ÜberSim: High-Fidelity Multi-Robot Simulation for Robot Soccer

Size: px
Start display at page:

Download "ÜberSim: High-Fidelity Multi-Robot Simulation for Robot Soccer"

Transcription

1 ÜberSim: High-Fidelity Multi-Robot Simulation for Robot Soccer Content Areas: common sense reasoning, knowledge representation Abstract High fidelity simulation can be a powerful tool for speeding up the robot control development cycle. For RoboCup robot soccer, a dynamic adversarial game where robot teams play soccer against one another, this is especially true. Developing control software for a robot soccer team is a difficult task, one that gains substantially from simulation. As yet, however, there are no publicly available simulators that are realistic or configurable enough to be of real use to a robot soccer team. In this paper, we describe ÜberSim, a new simulation engine built around a high-fidelity physics simulation engine that attempts to address this lack for the small size league, a division within RoboCup. The goal of ÜberSim is to provide a simulation engine that is extensible, re-configurable, and realistic enough to support easy transfer of robot control software directly from simulation to reality with a minimum of fuss. Here, we describe the implementation of ÜberSim for the small-size league, the design choices made and their reasoning. 1 Introduction Robot soccer is an international initiative developed to push the state-of-the-art in autonomous, multi-robot intelligence through autonomous robot teams playing games of soccer against one another. To develop a successful team, a wide range of challenges must be overcome. In particular, the adversarial nature of the task creates significant levels of uncertainty that must be dealt with by these control systems, while its fast-paced nature pushes robot control techniques to their limits and beyond. Without a doubt, developing robot soccer control systems is a difficult task, but it is a task that can and does benefit from the use of good simulation engines. Many teams (e.g.???) have made good use of simulators to enable them to greatly speed up their software development process. Developing software directly on robots is often a tedious and difficult process. Limited communications bandwidth often limits one s ability to gain insightful information during the debugging process. Likewise, robot experiments are limited in duration due to the need to regularly change batteries. None of these difficulties arise in simulation, where indeed the presence of the robots, which are often a contentious resource, is not even required. A simulator is useful if it speeds up development. Hence, a useful simulator is one that facilitates the transfer of simulation developed control to the real system with a minimum of fuss. The key here lies in capturing the important environment dynamics at reasonable computation speeds. Robot soccer, due to its fast pace, places greater requirements on accurate dynamics than does many other multi-robot research domains. Unfortunately, most freely available multi-robot simulators use lowfidelity dynamics models and rarely consider collisions making them unsuitable for this task. As a result researchers often develop custom simulators to fill this void. This development, however, is non-trivial and more than likely a low reward to cost venture. Moreover, if the simulator is not extensible it will most likely require substantial modifications as the robot hardware changes, which it often does from year to year. To address this issue, we have attempted to address this issue by developing a high-fidelity simulation engine, called ÜberSim. ÜberSim is extensible, reconfigurable, and publicly available under the Open Source LGPL license at [UberSim]. Specifically, Über- Sim has been developed for the small size league (SSL), a division of RoboCup robot soccer where robots are restricted in size and teams are allowed to use overhead cameras for global perception and off-field processing [Kitano et al., 98]. In this paper, we describe the development of ÜberSim, the design choices made, and the resulting system. The following section overviews the robot soccer, and the SSL with an eye to specifying the requirements for the simulator. Section 3 describes the details of the ÜberSim implementation. Section 4 presents some empirical evaluations to validate its performance, with Section 5 describing some related approaches. Finally, section 6 concludes the paper.

2 2 Overview In this section we briefly overview robot soccer and the small size league (SSL) with a view to defining the requirements for the ÜberSim simulator. 2.1 Robot Soccer RoboCup robot soccer [Kitano et al., 98] is an adversarial multi-robot research domain where autonomous teams of robots compete against one another in a game of soccer. The domain has a broad range of leagues, however, in the context of this paper our focus is on the small size league (SSL). A SSL game consists of two teams of five robots playing soccer with an orange golf ball on a 2.8m x 2.3m carpeted field, surrounded by short angled walls. Figure??? shows a picture of a game. Each robot is restricted in size, shape, and its use of ball manipulation devices. Robots must fit within an 18cm x 15cm cylinder and must not enclose the ball by more than 1cm. Ball manipulation devices are restricted, but many teams use dribbling and kicking mechanisms. A dribbler is a rubberized bar that rotates against the ball causing it to spin backwards against the robot thereby aiding manipulation, while a kicker is a linear or rotational actuator driving a metal bar. Given these restrictions, there is still much diversity in robot bases ranging in size and shape, speed and acceleration capabilities, ball manipulation mechanisms, through to the number of wheels, their type, and physical arrangement. Teams include differentially driven robots (diffbots), omni-directional robots (omnibots) and even four wheeled omni-directional robots. The SSL is further characterized by the allowance of off-field computers for processing and overhead cameras for global vision perception purposes. In this common case, the team as a whole is autonomous rather than each individual robot. Figure 1 shows the common system set up where global vision and control is processed off field and low-level velocity commands are sent to the individual robots with their own local servo control loops. Using this set up, many teams produce robot systems that are capable of high speed motions, approaching 2m/s, with ball speeds of 2m/s and even up to 5m/s! With such fast moving objects, accurate control is of paramount importance and this places many constraints on the accuracy of the simulation engine. 2.2 ÜberSim Requirements To be a useful SSL simulator, ÜberSim must provide identical software interfaces control software as well as simulating the physics of the robots and ball moving through the world. For SSL, or indeed any robot system, three interfaces are required: Perception, action, and a control interface for controlling the set up and execution of the simulation. The unique use of global vision and off field processing in the SSL creates a natural level to provide the perception and action interfaces. All teams that use global perception first process vision to produce estimates of where each object is on the field, potentially their orientation and velocities. Indeed, one of the current ideas within the league is to move to a unified vision server that produces this pose information for the teams to process. Likewise, at the opposite end of the spectrum, many teams make use of low-level control algorithms running on the robots proper to essentially create a remote controlled vehicle where velocity (or perhaps position) commands are sent to control it (see Figure???). These two facts conspire to make the output of the vision processing module and the input to the radio module the ideal location to simulate the interfaces to the real world. Overhead camera Figure 1. A typical small-size robot soccer game set up. The exact parameterization of each interface is dependent upon the system in question. In general, however, it will be a fairly simple remapping of the parameters (or ignoring some) and is therefore ignored for the remainder of this work. Small-Size System wireless <x,y,, conf> x 30Hz ~100ms latency <v x, v y, > x 30Hz ÜberSim Off field PC Figure 2. Perception and Action interfaces for ÜberSim. ÜberSim must also simulate the dynamical properties of each interface. Indeed, for the SSL where One of the keys to developing good SSL control systems is to account for the system latency and limited frame rate. Typical latencies are on the order of 100ms, which has substantial impact on a robot moving at 1 to 2m/s over such a small field. Consequently, the simulation engine must incorporate a latency mechanism in order for control performance to be similar between simulation and reality. Unless specialized hardware is used, frame rates are limited to 25, 30, 50, or 60Hz depending upon the type of camera used and whether the full interlaced frame rate is used or not. As with latency, accurately simulating this limited frame rate is critical to ensuring the transferability of software from simulation to reality.

