Planning and Control of Robot Motion at IROS 2009

[Summary by Kostas Bekris]

A large body of work at IROS related strongly to planning and control algorithms for robot motion. Thus, the following overview cannot be exhaustive but it attempts to highlight some of the challenges that the community, and especially the members of our committee, are trying to address with planning and control solutions.

Some of the papers focused on kinodynamic motion planning either by computing appropriate curves that directly satisfy dynamic constraints [1,2] or by studying how random linear projections of complex systems can be utilized to improve planning with dynamics [3]. A paper focused on complementarity-based dynamic simulation algorithms, which are often used as local planners in sampling-based algorithms [4]. The paper showed that complementarity-based solutions perform well in cluttered environments and are relatively insensitive to parameter choices. Such physically realistic simulation tools were also employed in order to solve disassembly problems [5] or in order to predict the appropriate actions by a real robot playing the game of Jenga [6].

Another focus area corresponded to collision-free motion coordination for multiple robots or agents. Previous work on velocity obstacles and reciprocal velocity obstacles was extended in order to account for constraints of car-like robots [7] and uncertainty in sensor data so as to avoid oscillations [8]. Another work dealt with "group control problems", such as simulating the operation of a shepherd agent that guides flocks of robots through obstacle-filled environments [9]. Other multi-agent works focused on pursuit-evasion problems. One paper dealt with the "lion and man" game in the presence of circular obstacles [10], while a different work introduced a simulation platform for studying such problems [11].

Many papers dealt with problems related to dynamic and/or uncertain environments and the need for anytime and real-time planning. A Bug-like algorithm for anytime planning was presented, which showed favorable performance in simulation [12]. Another paper extended the RRT algorithm by explicitly considering the robot's mobility in uneven terrains and the resulting uncertainty [13]. Other papers focused on manipulators. For example, one work dealt with the real-time computation of whether a manipulator's trajectory is collision-free without any minimum-distance calls [14]. Moreover, Willow Garage's PR2 robot was employed to show how to integrate 3D perception with real-time replanning for personal robotics applications [15].

Motion planning specifically for manipulation was the focus of multiple sessions. A paper showed how planning using the best single hypothesis under pose uncertainty for the manipulator often fails and how it could be remedied by the proposed approach [16]. A different paper dealt with pose uncertainty in the base of a mobile manipulator and proposed an approach based on the idea of Lazy-PRM [17]. The focus of another paper was to utilize ideas from control in order to address task-constrained manipulation and planning problems for redundant, possibly nonholonomic systems [18]. There were also manipulation papers that did not dealt with robotic arms. For example, one work presented a mechanism and a control strategy for non-contact manipulation of spherical objects in three dimensions using air-flow [19].

There were also multiple sessions devoted to humanoid robots and multi-arm systems, where many planning papers appeared. A paper decomposed a multi-arm system into parts and combined the results from different roadmap-based techniques for the overall solution [20]. A method for generating humanoid whole-body motion derived purely as a result of moving the head joint was also presented at IROS [21]. Other works focused on solving regrasping tasks for multi-arm systems [22] and controlling the internal forces and moments produced during multi-contact interactions between humanoid robots and the environment [23].

A series of planning papers dealt with challenges in medical robotics. For example, how to plan the motion of steerable medical needles [24] or active cannulas [25], medical devices composed of thin, pre-curved, telescoping tubes that may enable many new surgical procedures. Or, finally, how to automate a common surgical procedure of exposing an underlying area by grasping and lifting a thin layer of tissue [26].

References:

