Robot trajectory optimization using approximate inference

Toussaint, Marc

doi:10.1145/1553374.1553508

Cited by 277 publications

(338 citation statements)

References 9 publications

Supporting

Mentioning

337

Contrasting

Order By: Relevance

“…Our first contribution is to generate legible motion via functional gradient optimization in the space of trajectories (Sec. IV), echoing earlier works in motion planning [9,21,22,27,29,33,35], now with legibility as an optimization criterion. Fig.1 depicts this optimization process: by exaggerating the motion to the right, the robot makes the other goal option, G O , far less likely to be inferred by the observer that the correct goal G R .…”

Section: Introductionmentioning

confidence: 74%

Generating Legible Motion

Dragan¹,

Srinivasa²

2013

Robotics: Science and Systems IX

106

119

View full text Add to dashboard Cite

Abstract-Legible motion -motion that communicates its intent to a human observer -is crucial for enabling seamless human-robot collaboration. In this paper, we propose a functional gradient optimization technique for autonomously generating legible motion. Our algorithm optimizes a legibility metric inspired by the psychology of action interpretation in humans, resulting in motion trajectories that purposefully deviate from what an observer would expect in order to better convey intent. A trust region constraint on the optimization ensures that the motion does not become too surprising or unpredictable to the observer.Our studies with novice users that evaluate the resulting trajectories support the applicability of our method and of such a trust region. They show that within the region, legibility as measured in practice does significantly increase. Outside of it, however, the trajectory becomes confusing and the users' confidence in knowing the robot's intent significantly decreases.

show abstract

Section: Introductionmentioning

confidence: 74%

Generating Legible Motion

Dragan¹,

Srinivasa²

2013

Robotics: Science and Systems IX

106

119

View full text Add to dashboard Cite

show abstract

“…Despite recent efforts at reducing the computational costs of such global searches for a feasible path (Diankov and Kuffner, 2007;Burns and Brock, 2005;Toussaint, 2009), these methods cannot offer the reactivity sought for swiftly avoiding obstacles that appear suddenly.…”

Section: Related Workmentioning

confidence: 99%

A dynamical system approach to realtime obstacle avoidance

2012

View full text Add to dashboard Cite

This paper presents a novel approach to real-time obstacle avoidance based on dynamical systems (DS) that ensures impenetrability of multiple convex shaped objects. The proposed method can be applied to perform obstacle avoidance in Cartesian and Joint spaces and using both autonomous and non-autonomous DS-based controllers. Obstacle avoidance proceeds by modulating the original dynamics of the controller. The modulation is parameterizable and allows to determine a safety margin and to increase the robot's reactiveness in the face of uncertainty in the localization of the obstacle. The method is validated in simulation on different types of DS including locally and globally asymptotically stable DS, autonomous and non-autonomous DS, limit cycles, and unstable DS. Further, we verify it in several robot experiments on the 7 degrees of freedom Barrett WAM arm.

show abstract

“…When we collaborate, we make two inferences ( Fig.1,lower center) about our collaborator [12,54,56]: 1) we infer their goal based on their ongoing action (action-to-goal), and 2) if we know their goal, we infer their future action from it (goal-to-action). Our work Fig.…”

Section: Introductionmentioning

confidence: 99%

Integrating human observer inferences into robot motion planning

Dragan

Srinivasa

2014

Auton Robot

View full text Add to dashboard Cite

Abstract-Our goal is to enable robots to produce motion that is suitable for human-robot collaboration and coexistence. Most motion in robotics is purely functional, ideal when the robot is performing a task in isolation. In collaboration, however, the robot's motion has an observer, watching and interpreting the motion.In this work, we move beyond functional motion, and introduce the notion of an observer into motion planning, so that robots can generate motion that is mindful of how it will be interpreted by a human collaborator.We formalize predictability and legibility as properties of motion that naturally arise from the inferences in opposing directions that the observer makes, drawing on action interpretation theory in psychology. We propose models for these inferences based on the principle of rational action, and derive constrained functional trajectory optimization techniques for planning motion that is predictable or legible.Finally, we present experiments that test our work on novice users, and discuss the remaining challenges in enabling robots to generate such motion online in complex situations. 1

show abstract

Robot trajectory optimization using approximate inference

Cited by 277 publications

References 9 publications

Generating Legible Motion

Generating Legible Motion

A dynamical system approach to realtime obstacle avoidance

Integrating human observer inferences into robot motion planning

Contact Info

Product

Resources

About