3 3 ÜBERSIM IMPLEMENTATION We now describe the specifics of the ÜberSim implementation for the SSL beginning with an overview of the ÜberSim approach. Throughout we emphasize the need to produce accurate but fast simulations and the need for extensible and easily reconfigurable simulation. 3.1 Simulator Overview With ÜberSim, we take a distinctly different to most robot simulation engines. Rather than developing our own custom rigid body dynamics and collision software, which is the approach used by most researchers, we have taken the approach of using a publicly available Open Source Software (OSS) library. Our motivation for this approach is simple. A mature OSS library with a community of users can be considered proven software, relatively bug free, and both functional and capable. Additionally, if the project is supported, new functionality will get added while existing bugs will get fixed. The alternative, of developing custom software, is a nontrivial task as evidenced by the surprisingly few OSS rigid body simulation libraries available. Of the mature OSS libraries, of which there are few, the Open Dynamics Engine (ODE) [ODE] meets the requirements and has a community of users that continues to grow in size. We describe further the relevant details of the simulator in the following sections. To achieve extensibility, we have devised object class structure combined with text-based configuration files. For creating a new robot, or some other object type, one defines the constructor to create and fill in the data structures that define the robot and a handle_event method to translate robot commands into force or speed settings for the robot s actuators. The simulator supports text-based configuration files for generically setting robot defined vector-based parameters thereby providing easy reconfiguration of robot parameters for a given robot class. Collision Manager ODE Dynamics, Collisions, Graphics Field Ball Diff Robot Scene Graph Figure 2. ÜberSim system components. Omni Robot Main Controller IFace Vision Radio Control Figure?? shows the major ÜberSim components. The primary data structure is the scene graph (SG), actually it is a tree but we use its more commonly accepted name. The SG maintains the position, dynamics, and collision information for all the physical objects in the environment. The collision manager interacts between ODE and the SG to implement collision detection routines and update the appropriate data structures. The Interface module provides the socket interface to the user software (or visualization tools). Finally, the main controller runs the show. In the following sections we describe these modules in greater detail. 3.2 Open Dynamics Engine The physics engine is the key to making ÜberSim useful in terms of high performance robot control. To be suitable, the physics engine must provide accurate rigid body dynamics, a suitable range of joint types, a suitable range of actuators, elastic contact collisions between rigid body objects, finite static and dynamic friction between touching surfaces and the transition between friction modes. As mentioned above, there are few OSS rigid body simulation engines that meet these criteria. Although there a number of free body simulators, most of these are intended for fast pseudo realistic simulation for computer games. Although visually realistic, accurate collision dynamics and finite friction limits are often ignored. For the task at hand, we need both accuracy and sufficient speed to operate on a standard PC in real time or faster. For ÜberSim we chose the Open Dynamics Engine (ODE) (version was used for this paper). ODE is a free rigid body simulator, developed by Russell Smith, and is available at [ODE]. It has been used in a number of other projects (see the site for details), is reasonably documented, and has reached a maturity level ensuring that the code is stable and useable. ODE provides support for rigid body motion with non-zero mass and rotational inertia, and non-uniform mass distributions. The simulator uses Euler integration with a settable simulation step size. Euler integration is accurate, fast, and relatively stable. To aid in simulation stability, a restitution force is allowed which forces wayward joints back into position. This is required as integration errors invariably accumulate. For a simulation such as ÜberSim, trading a little accuracy for long-term stability is necessary feature. For a number of simple object types (rectangular prisms, spheres and cylinders), ODE provides collision detection primitives. Most importantly, ODE simulates both static and dynamic friction for contacting surfaces with bounded stiction limits. Further, the frictional forces can be directionally dependent, which is critical for omni-directional wheels where the wheels free roll in one direction and have traction in the other. A number of joint types and actuator types are supported. Although the list of types is small, they cover the full range of joints and actuators found in robot soccer including the more complex Sony AIBO and humanoid robots. As a bonus feature, ODE includes OpenGL routines to render the 3D simulated environment thereby providing support for useful visualization tools. Put together, ODE meets our requirements for the simulation engine. Its main limitation stems from its collision detection package, which only provides a small set of primitive object shapes rather than a general polygon soup. Within ODE, objects consist of rigid bodies, geometries, and joints. Rigid bodies are the dynamical objects (with mass etc) and are used in the integration process. In

4 contrast, geometries and joints are not dynamical objects. Geometries are used to determine when and where collisions occur and what resultant forces are transferred to the connected rigid bodies. Joints are special geometries used to specify how two connected rigid bodies can move in relation to one another. Joints can also be powered, which provides the primary mechanism for controlling how the connected body behaves. 3.3 Scene Graph The main data structure used in ÜberSim is the Scene Graph (SG). The SG maintains all the information required by ODE to carry out integration steps, and all the information necessary to perform collision detection. The SG is actually a tree, although we use its more common name, that stores each object type in a hierarchical relationship. Figure 3 shows a simple SG. The SG uses the same object types identified by ODE: rigid bodies, geometries, joints, and contact surfaces. Geometries consist of one of the primitive shapes supported by ODE namely rectangular prisms, spheres, cylinders. Likewise, joints map directly to the available joint types found in ODE. root sphere ball rigid shell Robot1 wheel wheel Robot2 cylinder Figure 3. A simple scene graph. Rigid bodies are boxes, nodes are ovals, and hexagons are geometries. In addition to encapsulating all the physical parameters required by ÜberSim, it significantly simplifies the interaction between ÜberSim and ODE. In short, with the exception of the collision manager, all other modules interact with the simulation by manipulating or reading the SG data structure. We discuss this in further detail in section 3.6 for the main control loop. 3.4 Collision Detection ODE provides only primitive collision handling support. The Collision Manager (CM) interacts with ODE and the SG to detect and handle collisions. Collision detection, even for the primitive objects supported by ODE, is known to be processor intensive. We can avoid much wasted computation by recognizing that most objects in the environment are not colliding with one another. Further, the hierarchical nature of the SG enables us to make a significant improvement by using hierarchical bounding boxes, or in this case bounding spheres. This approach, which is common in collision detection packages (e.g. [Reggianni, 02]. Each node within the SG contains a bounding sphere that encapsulates the geometries at that node and all of its children. Very fast collision detection can then be performed between each bounding sphere within the SG. Collision detection can be performed by walking through the tree considering each node and checking for bounding sphere overlaps between the node and all nodes that have not been checked and are not a child or ancestor of the current node. Further optimizations are possible by realizing that if two nodes do not overlap, then none of their children overlap and so do not need to be compared. Once potential collisions have been winnowed, the more expensive ODE collision detection operations are applied. For any contacting surfaces, ODE generates the appropriate restoring forces. Depending upon the surfaces in question, these restoration forces may include friction components the parameters of which are stored in the SG as surface properties. 3.5 Object Classes Each object type in the environment is encapsulated into a class and has an associated text-based configuration file defining its parameter values. RigidBodyEntity is the base object and it encapsulates the notion of a rigid body with geometry providing base methods for adding objects to the scene graph, accessing its position, and for handling manipulation commands from ÜberSim. All other objects, the field, ball, and robot classes, are derived from RigidBodyEntity and therefore inherit its basic capabilities. As a result, each more complex object need only define a constructor to create its relevant SG objects, and a handle-event method to respond to ÜberSim generated events. The range of events include position events used to manually move objects around, command events generated when robot control programs send commands to a robot and query events that respond with the object position, orientation, and velocity in the world. The latter is used by ÜberSim to generate its perceptual output to the controlling software. The field and ball, which are trivial, are derived from the RibidBodyEntity class. Both classes have a constructor for creating the necessary geometries and contact surfaces and adding them to the SG. The parameters for these values are derived from the text-based configuration files. These text-based files are much like script files and allow assignment of a known parameter name to a vector list of values, which may be Boolean, strings, integers, or floats. The ball object additionally has a single rigid body object to define its mass and rotational inertia. No handle-event is required for the field (as it never moves), while the ball must only handles position events. One of the core challenges in the ÜberSim concept is the problem of how to enable new robot configurations to be easily added and/or parameters changed. In terms of robot hardware and capabilities, RoboCup is a very dynamic domain. Robots are often completely rebuilt from one year to the next. Hardware innovations may result in completely new drive configurations, new complex ball

