RSS 2014 Workshop

Constrained decision-making in robotics: models, algorithms, and applications 

Berkeley, July 13, 2014

Organizers:

Stefano Carpin
School of Engineering
University of California, Merced
Marco Pavone
Dept. of. Aeronautics and Astronautics
Stanford University
Schedule Call for Papers Special Issue

 

Abstract

As the complexity of robotic tasks grows, robotic decision makers increasingly face the problem of trading off different objectives. For example, a rescue robot might be required to plan trajectories so as to maximize the probability of success to reach a victim and, at the same time, minimize the duration of the traversal (a task often contrasting the one of ensuring safety). A natural framework for this class of problems is constrained decision-making, whereby a decision maker seeks to optimize a given cost function (often stochastic) while keeping other costs (usually involving risk assessments) below given bounds. In the last decade, the operations research community has made significant strides on the topics of constrained decision-making (notoriously more challenging than the unconstrained counterpart) and risk assessments in dynamic scenarios. The result is a comprehensive theory and a set of algorithmic tools for (risk)-constrained decision-making. Yet, despite their relevance, these results have seen limited application within the robotics domain. Accordingly, the objective of this workshop is threefold: 
(1) To convene together, arguably for the first time, researchers working in the areas of decision-making, risk theory, and robotics to facilitate a joint discussion on (risk)-constrained decision-making in robotics.  
(2) To inform robotic researchers about the state of the art in constrained decision-making and modern risk theory.
(3) To formulate a research agenda on the topics of risk modeling for robotics applications, algorithmic approaches for robotic decision-making under constraints, and application of these results to robotic planning.
 

Confirmed Speakers

- Prof. Emilio Frazzoli, Massachusetts Institute of Technology. Background: aeronautics and astronautics/operations research. Focus of the talk: risk-constrained planning for autonomous driving. Confirmed speaker.
 
- Prof. Rahul Jain, University of Southern California. Background: electrical engineering/operations research. Focus of the talk: algorithms for risk-constrained decision-making. Confirmed speaker.
 
- Dr.   Masahiro Ono, NASA Jet Propulsion Laboratory. Background: aeronautics and astronautics. Focus of the talk: risk-constrained planning for autonomous spacecraft control. Confirmed speaker.
 
- Prof. Marco Pavone, Stanford University. Background: aeronautics and astronautics. Focus of the talk: algorithms for risk-constrained planning and control. Confirmed speaker.
 
- Dr.   Alessandro Pinto, United Technologies Research Center. Background: electrical engineering/computer science. Focus of the talk: risk-constrained planning for strategic deployment of ground robots. Confirmed speaker.
 
- Prof. Gaurav Sukhatme, University of Southern California. Background: computer science. Focus of the talk: risk-constrained planning for underwater applications. Confirmed speaker.
 
- Prof. Mykel Kochenderfer, Stanford University. Background: aeronautics and astronautics. Focus of the talk: decision theoretic approaches for next generation aircraft collision avoidance. Confirmed speaker.

 

Schedule

The worshop will take place in room Wheeler 210 (see http://ets.berkeley.edu/classrooms/210-wheeler for details).

Time  
8:20-8:30 Welcome address -- Stefano Carpin, Marco Pavone
8:30-9:00 Marco Pavone
Risk-averse and risk-constrained stochastic optimal control
9:00-9:30 Rahul Jain
Risk-aware Stochastic Optimization
9:30-10:00 Emilio Frazzoli &  Pratik Chaudhari
Sampling-based Algorithms for Risk-constrained Autonomous Urban Navigation
10:00-10:30 Coffee Break
10:30-11:00 Masahiro Ono
Chance-constrained optimal control: from smart grid to Mars EDL and mobility planning
11:00-11:30 Mykel Kochenderfer
A decision theoretic approach for next generation aircraft collision avoidance
11:30-15:00 Lunch Break (FIFA World Cup Final)
15:00-15:30

Poster session

S.D. Bopardikar, B. Englot, A. Speranzon
Chance-Constrained Multi-Objective Path Planning Under State Uncertainty

Y. Cui, J.T. Lane, R. Voyles
Real-time, On-board System Health Management for Resource Constrained Field Robotics

N. Demir, B. Acikmese
Probabilistic Density Control of Swarms of Mobile Agents with Generalized Safety Constraints

C. Erdogan, M. Stilman
Ensuring Buildability of Simple Machine Designs with Task-Constrained Motion Planning

S. Feyzabadi, S. Carpin
Risk-aware Path Planning Using Hiererachical Constrained Markov Decision Processes

15:30-16:00 Alessandro Pinto
Application of Constrained Decision Processes to Mission Planning
16:00-16:30 Gaurav Sukhatme
Beyond information gathering: physical sampling as a constrained decision-making problem
16:30-17:00 Coffe break
17:00-17:30 Discussion and concluding remarks

 

Followup Journal Special Issue

 

A followup special issue of the Autonomous Robots journal on the topic of constrained decision making in robotics has been approved. Click here for the Call for Papers. Initial Submission deadline is October 15, 2014. Submission will be open to everyone, but workshop presenters and attendees are particularly encouraged to submit their contributions.