  1. Kinodynamic Motion Planning for Mobile Robots Using Splines
    Boris Lau, Cristoph Sprunk, Wolfram Burgard - Univ. of Freiburg
  2. Computing Clothoid Segments for Trajectory Generation
    Doran Wilde - Bringham Young University
  3. On the performance of Random Linear Projections for Sampling-based Motion Planning
    Ioan Alexandru Sucan, Lydia Kavraki - Rice University
  4. Complementarity-based Dynamic Simulation for Kinodynamic Motion Planning
    Nilanjan Chakraborty, Srinivbas Akella, Jeff Trinkle - RPI & CMU
  5. Efficient Planning of Disassembly Sequences in Physics-based Animation
    Aleotti Jacopo, Caselli Stefano - University of Parma
  6. Robot Jenga: Autonomous and Strategic Block Extraction
    Jiuguang Wang, Philip Rogers, Lonnie Parker, Douglas Brooks, Mike Stilman - Georgia Tech.
  7. Generalized Velocity Obstacles
    David Wilkie, Jur van den Berg, Dinesh Manocha - UNC Chapel Hill
  8. Independent Navigation of Multiple Mobile Robots with Hybrid Reciprocal Velocity Obstacles
    Jamie Snape, Jur van den Berg, Stephen J. Guy, Dinesh Manocha - UNC Chapel Hill
  9. Behavior-Based Motion Planning for Group Control
    Chirstopher Vo, Joseph Harrison, Jyh-Ming Lien - George Mason
  10. Lion and Man Game in the Presence of a Circular Obstacle
    Nikhil Karnad, Volkan Isler - University of Minnesota
  11. Intelligent Pursuit and Evasion in an Unknown Environment
    Jonathan Annas, Jing Xiao - UNC Charlotte
  12. A Bug-Inspired Algorithm for Efficient Anytime Path Planning
    Javier Antich, Alberto Ortiz and Javier Minguez - University of the Balearic Islands & University of Zaragoza
  13. Stochastic Mobility-based Path Planning in Uncertain Environments
    Gaurav Kewlani, Genya Ishigami, Karl Iagnemma - MIT
  14. Perceiving Guaranteed Continuously Collision-Free Robot Trajectories in an Unknown and Unpredictable Environment
    Rayomand Vatcha, Jing Xiao - UNC Charlotte
  15. Real-Time Perception-Guided Motion Planning for a Personal Robot
    Radu Bogdan Rusu, Ioan Alexandru Sucan, Brian Gerkey, Sachin Chitta, Michael Beetz, Lydia Kavraki - Tech. Univ. Munchen, Rice University & Willow Garage
  16. Addressing Pose Uncertainty in Manipulation Planning Using Task Space Regions
    Dmitry Berenson, Siddhartha Srinivasa, James Kuffner - CMU
  17. Lazy-PRM for a Manipulator with Base Pose Uncertainty
    Yifeng Huang, Kamal Gupta - Simon Fraser University
  18. A Control-based Approach to Task-Constrained Motion Planning
    Giuseppe Oriolo, Marilena Vendittelli - Univ. di Roma "La Sapienza"
  19. Automated Manipulation of Spherical Objects in Three Dimensions using a Gimbaled Air Jet
    Aaron Becker, Robert Sandheinrich, Timothy Bretl - UIUC
  20. Roadmap Composition for Multi-Arm Systems Path Planning
    Mokhtar Gharbi, Juan Cortes, Thierry Simeon - LAAS-CNRS
  21. Steering a humanoid robot by its head
    Manish Sreenivasa, Phillippe Soueres, Jean-Paul Laumond, Alain Berthoz - CNRS
  22. Humanoid Motion Planning for Dual-Arm Manipulation and Re-Grasping Tasks
    Nikolaus Vahrenkamp, Dmitry Berenson, Tamim Asfour, James Kuffner, Rudiger Dillmann - Univ. of Karlsruhe & CMU
  23. Modeling and Control of Multi-Contact Centers of Pressure and Internal Forces in Humanoid Robots
    Luis Sentis, Jaeheung Park, Oussama Khatib - Stanford
  24. Planning Fireworks Trajectories for Steerable Medical Needles to Reduce Patient Trauma
    Jijie Xu, Vincent Duindam, Ron Alterovitz, Jean Pouliot, J. Adam M. Cunha, I Chow-Hsu and Ken Goldberg - Rochester IT, UC Berkeley, UNC Chapel Hill, UC San Francisco
  25. Motion Planning for Active Cannulas
    Lisa Lyons, Robert Webster, Ron Alterovitz - UNC Chapel Hill & Vanderbilt
  26. Surgical Retraction of Non-Uniform Deformable Layers of Tissue: 2D Robot Grasping and Path Planning
    Rik Jansen, Kris Hauser, Nuttapong Chentanez, Frank van der Stappen, Ken Goldberg - Utrecht University, UC Berkeley