5 manipulation devices, and at the very least substantial changes in the parameters that describe the physical characteristics of a robot. For the ÜberSim concept to be useful to a team, or to be capable of providing a simulation environment for comparing two teams, there must be mechanisms to add new robot types easily. Moreover, given the argued need for simulation accuracy, there must be a mechanism to easily adapt robot parameters to closely match those of the physical robot. To achieve this goal, ÜberSim uses multiple robot objects where each object encapsulates a particular robot type. Each robot class is derived from the usual Rigid- BodyEntity class giving it the base abilities for manipulation. Each robot object encapsulates a parameterized robot configuration, where the parameters describe the physical characteristics of a robot. In the current implementation of ÜberSim, two robot objects are defined: OmniRobot and DiffRobot. Each robot type, with appropriate parameters, can represent nearly all the robots found in small-size RoboCup competitions. The two types are distinguishable based on the drive configuration where the DiffRobot uses two wheels, while the Omni- Robot uses three specialized omni wheels. Each robot has an optional dribbler and kicker mechanism. The robot types define a generic differential or omni directional robot base with a kicker and dribbler (if selected). The particular parameters that define the dimensions and positioning of each part, the physical properties of the major robot components (mass, inertia etc.) and their contact parameters are read from human readable/editable ASCII text configuration files. 3.6 Putting it all together The main controller loop consists of taking the SG at the beginning of each simulation step and generating a list of rigid bodies. This list is sent to ODE to integrate ahead one time step. The SG is then updated with results of the integration. Collision detection, discussed next, is then performed and the appropriate actions taken. The main controller then extracts from the SG the positions of the objects in the world via the object classes, discussed below, and generates the appropriate perceptual output to the robot control software. Any commands sent by the control software are transformed appropriately and sent to the joint objects in the SG to be taken into account during the next simulation step. 4 RESULTS & DISCUSSION ÜberSim is able to simulate two teams of 5 robots and the golf ball playing soccer. With two simulation steps per frame of vision, running on a 1GHz Pentium processor, the simulator is able to execute at full frame rate (30Hz output, 60Hz simulation steps) using around 60% of the processor for the two full teams. To examine the performance of the system, we have compared the trajectories generated by ÜberSim for a robot performing a set pattern and performing acceleration tests. The robots were controlled using our CMDragons software [CMDragons]. The set pattern is a fixed sequence of target points fed into the navigation/motion control system where the target speed at the destination is set to zero. For details on the navigation and motion control see [Bruce et al., 02]. Figure 4 shows the acceleration tests at 2 m.s -2 for a differential drive robot in simulation and reality. The trials were repeated 5 times. Note that there will be some variation in the observed trajectory as the real system has noise and natural variation in the robot parameters. Figure 5 shows trajectory comparisons for a differential drive driving in a figure 8 at around 0.8 m.s -1 for a simulated robot versus a real one. In terms of running a team, simulation accuracy is appears visually reasonable, however, it is difficult to perform experiments analyzing how accurate simulation collisions are to the real thing and this remains an area of future work. Speed m/s Figure 4. Acceleration comparison for a simulated Diff Robot (. ) versus a real robot ( + ) repeated 5 times Linear Acceleration Test Time (s) Figure 5. Trajectory comparison for simulator (line) and real robot (. ) performing a set figure 8 shape. 5 RELATED WORK A plethora of robot simulators have been created to support robot research. Although some are available publicly, most are not. Most simulators do not model robot dynamics at any detailed way and are intended for generic robot research. Prime examples of this approach are

6 Player/Stage [Vaughan, 00] and TeamBots [Balch, 98]. Both Player/Stage and TeamBots is a distributed multirobot simulator and can simulate large numbers of different robots interacting in complex environments with a variety of sensors. Both simulators have limited, or no, dynamics simulation capabilities. Although TeamBots has been used to simulate games of robot soccer, the simulator again uses low-fidelity dynamics approximations and its collision detection abilities are limited. The Robot Soccer Server [Noda et al., 98], which is the official simulator for the RoboCup simulator robot soccer league, provides a simulation environment dedicated to investigating the high-level control issues for a team of distributed soccer agents. In contrast to TeamBots and Player/Stage, the Soccer Server simulates each agent as a high-level abstract robot/human and does not address robot-like motion in any way. Similarly, it lacks any meaningful collision detection capabilities. Within the robot soccer community, many develop customized simulation software. In general though, these simulators meet the minimum requirements for the task at hand and must often be rewritten as robot hardware changes. Finally, these simulators are almost never made publicly available or their details published. There are some exceptions, however. M-ROSE [Buck et al. 02] is a novel simulation engine that uses a neural network (NN) to learn the forward dynamics model using backpropagation. The forward model is trained from real robot data, and is used in its to simulate the dynamics of the robot by predicting its next state based on the current state and velocity command. This technique is quite novel and captures the robot dynamics without the need for parameterization. The drawback relates to collision dynamics, which must be implemented by hand. Indeed, the authors only implement ball-robot collision detection. It is unclear how robot-robot collision detection could be integrated in any reliable fashion to the NN. 6 CONCLUSIONS In this paper we described the development of a new Open Source, high fidelity simulation engine for robot soccer, in particular for the small size league. Our goal has been to develop an extensible simulation engine that is accurate, fast, as well as being easily reconfigured for parameter changes. To achieve these goals we have incorporated a capable Open Source physics simulation core with extensible robot classes and text based configuration files. We demonstrated the validity of our approach by applying ÜberSim to simulate our own small size team, the CMDragons small size team participants at RoboCup The empirical evidence convinces us that ÜberSim is both capable and that the concept has great potential, which we hope to harness in our future work. ACKNOWLEDGMENTS The authors would like to thank Prof. Manuela Veloso and Michael Bowling for the help and support for the development of the research described in this paper. This research was sponsored by Grants No. DABT and F The views and conclusions contained in this document are those of the authors and should not be interpreted as necessarily representing the official policies or endorsements, either expressed or implied, of the funding agencies. REFERENCES [Balch, 98] Balch, T. JavaSoccer. RoboCup-97: Robot Soccer World Cup I, Springer-Verlag, [Bruce & Veloso, 03] Bruce, J., & Veloso, M. Fast and accurate vision-based pattern detection and identification. In Proceedings of the 2003 IEEE International Conference on Robotics and Automation (ICRA 03), 2003, under submission. [Bruce et al., 02] Bruce, J., Bowling, M., Browning, B., & Veloso, M. Multi-robot team response to a multi-robot opponent team. In Proceedings of IEEE/RSJ International Conference on Robotics and Autonomous Systems, workshop on Collaborative Robots, [Buck et al., 02] Buck, S., Beetz, M., & Schmitt, T. M- ROSE: A multi robot simulation environment for learning cooperative behavior. In H. Asama, T. Arai, T. Fukuda, and T. Hasegawa (eds.): Distributed Autonomous Robotic Systems 5, Springer, [Kitano et al., 98] Kitano, H., Asada, M., Kuniyoshi, Y., Noda, I., Osawa, E., & Matsubara, H., RoboCup: A Challenge Problem for AI and Robotics. RoboCup-97: Robot Soccer World Cup I, Nagoya, L.N. on A.I., Springer Verlag, 1998, [Noda et al., 98] Noda, I., Matsubara, H., Hiraki, K. & Frank, I. Soccer Server: A tool for research on multiagent systems. Applied Artificial Intelligence, 12: , [ODE] [Reggiani et al., 02] Reggiani, M., Mazzoli, M., Caselli, S. An experimental evaluation of collision detection packages for robot motion planning. In Proceedings of the 2002 IEEE/RSJ International Conference on Intelligent Robots and Systems, (IROS 2002), [ÜberSim] [Vaughan, 00] Vaughan, R. Stage: a multiple robot simulator. Technical Report IRIS , Institute for Robotics and Intelligent Systems, School of Engineering, University of Southern California, 2000.

