ICRA 2015 Workshop: Beyond Geometric Constraints: Planning for Solving Complex Tasks, Reducing Uncertainty, and Generating Informative Paths & Policies

May 30, 2015, 1:30-5:00, Seattle, Washington, USA

The workshop will highlight the latest advances in resolving general robotic planning problems that march beyond robot motion and path planning under geometric and differential constraints. Equipped with a solid foundation in solving fundamental tasks such sensing/perception and decision making/motion planning, researchers have begun to tackle issues that appear when multiple aspects of a task must be handled simultaneously in an integrated application. Complex interactions frequently arise due to non-linear dependencies between the sub-problems. Alternatively, conflicting priorities may appear when multiple robots couple through geometric, communication, and other resource-sharing induced constraints. In recent years, efforts geared toward understanding such interactions have begun to pay off, as will be made evident by the presentations at the workshop. Nevertheless, much work remains to be done. Through invited talks, interactive presentations of contributed work, and open discussion sessions, the objective of the workshop is to bring together experts to establish where we are at solving these general planning problems, reach a better understanding of the inherent structures that give rise to complex interactions in such problems, pinpoint the universal and immediate key challenges that hinder the untangling of the interactions, and develop ideas for new techniques and algorithmic solutions for solving these challenges. We look forward to seeing you at the ICRA 2015 workshop on Beyond Geometric Constraints: Planning for Solving Complex Tasks, Reducing Uncertainty, and Generating Informative Paths & Policies.

 1:30 - 1:35 Welcome & Introduction
 1:35 - 1:55  Invited Talk: Integrated Task and Motion Planning in Belief Space
Leslie Kaelbling & Tomas Lozano-Perez, Massachusetts Institute of Technology
 1:55 - 2:15  Invited Talk: Exploiting Task and Constraint Structure in Motion Planning
Kris Hauser, Duke University
 2:15 - 2:35  Invited Talk: Monte Carlo Motion Planning for Robot Motion Optimization Under Uncertainty
Marco Pavone, Stanford University
 2:35 - 2:55  Invited Talk: Multi-robot Routing Algorithms for Coverage in Rich Environments
Vijay Kumar & Pratap Tokekar, University of Pennsylvania
 2:55 - 3:30 
Highlights from Contributed Papers & Poster Session
  1. POMDP Based Manipulation Planning in Object Composition Space
    Joni Pajarinen and Ville Kyrki

  2. Understanding the Past to Predict the Future: Multipolicy Decision-Making using Changepoints for Autonomous Driving
    Enric Galceran, Alexander G. Cunningham, Ryan M. Eustice and Edwin Olson

  3. Image-Based Visual Servoing for Differentially Flat Underactuated Systems using Cross-Entropy Motion Planning
    Matthew Sheckells and Marin Kobilarov

  4. Variable Density PRM Waypoint Generation and Connection Radii for Energy-Efficient Flight through Wind Fields
    Jen Jen Chung, Nicholas R.J. Lawrance, Seng Keat Gan, Zhe Xu, Robert Fitch and Salah Sukkarieh

  5. Failure Avoidance through Learning Guided Probabilistic Planning
    Melis Kapotoglu, Cagatay Koc, and Sanem Sariel

  6. Risk Aware Graph Search With Uncertain Edge Costs
    Ryan Skeele, Jen Jen Chung, Geoffrey A. Hollinger and Kagan Tumer

  7. Shortest Path Exploration with Fast Marching
    Nicholas R.J. Lawrance, Jen Jen Chung and Geoffrey A. Hollinger

  8. Active Object Recognition via Monte Carlo Tree Search
    Mikko Lauri, Nikolay Atanasov, George J. Pappas and Risto Ritala

  9. Tunable-Risk Sampling-Based Path Planning Using a Cost Hierarchy
    Tixiao Shan and Brendan Englot

  10. Theory and Implementation of Dual-Arm Manipulation Planning
    Kensuke Harada

  11. Supervised Learning: A potential tool for feasible motion planning in State Space
    Mukunda Bharatheesha and Martijn Wisse

 3:30 - 3:50  Invited Talk: Adaptive Informative Path Planning in Metric Spaces
David Hsu, The National University of Singapore
 3:50 - 4:10  Invited Talk: Constraints in Informative Path Planning
Gaurav Sukhatme, University of Southern California
 4:10 - 4:30  Invited Talk: Human Perception Meets Robot Perception
Steve LaValle, University of Illinois at Urbana Champaign
 4:30 - 4:45  Invited Talk: Codesign of sensing, computation, and actuation
Andrea Censi & Emilio Frazzoli, Massachusetts Institute of Technology
 4:45 - 5:00  Closing Discussion

Sertac Karaman (MIT),  Daniela Rus (MIT),  Mac Schwager (BU), and Jingjin Yu* (MIT).