human body simulation. To provide this generality, we have developed I-GMS in an object-oriented framework. The user interactivity is supported through a haptic interface for articulated bodies, introducing interactive dynamic simulation schemes. This user-interaction is achieved by performing push and pull operations via the PHANToM haptic device, which runs as an integrated part of I-GMS. Also, a hybrid scheme was used for simulating internal contacts (between bodies in the multirigid-body system) in the presence of friction, which could avoid the nonexistent solution problem often faced when solving contact problems with Coulomb friction. In our hybrid scheme, two impulse-based methods are exploited so that different methods are applied adaptively, depending on whether the current contact situation is characterized as "bouncing" or "steady." We demonstrate the user-interaction capability of I-GMS through online editing of trajectories of a 6-degree of freedom (dof) articulated structure.

Parallel protein folding with STAPL

Abstract: Summary form only given. The protein folding problem is to study how a protein dynamically folds to its so-called native state - an energetically stable, three-dimensional configuration. Understanding this process is of great practical importance since some devastating diseases such as Alzheimer's and bovine spongiform encephalopathy (Mad Cow) are associated with the misfolding of proteins. In our group, we have developed a new computational technique for studying protein folding that is based on probabilistic roadmap methods for motion planning. Our technique yields an approximate map of a protein's potential energy landscape that contains thousands of feasible folding pathways. We have validated our method against known experimental results. Other simulation techniques, such as molecular dynamics or Monte Carlo methods, require many orders of magnitude more time to produce a single, partial, trajectory. We report on our experiences parallelizing our method using STAPL (the standard template adaptive parallel library), that is being developed in the Parasol Lab at Texas A&M. An efficient parallel version enables us to study larger proteins with increased accuracy. We demonstrate how STAPL enables portable efficiency across multiple platforms without user code modification. We show performance gains on two systems: a dedicated Linux cluster and an extremely heterogeneous multiuser Linux cluster.

Enveloping multi-pocket obstacles with hexagonal metamorphic robots

Abstract: The problem addressed is reconfiguration planning for a metamorphic robotic system composed of any number of hexagonal robots when a single obstacle with multiple indentations or "pockets" is embedded in the goal environment. We extend our earlier work on filling a single pocket in an obstacle to the case where the obstacle surface may contain multiple pockets. The planning phase of our algorithm first determines whether the obstacle pockets provide sufficient clearance for module movement, i.e., whether the obstacle is "admissible". In this paper, we present algorithms that sequentially order individual pockets and order module placement inside each pocket. These algorithms ensure that every cell in each pocket is filled and that module deadlock and collision do not occur during reconfiguration. This paper also provides a complete overview of the planning stage that is executed prior to reconfiguration and presents a distributed reconfiguration schema for filling more than one obstacle pocket concurrently, followed by the envelopment of the entire obstacle. Lastly, we present examples of obstacles with multiple pockets that were successfully filled using our distributed reconfiguration simulator.

Complexity analysis and approximate solutions for two multiple-robot localization problems

Abstract: We consider the localization problem for a system of mobile robots using inexpensive range sensors. Among many issues for multi-robot systems, two problems are identified and formally defined. The first problem is sensing ranges from all robots as quickly as possible while avoiding sensor cross-talk, and the second problem is to localize a multi-robot system using a minimal number of range sensings. We show that both these problems are NP-complete, and we propose an approximate method for the multi-robot localization problem that takes advantage of the robots pose uncertainty information. Simulation results show the effectiveness of our method for localizing multiple robots.

A framework for roadmap-based navigation and sector-based localization of mobile robots

Abstract: Personal robotics applications require autonomous mobile robot navigation methods that are safe, robust, and inexpensive. Two requirements for autonomous use of robots for such applications are an automatic motion planner to select paths and a robust way of ensuring that the robot can follow the selected path given the unavoidable odometer and control errors that must be dealt with for any inexpensive robot. Additional difficulties are faced when there is more than one robot involved. In this dissertation, we describe a new roadmap-based method for mobile robot navigation. It is suitable for partially known indoor environments and requires only inexpensive range sensors. The navigator selects paths from the roadmap and designates localization points on those paths. In particular, the navigator selects feasible paths that are sensitive to the needs of the application (e.g., no sharp turns) and of the localization algorithm (e.g., within sensing range of two features). We present a new sector-based localizer that is robust in the presence of sensor limitations and unknown obstacles while still maintaining computational efficiency. We extend our approach to teams of robots focusing on quickly sensing ranges from all robots while avoiding sensor cross-talk, and reducing the pose uncertainties of all robots while using a minimal number of sensing rounds. We present experimental results for mobile robots and describe a web-based route planner for the Texas A&M campus that utilizes our navigator.