ÜberSim: A Multi-Robot Simulator for Robot Soccer

ÜberSim: A Multi-Robot Simulator for Robot Soccer ÜberSim: A Multi-obot Simulator for obot Soccer Brett Browning * and Erick Tryzelaar ** School of Computer Science Carnegie Mellon University 5 Forbes Avenue, Pittsburgh, PA, 523, USA + 42 268 62 * brettb@cs.cmu.edu,

More information

Motion Control in Dynamic Multi-Robot Environments

Motion Control in Dynamic Multi-Robot Environments Motion Control in Dynamic Multi-Robot Environments Michael Bowling mhb@cs.cmu.edu Manuela Veloso mmv@cs.cmu.edu Computer Science Department Carnegie Mellon University Pittsburgh, PA 15213-3890 Abstract

More information

(a) (b) (c) Fig. 1. Omnidirectional camera: (a) principle; (b) physical construction; (c) captured. of a local vision system is more challenging than

(a) (b) (c) Fig. 1. Omnidirectional camera: (a) principle; (b) physical construction; (c) captured. of a local vision system is more challenging than An Omnidirectional Vision System that finds and tracks color edges and blobs Felix v. Hundelshausen, Sven Behnke, and Raul Rojas Freie Universität Berlin, Institut für Informatik Takustr. 9, 14195 Berlin,

More information

Behavior Learning for a Mobile Robot with Omnidirectional Vision Enhanced by an Active Zoom Mechanism

Behavior Learning for a Mobile Robot with Omnidirectional Vision Enhanced by an Active Zoom Mechanism Behavior Learning for a Mobile Robot with Omnidirectional Vision Enhanced by an Active Zoom Mechanism Sho ji Suzuki, Tatsunori Kato, Minoru Asada, and Koh Hosoda Dept. of Adaptive Machine Systems, Graduate

More information

Dept. of Adaptive Machine Systems, Graduate School of Engineering Osaka University, Suita, Osaka , Japan

Dept. of Adaptive Machine Systems, Graduate School of Engineering Osaka University, Suita, Osaka , Japan An Application of Vision-Based Learning for a Real Robot in RoboCup - A Goal Keeping Behavior for a Robot with an Omnidirectional Vision and an Embedded Servoing - Sho ji Suzuki 1, Tatsunori Kato 1, Hiroshi

More information

Towards a Calibration-Free Robot: The ACT Algorithm for Automatic Online Color Training

Towards a Calibration-Free Robot: The ACT Algorithm for Automatic Online Color Training Towards a Calibration-Free Robot: The ACT Algorithm for Automatic Online Color Training Patrick Heinemann, Frank Sehnke, Felix Streichert, and Andreas Zell Wilhelm-Schickard-Institute, Department of Computer

More information

Robust Color Choice for Small-size League RoboCup Competition

Robust Color Choice for Small-size League RoboCup Competition Robust Color Choice for Small-size League RoboCup Competition Qiang Zhou Limin Ma David Chelberg David Parrott School of Electrical Engineering and Computer Science, Ohio University Athens, OH 45701, U.S.A.

More information

Nao Devils Dortmund. Team Description Paper for RoboCup Matthias Hofmann, Ingmar Schwarz, and Oliver Urbann

Nao Devils Dortmund. Team Description Paper for RoboCup Matthias Hofmann, Ingmar Schwarz, and Oliver Urbann Nao Devils Dortmund Team Description Paper for RoboCup 2017 Matthias Hofmann, Ingmar Schwarz, and Oliver Urbann Robotics Research Institute Section Information Technology TU Dortmund University 44221 Dortmund,

More information

Continuous Valued Q-learning for Vision-Guided Behavior Acquisition

Continuous Valued Q-learning for Vision-Guided Behavior Acquisition Continuous Valued Q-learning for Vision-Guided Behavior Acquisition Yasutake Takahashi, Masanori Takeda, and Minoru Asada Dept. of Adaptive Machine Systems Graduate School of Engineering Osaka University

More information

Using Layered Color Precision for a Self-Calibrating Vision System

Using Layered Color Precision for a Self-Calibrating Vision System ROBOCUP2004 SYMPOSIUM, Instituto Superior Técnico, Lisboa, Portugal, July 4-5, 2004. Using Layered Color Precision for a Self-Calibrating Vision System Matthias Jüngel Institut für Informatik, LFG Künstliche

More information

Tailored Real-Time Simulation for Teams of Humanoid Robots

Tailored Real-Time Simulation for Teams of Humanoid Robots Tailored Real-Time Simulation for Teams of Humanoid Robots Martin Friedmann, Karen Petersen, and Oskar von Stryk Simulation and Systems Optimization Group Technische Universität Darmstadt Hochschulstr.

More information

Eagle Knights 2007: Four-Legged League

Eagle Knights 2007: Four-Legged League Eagle Knights 2007: Four-Legged League Alfredo Weitzenfeld 1, Alonso Martínez 1, Bernardo Muciño 1, Gabriela Serrano 1, Carlos Ramos 1, and Carlos Rivera 1 1 Robotics Laboratory Lab, ITAM, Rio Hondo 1,

More information

Machine Learning on Physical Robots

Machine Learning on Physical Robots Machine Learning on Physical Robots Alfred P. Sloan Research Fellow Department or Computer Sciences The University of Texas at Austin Research Question To what degree can autonomous intelligent agents

More information

BabyTigers-98: Osaka Legged Robot Team

BabyTigers-98: Osaka Legged Robot Team BabyTigers-98: saka Legged Robot Team Noriaki Mitsunaga and Minoru Asada and Chizuko Mishima Dept. of Adaptive Machine Systems, saka niversity, Suita, saka, 565-0871, Japan Abstract. The saka Legged Robot

More information

Acquiring Observation Models through Reverse Plan Monitoring

Acquiring Observation Models through Reverse Plan Monitoring Acquiring Observation Models through Reverse Plan Monitoring Sonia Chernova, Elisabeth Crawford, and Manuela Veloso Computer Science Department, Carnegie Mellon University, Pittsburgh PA 15213, USA Abstract.

More information

Playing Creative Soccer: Randomized Behavioral Kinodynamic Planning of Robot Tactics

Playing Creative Soccer: Randomized Behavioral Kinodynamic Planning of Robot Tactics Playing Creative Soccer: Randomized Behavioral Kinodynamic Planning of Robot Tactics Stefan Zickler and Manuela Veloso Computer Science Department Carnegie Mellon University Pittsburgh, Pennsylvania, USA

More information

The University of Pennsylvania Robocup 2011 SPL Nao Soccer Team

The University of Pennsylvania Robocup 2011 SPL Nao Soccer Team The University of Pennsylvania Robocup 2011 SPL Nao Soccer Team General Robotics Automation, Sensing and Perception (GRASP) Laboratory University of Pennsylvania, Philadelphia, PA 19104 Abstract. This

More information

John Hsu Nate Koenig ROSCon 2012

John Hsu Nate Koenig ROSCon 2012 John Hsu Nate Koenig ROSCon 2012 Outline What is Gazebo, and why should you use it Overview and architecture Environment modeling Robot modeling Interfaces Getting Help Simulation for Robots Towards accurate

More information

Toward an Undergraduate League for RoboCup.

Toward an Undergraduate League for RoboCup. Toward an Undergraduate League for RoboCup. John Anderson, Jacky Baltes, David Livingston, Elizabeth Sklar, and Jonah Tower Department of Computer Science Columbia University New York, NY, 10027, USA sklar,jpt2002@cs.columbia.edu

More information

Behavior Programming Language and Automated Code Generation for Agent Behavior Control

Behavior Programming Language and Automated Code Generation for Agent Behavior Control Behavior Programming Language and Automated Code Generation for Agent Behavior Control Thuc Vu, Manuela Veloso tdv@andrew.cmu.edu, mmv@cs.cmu.edu Carnegie Mellon University 5000 Forbes Avenue Pittsburgh,

More information

Using surface markings to enhance accuracy and stability of object perception in graphic displays

Using surface markings to enhance accuracy and stability of object perception in graphic displays Using surface markings to enhance accuracy and stability of object perception in graphic displays Roger A. Browse a,b, James C. Rodger a, and Robert A. Adderley a a Department of Computing and Information

More information

NuBot Team Description Paper 2013

NuBot Team Description Paper 2013 NuBot Team Description Paper 2013 Zhiwen Zeng, Dan Xiong, Qinghua Yu, Kaihong Huang, Shuai Cheng, Huimin Lu, Xiangke Wang, Hui Zhang, Xun Li, Zhiqiang Zheng College of Mechatronics and Automation, National

More information

An Efficient Need-Based Vision System in Variable Illumination Environment of Middle Size RoboCup

An Efficient Need-Based Vision System in Variable Illumination Environment of Middle Size RoboCup An Efficient Need-Based Vision System in Variable Illumination Environment of Middle Size RoboCup Mansour Jamzad and Abolfazal Keighobadi Lamjiri Sharif University of Technology Department of Computer

More information

Real-Time Randomized Motion Planning for Multiple Domains

Real-Time Randomized Motion Planning for Multiple Domains Real-Time Randomized Motion Planning for Multiple Domains James Bruce and Manuela Veloso Computer Science Department Carnegie Mellon University Pittsburgh, Pennsylvania, USA {jbruce,mmv}@cs.cmu.edu Abstract.

More information

Improving Vision-Based Distance Measurements using Reference Objects

Improving Vision-Based Distance Measurements using Reference Objects Improving Vision-Based Distance Measurements using Reference Objects Matthias Jüngel, Heinrich Mellmann, and Michael Spranger Humboldt-Universität zu Berlin, Künstliche Intelligenz Unter den Linden 6,

More information

Soccer Server and Researches on Multi-Agent Systems. NODA, Itsuki and MATSUBARA, Hitoshi ETL Umezono, Tsukuba, Ibaraki 305, Japan

Soccer Server and Researches on Multi-Agent Systems. NODA, Itsuki and MATSUBARA, Hitoshi ETL Umezono, Tsukuba, Ibaraki 305, Japan Soccer Server and Researches on Multi-Agent Systems NODA, Itsuki and MATSUBARA, Hitoshi ETL 1-1-4 Umezono, Tsukuba, Ibaraki 305, Japan fnoda,matsubarg@etl.go.jp Abstract We have developed Soccer Server,

More information

Ball tracking with velocity based on Monte-Carlo localization

Ball tracking with velocity based on Monte-Carlo localization Book Title Book Editors IOS Press, 23 1 Ball tracking with velocity based on Monte-Carlo localization Jun Inoue a,1, Akira Ishino b and Ayumi Shinohara c a Department of Informatics, Kyushu University

More information

Horus: Object Orientation and Id without Additional Markers

Horus: Object Orientation and Id without Additional Markers Computer Science Department of The University of Auckland CITR at Tamaki Campus (http://www.citr.auckland.ac.nz) CITR-TR-74 November 2000 Horus: Object Orientation and Id without Additional Markers Jacky

More information

Electromagnetic Platform Stabilization for Mobile Robots

Electromagnetic Platform Stabilization for Mobile Robots Electromagnetic Platform Stabilization for Mobile Robots Eric Deng and Ross Mead University of Southern California 3710 McClintock Avenue, Los Angeles, CA 90089-0781 denge@usc.edu, rossmead@usc.edu Abstract

More information

Real time game field limits recognition for robot self-localization using collinearity in Middle-Size RoboCup Soccer

Real time game field limits recognition for robot self-localization using collinearity in Middle-Size RoboCup Soccer Real time game field limits recognition for robot self-localization using collinearity in Middle-Size RoboCup Soccer Fernando Ribeiro (1) Gil Lopes (2) (1) Department of Industrial Electronics, Guimarães,

More information

Neuro-adaptive Formation Maintenance and Control of Nonholonomic Mobile Robots

Neuro-adaptive Formation Maintenance and Control of Nonholonomic Mobile Robots Proceedings of the International Conference of Control, Dynamic Systems, and Robotics Ottawa, Ontario, Canada, May 15-16 2014 Paper No. 50 Neuro-adaptive Formation Maintenance and Control of Nonholonomic

More information

Motion Capture & Simulation

Motion Capture & Simulation Motion Capture & Simulation Motion Capture Character Reconstructions Joint Angles Need 3 points to compute a rigid body coordinate frame 1 st point gives 3D translation, 2 nd point gives 2 angles, 3 rd

More information

Particle-Filter-Based Self-Localization Using Landmarks and Directed Lines

Particle-Filter-Based Self-Localization Using Landmarks and Directed Lines Particle-Filter-Based Self-Localization Using Landmarks and Directed Lines Thomas Röfer 1, Tim Laue 1, and Dirk Thomas 2 1 Center for Computing Technology (TZI), FB 3, Universität Bremen roefer@tzi.de,

More information

Cooperative Conveyance of an Object with Tethers by Two Mobile Robots

Cooperative Conveyance of an Object with Tethers by Two Mobile Robots Proceeding of the 11th World Congress in Mechanism and Machine Science April 1-4, 2004, Tianjin, China China Machine Press, edited by Tian Huang Cooperative Conveyance of an Object with Tethers by Two

More information

XABSL - A Pragmatic Approach to Behavior Engineering

XABSL - A Pragmatic Approach to Behavior Engineering XABSL - A Pragmatic Approach to Behavior Engineering Martin Loetzsch 1,3 1 Sony CSL Paris 6, rue Amyot, F-75005 Paris, France martin.loetzsch@csl.sony.fr Max Risler 2 2 Technische Universität Darmstadt

More information

Elastic Bands: Connecting Path Planning and Control

Elastic Bands: Connecting Path Planning and Control Elastic Bands: Connecting Path Planning and Control Sean Quinlan and Oussama Khatib Robotics Laboratory Computer Science Department Stanford University Abstract Elastic bands are proposed as the basis

More information

Processing Interaction Protocols in Parallel: a Logic Programming implementation for Robotic Soccer

Processing Interaction Protocols in Parallel: a Logic Programming implementation for Robotic Soccer Processing Interaction Protocols in Parallel: a Logic Programming implementation for Robotic Soccer Mariano Tucat 1 Alejandro J. García 2 Artificial Intelligence Research and Development Laboratory Department

More information

Prof. Fanny Ficuciello Robotics for Bioengineering Visual Servoing

Prof. Fanny Ficuciello Robotics for Bioengineering Visual Servoing Visual servoing vision allows a robotic system to obtain geometrical and qualitative information on the surrounding environment high level control motion planning (look-and-move visual grasping) low level

More information

Realtime Object Recognition Using Decision Tree Learning

Realtime Object Recognition Using Decision Tree Learning Realtime Object Recognition Using Decision Tree Learning Dirk Wilking 1 and Thomas Röfer 2 1 Chair for Computer Science XI, Embedded Software Group, RWTH Aachen wilking@informatik.rwth-aachen.de 2 Center

More information

Learning Inverse Dynamics: a Comparison

Learning Inverse Dynamics: a Comparison Learning Inverse Dynamics: a Comparison Duy Nguyen-Tuong, Jan Peters, Matthias Seeger, Bernhard Schölkopf Max Planck Institute for Biological Cybernetics Spemannstraße 38, 72076 Tübingen - Germany Abstract.

More information

Plymouth Humanoids Team Description Paper for RoboCup 2013

Plymouth Humanoids Team Description Paper for RoboCup 2013 Plymouth Humanoids Team Description Paper for RoboCup 2013 Arron Griffiths 1, Clare Simpson, Paul Eastham, Alex Smith, Guido Bugmann, Chenguang Yang, Peter Gibbons, and Phil F. Culverhouse Centre for Robotics

More information

ACE Project Report. December 10, Reid Simmons, Sanjiv Singh Robotics Institute Carnegie Mellon University

ACE Project Report. December 10, Reid Simmons, Sanjiv Singh Robotics Institute Carnegie Mellon University ACE Project Report December 10, 2007 Reid Simmons, Sanjiv Singh Robotics Institute Carnegie Mellon University 1. Introduction This report covers the period from September 20, 2007 through December 10,

More information

Robust and Accurate Detection of Object Orientation and ID without Color Segmentation

Robust and Accurate Detection of Object Orientation and ID without Color Segmentation 0 Robust and Accurate Detection of Object Orientation and ID without Color Segmentation Hironobu Fujiyoshi, Tomoyuki Nagahashi and Shoichi Shimizu Chubu University Japan Open Access Database www.i-techonline.com

More information

Processing Interaction Protocols in Parallel: a Logic Programming implementation for Robotic Soccer

Processing Interaction Protocols in Parallel: a Logic Programming implementation for Robotic Soccer Processing Interaction Protocols in Parallel: a Logic Programming implementation for Robotic Soccer Mariano Tucat Alejandro J. García mt@cs.uns.edu.ar ajg@cs.uns.edu.ar Consejo Nacional de Investigaciones

More information

SHARPKUNGFU TEAM DESCRIPTION 2006

SHARPKUNGFU TEAM DESCRIPTION 2006 SHARPKUNGFU TEAM DESCRIPTION 2006 Qining Wang, Chunxia Rong, Yan Huang, Guangming Xie, Long Wang Intelligent Control Laboratory, College of Engineering, Peking University, Beijing 100871, China http://www.mech.pku.edu.cn/robot/fourleg/

More information

Object Recognition in Robot Football Using a one Dimensional Image

Object Recognition in Robot Football Using a one Dimensional Image Object Recognition in Robot Football Using a one Dimensional Image Hatice Köse and H. Levent Akin Bogaziçi University, Department of Computer Engineering, 80815 Bebek, Istanbul, TURKEY {kose,akin}@boun.edu.tr

More information

using an omnidirectional camera, sufficient information for controlled play can be collected. Another example for the use of omnidirectional cameras i

using an omnidirectional camera, sufficient information for controlled play can be collected. Another example for the use of omnidirectional cameras i An Omnidirectional Vision System that finds and tracks color edges and blobs Felix v. Hundelshausen, Sven Behnke, and Raul Rojas Freie Universität Berlin, Institut für Informatik Takustr. 9, 14195 Berlin,

More information

Motion Planning for Dynamic Knotting of a Flexible Rope with a High-speed Robot Arm

Motion Planning for Dynamic Knotting of a Flexible Rope with a High-speed Robot Arm The 2010 IEEE/RSJ International Conference on Intelligent Robots and Systems October 18-22, 2010, Taipei, Taiwan Motion Planning for Dynamic Knotting of a Flexible Rope with a High-speed Robot Arm Yuji

More information

Multi-Robot Navigation and Coordination

Multi-Robot Navigation and Coordination Multi-Robot Navigation and Coordination Daniel Casner and Ben Willard Kurt Krebsbach, Advisor Department of Computer Science, Lawrence University, Appleton, Wisconsin 54912 daniel.t.casner@ieee.org, benjamin.h.willard@lawrence.edu

More information

Visual Attention Control by Sensor Space Segmentation for a Small Quadruped Robot based on Information Criterion

Visual Attention Control by Sensor Space Segmentation for a Small Quadruped Robot based on Information Criterion Visual Attention Control by Sensor Space Segmentation for a Small Quadruped Robot based on Information Criterion Noriaki Mitsunaga and Minoru Asada Dept. of Adaptive Machine Systems, Osaka University,

More information

beestanbul RoboCup 3D Simulation League Team Description Paper 2011

beestanbul RoboCup 3D Simulation League Team Description Paper 2011 beestanbul RoboCup 3D Simulation League Team Description Paper 2011 Shahriar Asta, Tuna Sonmez, Onuralp Ulusoy, Alper Alimoglu, Mustafa Ersen, Ozdemir Sozmen, and Sanem Sariel-Talay Artificial Intelligence

More information

A Real Time Vision System for Robotic Soccer

A Real Time Vision System for Robotic Soccer A Real Time Vision System for Robotic Soccer Chunmiao Wang, Hui Wang, William Y. C. Soh, Han Wang Division of Control & Instrumentation, School of EEE, Nanyang Technological University, 4 Nanyang Avenue,

More information

Model-Based Human Motion Capture from Monocular Video Sequences

Model-Based Human Motion Capture from Monocular Video Sequences Model-Based Human Motion Capture from Monocular Video Sequences Jihun Park 1, Sangho Park 2, and J.K. Aggarwal 2 1 Department of Computer Engineering Hongik University Seoul, Korea jhpark@hongik.ac.kr

More information

Mobile Robots Locomotion

Mobile Robots Locomotion Mobile Robots Locomotion Institute for Software Technology 1 Course Outline 1. Introduction to Mobile Robots 2. Locomotion 3. Sensors 4. Localization 5. Environment Modelling 6. Reactive Navigation 2 Today

More information

Project 1 : Dead Reckoning and Tracking

Project 1 : Dead Reckoning and Tracking CS3630 Spring 2012 Project 1 : Dead Reckoning and Tracking Group : Wayward Sons Sameer Ansari, David Bernal, Tommy Kazenstein 2/8/2012 Wayward Sons CS3630 Spring 12 Project 1 Page 2 of 12 CS 3630 (Spring

More information

Multi Robot Object Tracking and Self Localization Using Visual Percept Relations

Multi Robot Object Tracking and Self Localization Using Visual Percept Relations Multi Robot Object Tracking and Self Localization Using Visual Percept Relations Daniel Göhring and Hans-Dieter Burkhard Department of Computer Science Artificial Intelligence Laboratory Humboldt-Universität

More information

Scene Management. Video Game Technologies 11498: MSc in Computer Science and Engineering 11156: MSc in Game Design and Development

Scene Management. Video Game Technologies 11498: MSc in Computer Science and Engineering 11156: MSc in Game Design and Development Video Game Technologies 11498: MSc in Computer Science and Engineering 11156: MSc in Game Design and Development Chap. 5 Scene Management Overview Scene Management vs Rendering This chapter is about rendering

More information

Graph-based Planning Using Local Information for Unknown Outdoor Environments

Graph-based Planning Using Local Information for Unknown Outdoor Environments Graph-based Planning Using Local Information for Unknown Outdoor Environments Jinhan Lee, Roozbeh Mottaghi, Charles Pippin and Tucker Balch {jinhlee, roozbehm, cepippin, tucker}@cc.gatech.edu Center for

More information

Worksheet Answer Key: Scanning and Mapping Projects > Mine Mapping > Investigation 2

Worksheet Answer Key: Scanning and Mapping Projects > Mine Mapping > Investigation 2 Worksheet Answer Key: Scanning and Mapping Projects > Mine Mapping > Investigation 2 Ruler Graph: Analyze your graph 1. Examine the shape formed by the connected dots. i. Does the connected graph create

More information

HEXAPODS FOR PRECISION MOTION AND VIBRATION CONTROL

HEXAPODS FOR PRECISION MOTION AND VIBRATION CONTROL HEXAPODS FOR PRECISION MOTION AND VIBRATION CONTROL Eric H. Anderson, Michael F. Cash, Jonathan L. Hall and Gregory W. Pettit CSA Engineering Inc., Mountain View, CA Introduction Parallel kinematic manipulators

More information

Adaptive Robotics - Final Report Extending Q-Learning to Infinite Spaces

Adaptive Robotics - Final Report Extending Q-Learning to Infinite Spaces Adaptive Robotics - Final Report Extending Q-Learning to Infinite Spaces Eric Christiansen Michael Gorbach May 13, 2008 Abstract One of the drawbacks of standard reinforcement learning techniques is that

More information

An Illumination Identification System for the AIBO Robot

An Illumination Identification System for the AIBO Robot An Illumination Identification System for the AIBO Robot Michael S. Zehmeister, Yang Sok Kim and Byeong Ho Kang School of Computing, University of Tasmania Sandy Bay, Tasmania 7001, Australia {yangsokk,

More information

Thomas Bräunl EMBEDDED ROBOTICS. Mobile Robot Design and Applications with Embedded Systems. Second Edition. With 233 Figures and 24 Tables.

Thomas Bräunl EMBEDDED ROBOTICS. Mobile Robot Design and Applications with Embedded Systems. Second Edition. With 233 Figures and 24 Tables. Thomas Bräunl EMBEDDED ROBOTICS Mobile Robot Design and Applications with Embedded Systems Second Edition With 233 Figures and 24 Tables Springer CONTENTS PART I: EMBEDDED SYSTEMS 1 Robots and Controllers

More information

Implementation of Path Planning using Genetic Algorithms on Mobile Robots

Implementation of Path Planning using Genetic Algorithms on Mobile Robots Implementation of Path Planning using Genetic Algorithms on Mobile Robots Hagen Burchardt, Ralf Salomon Abstract This paper presents the implementation of a genetic algorithm based path planning on RoboCup's

More information

CMDragons 2007 Team Description

CMDragons 2007 Team Description CMDragons 2007 Team Description James Bruce Stefan Zickler Mike Licitra Manuela Veloso December 2007 CMU-CS-07-173 School of Computer Science Carnegie Mellon University Pittsburgh, PA 15213 Abstract We

More information

Big Picture: Overview of Issues and Challenges in Autonomous Robotics + Impact on Practical Implementation Details

Big Picture: Overview of Issues and Challenges in Autonomous Robotics + Impact on Practical Implementation Details Big Picture: Overview of Issues and Challenges in Autonomous Robotics + Impact on Practical Implementation Details August 27, 2002 Class Meeting 2 Announcements/Questions Course mailing list set up: cs594-sir@cs.utk.edu

More information

y 2 x 1 Simulator Controller client (C,Java...) Client Ball position joint 2 link 2 link 1 0 joint 1 link 0 (virtual) State of the ball

y 2 x 1 Simulator Controller client (C,Java...) Client Ball position joint 2 link 2 link 1 0 joint 1 link 0 (virtual) State of the ball Vision-based interaction with virtual worlds for the design of robot controllers D. d'aulignac, V. Callaghan and S. Lucas Department of Computer Science, University of Essex, Colchester CO4 3SQ, UK Abstract

More information

Modeling the Virtual World

Modeling the Virtual World Modeling the Virtual World Joaquim Madeira November, 2013 RVA - 2013/2014 1 A VR system architecture Modeling the Virtual World Geometry Physics Haptics VR Toolkits RVA - 2013/2014 2 VR object modeling

More information

Collision Detection. Pu Jiantao.

Collision Detection. Pu Jiantao. Collision Detection Pu Jiantao. 12-09 Content Introduction CD with rays Dynamic CD using BSP Trees Hierarchical Method OBBTree Method Some Other Topics 12-1717 2 Introduction Collision Collision is a fundamental

More information

COMBINING TASK AND MOTION PLANNING FOR COMPLEX MANIPULATION

COMBINING TASK AND MOTION PLANNING FOR COMPLEX MANIPULATION COMBINING TASK AND MOTION PLANNING FOR COMPLEX MANIPULATION 16-662: ROBOT AUTONOMY PROJECT Team: Gauri Gandhi Keerthana Manivannan Lekha Mohan Rohit Dashrathi Richa Varma ABSTRACT Robots performing household

More information

FutBotIII: Towards a Robust Centralized Vision System for RoboCup Small League

FutBotIII: Towards a Robust Centralized Vision System for RoboCup Small League RoboCup-99 Team Descriptions Small Robots League, Team FutBotIII, pages 43 47 http: /www.ep.liu.se/ea/cis/1999/006/07/ 43 FutBotIII: Towards a Robust Centralized Vision System for RoboCup Small League

More information

Virtual Environments

Virtual Environments ELG 524 Virtual Environments Jean-Christian Delannoy viewing in 3D world coordinates / geometric modeling culling lighting virtualized reality collision detection collision response interactive forces

More information

UPennalizers. RoboCup Standard Platform League Team Report 2010

UPennalizers. RoboCup Standard Platform League Team Report 2010 UPennalizers RoboCup Standard Platform League Team Report 2010 Jordan Brindza, Levi Cai, Ashleigh Thomas, Ross Boczar, Alyin Caliskan, Alexandra Lee, Anirudha Majumdar, Roman Shor, Barry Scharfman, Dan

More information

Jolly Pochie 2005 in the Four Legged Robot League

Jolly Pochie 2005 in the Four Legged Robot League Jolly Pochie 2005 in the Four Legged Robot League Jun Inoue 1, Hayato Kobayashi 1, Akira Ishino 2, and Ayumi Shinohara 3 1 Department of Informatics, Kyushu University {j-inoue, h-koba}@i.kyushu-u.ac.jp

More information

Zero Robotics Autonomous Space Capture Challenge Manual

Zero Robotics Autonomous Space Capture Challenge Manual Zero Robotics Autonomous Space Capture Challenge Manual v1.3 1 Introduction 1.1 Conventions Vectors All vectors in this document are denoted with a bold face font. Of special note is the position vector

More information

Eagle Knights: RoboCup Small Size League. Dr. Alfredo Weitzenfeld ITAM - Mexico

Eagle Knights: RoboCup Small Size League. Dr. Alfredo Weitzenfeld ITAM - Mexico Eagle Knights: RoboCup Small Size League Dr. Alfredo Weitzenfeld ITAM - Mexico Playing Field and Robot Size camera 15 cm 4 m 18 cm 4 m 5.5 m Distributed Robot Control Eagle Knights (2003-2006) 2003 2004

More information

Hand-Eye Calibration from Image Derivatives

Hand-Eye Calibration from Image Derivatives Hand-Eye Calibration from Image Derivatives Abstract In this paper it is shown how to perform hand-eye calibration using only the normal flow field and knowledge about the motion of the hand. The proposed

More information

Decision Algorithm for Pool Using Fuzzy System

Decision Algorithm for Pool Using Fuzzy System Decision Algorithm for Pool Using Fuzzy System S.C.Chua 1 W.C.Tan 2 E.K.Wong 3 V.C.Koo 4 1 Faculty of Engineering & Technology Tel: +60-06-252-3007, Fax: +60-06-231-6552, E-mail: scchua@mmu.edu.my 2 Faculty

More information

Simulation Coercion Applied to Multiagent DDDAS

Simulation Coercion Applied to Multiagent DDDAS Simulation Coercion Applied to Multiagent DDDAS Yannick Loitière, David Brogan, and Paul Reynolds Computer Science Department University of Virginia, Charlottesville, VA 22901, USA {ycl2r,dbrogan,pfr}@virginia.edu

More information

Lesson 1: Introduction to Pro/MECHANICA Motion

Lesson 1: Introduction to Pro/MECHANICA Motion Lesson 1: Introduction to Pro/MECHANICA Motion 1.1 Overview of the Lesson The purpose of this lesson is to provide you with a brief overview of Pro/MECHANICA Motion, also called Motion in this book. Motion

More information

Environments. Michael Bowling and Manuela Veloso. Carnegie Mellon University. Pittsburgh, PA

Environments. Michael Bowling and Manuela Veloso. Carnegie Mellon University. Pittsburgh, PA Motion Control in Dynamic Multi-Robot Environments Michael Bowling and Manuela Veloso Computer Science Department Carnegie Mellon University Pittsburgh, PA 15213-3890 Abstract. All mobile robots require

More information

Line of Sight Stabilization Primer Table of Contents

Line of Sight Stabilization Primer Table of Contents Line of Sight Stabilization Primer Table of Contents Preface 1 Chapter 1.0 Introduction 3 Chapter 2.0 LOS Control Architecture and Design 11 2.1 Direct LOS Stabilization 15 2.2 Indirect LOS Stabilization

More information

INTERACTIVE ENVIRONMENT FOR INTUITIVE UNDERSTANDING OF 4D DATA. M. Murata and S. Hashimoto Humanoid Robotics Institute, Waseda University, Japan

INTERACTIVE ENVIRONMENT FOR INTUITIVE UNDERSTANDING OF 4D DATA. M. Murata and S. Hashimoto Humanoid Robotics Institute, Waseda University, Japan 1 INTRODUCTION INTERACTIVE ENVIRONMENT FOR INTUITIVE UNDERSTANDING OF 4D DATA M. Murata and S. Hashimoto Humanoid Robotics Institute, Waseda University, Japan Abstract: We present a new virtual reality

More information

Mobile Robots: An Introduction.

Mobile Robots: An Introduction. Mobile Robots: An Introduction Amirkabir University of Technology Computer Engineering & Information Technology Department http://ce.aut.ac.ir/~shiry/lecture/robotics-2004/robotics04.html Introduction

More information

A Modular Software Framework for Eye-Hand Coordination in Humanoid Robots

A Modular Software Framework for Eye-Hand Coordination in Humanoid Robots A Modular Software Framework for Eye-Hand Coordination in Humanoid Robots Jurgen Leitner, Simon Harding, Alexander Forster and Peter Corke Presentation: Hana Fusman Introduction/ Overview The goal of their

More information

TURN AROUND BEHAVIOR GENERATION AND EXECUTION FOR UNMANNED GROUND VEHICLES OPERATING IN ROUGH TERRAIN

TURN AROUND BEHAVIOR GENERATION AND EXECUTION FOR UNMANNED GROUND VEHICLES OPERATING IN ROUGH TERRAIN 1 TURN AROUND BEHAVIOR GENERATION AND EXECUTION FOR UNMANNED GROUND VEHICLES OPERATING IN ROUGH TERRAIN M. M. DABBEERU AND P. SVEC Department of Mechanical Engineering, University of Maryland, College

More information

Adaptive Motion Control: Dynamic Kick for a Humanoid Robot

Adaptive Motion Control: Dynamic Kick for a Humanoid Robot Adaptive Motion Control: Dynamic Kick for a Humanoid Robot Yuan Xu and Heinrich Mellmann Institut für Informatik, LFG Künstliche Intelligenz Humboldt-Universität zu Berlin, Germany {xu,mellmann}@informatik.hu-berlin.de

More information

Robot Controller. Sensor Processing Module. Locomotion Module. Sensors. Environment

Robot Controller. Sensor Processing Module. Locomotion Module. Sensors. Environment Evolution of Controllers from a High-Level Simulator to a High DOF Robot G. S. Hornby 1, S. Takamura 2, O. Hanagata 3, M. Fujita 3, and J. Pollack 1 1 Computer Science Dept., Brandeis University, Waltham,

More information

CREATION OF THE TREE MODEL ADAPTING TO ENVIRONMENT

CREATION OF THE TREE MODEL ADAPTING TO ENVIRONMENT CREATION OF THE TREE MODEL ADAPTING TO ENVIRONMENT Ryota Ueno Yoshio Ohno {ryota ohno}@on.ics.keio.ac.jp Graduate School of Science and Technology, Keio University 3-14-1 Hiyoshi, Kohoku-ku Yokohama 223-8522

More information

Figure 2.1: High level diagram of system.

Figure 2.1: High level diagram of system. Basile and Choudhury 6.111 Final Project: La PC-na Project Proposal 1 Introduction The luxury of purchasing separate pool tables, foosball tables, and air hockey tables is beyond the budget of many, particularly

More information

Interactive Educational Software for Dynamic Systems Control

Interactive Educational Software for Dynamic Systems Control Interactive Educational Software for Dynamic Systems Control José L. Lima 1, José C. Gonçalves, Paulo G. Costa 3, António P. Moreira 4 Abstract - This paper describes a tool for interactive learning that

More information

ROBOTICS 01PEEQW. Basilio Bona DAUIN Politecnico di Torino

ROBOTICS 01PEEQW. Basilio Bona DAUIN Politecnico di Torino ROBOTICS 01PEEQW Basilio Bona DAUIN Politecnico di Torino Control Part 4 Other control strategies These slides are devoted to two advanced control approaches, namely Operational space control Interaction

More information

Merging Gaussian Distributions for Object Localization in Multi-Robot Systems

Merging Gaussian Distributions for Object Localization in Multi-Robot Systems Merging Gaussian Distributions for Object Localization in Multi-Robot Systems Ashley W. Stroupe, Martin C. Martin, and Tucker Balch Robotics Institute, Carnegie Mellon University Pittsburgh, Pennsylvania,

More information

Robotics Project. Final Report. Computer Science University of Minnesota. December 17, 2007

Robotics Project. Final Report. Computer Science University of Minnesota. December 17, 2007 Robotics Project Final Report Computer Science 5551 University of Minnesota December 17, 2007 Peter Bailey, Matt Beckler, Thomas Bishop, and John Saxton Abstract: A solution of the parallel-parking problem

More information

Complexity. Object Orientated Analysis and Design. Benjamin Kenwright

Complexity. Object Orientated Analysis and Design. Benjamin Kenwright Complexity Object Orientated Analysis and Design Benjamin Kenwright Outline Review Object Orientated Programming Concepts (e.g., encapsulation, data abstraction,..) What do we mean by Complexity? How do

More information

Chapter 1: Introduction

Chapter 1: Introduction Chapter 1: Introduction This dissertation will describe the mathematical modeling and development of an innovative, three degree-of-freedom robotic manipulator. The new device, which has been named the

More information

SPATIAL GUIDANCE TO RRT PLANNER USING CELL-DECOMPOSITION ALGORITHM

SPATIAL GUIDANCE TO RRT PLANNER USING CELL-DECOMPOSITION ALGORITHM SPATIAL GUIDANCE TO RRT PLANNER USING CELL-DECOMPOSITION ALGORITHM Ahmad Abbadi, Radomil Matousek, Pavel Osmera, Lukas Knispel Brno University of Technology Institute of Automation and Computer Science

More information

Instance-Based Action Models for Fast Action Planning

Instance-Based Action Models for Fast Action Planning Instance-Based Action Models for Fast Action Planning Mazda Ahmadi and Peter Stone Department of Computer Sciences The University of Texas at Austin 1 University Station C0500, Austin, TX 78712-0233 Email:{mazda,pstone}@cs.utexas.edu

More information

MatODE. Wouter Caarls and Erik Schuitema {w.caarls, December 21, 2011

MatODE. Wouter Caarls and Erik Schuitema {w.caarls, December 21, 2011 MatODE Wouter Caarls and Erik Schuitema {w.caarls, e.schuitema}@tudelft.nl December 21, 2011 Contents 1 Introduction 1 1.1 What is ODE?............................ 1 1.2 What is matode?..........................

More information

Real-Time Randomized Path Planning for Robot Navigation

Real-Time Randomized Path Planning for Robot Navigation Real-Time Randomized Path Planning for Robot Navigation James Bruce (jbruce@cs.cmu.edu) Manuela Veloso (mmv@cs.cmu.edu) Computer Science Department Carnegie Mellon University 5000 Forbes Avenue Pittsburgh

More information