Generating Legible Motion

By Anca Dragan, Siddhartha S. Srinivasa. 2013

200 citations recorded by the source corpus.

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.

Subfields: Control and Systems Engineering, Natural Language Processing, Robotics & Control

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Control and Systems Engineering

In Natural Language Processing

In Robotics & Control

Citations in the corpus

Learning Dynamic Robot-to-Human Object Handover from Human Feedback (2016, background): “In the second stage, the giver and taker signal to each other and exchange information, often through non-verbal communication, such as motion [12], eye gaze, or head orientation [13], in order to establish joint intention of handover.”
Learning Trajectory Preferences for Manipulators via Iterative Improvement (2013, methodology): “Furthermore, humans like to anticipate robots move and to gain users’ confidence, robot should produce predictable and legible robot motion [8].”
Plan explicability and predictability for robot task planning (2015, methodology): “Our problem settings are as follows: an intelligent agent is given a goal by a human (so that the human knows the goal of the agent) working in the same environment and it needs to synthesize a plan to achieve the goal.”
Cooperative Inverse Reinforcement Learning (2016, background): “Examples include ﬁnding a motion that best communicates an agent’s intention (Dragan & Srinivasa, 2013), or ﬁnding a natural language utterance that best communicates a particular grounding (Golland et al., 2010).”
Asking for Help Using Inverse Semantics (2014, background): “To address these problems, we propose an alternative approach to recovering from the inevitable failures which occur when robots execute complex tasks in real-world environments: when the robot encounters failure, it verbally requests help from a human partner.”
Effects of Robot Motion on Human-Robot Collaboration (2015, methodology): “This section summarizes the three motion planning paradigms compared in the study: purely functional motion planning, which is typically achieved via sampling-based methods, and predictable and legible motion planning as defined in our prior work [4], achieved via trajectory optimization.”
Shared autonomy via hindsight optimization for teleoperation and teaming (2017, methodology): “Blending methods (Dragan and Srinivasa, 2013b) attempt to bridge the gap between highly assistive methods with little user control, and minimal assistance with higher user burden.”
Translating Neuralese (2017, background): “On the control side, a long line of work considers nonverbal communication strategies in multiagent policies (Dragan and Srinivasa, 2013).”
Learning preferences for manipulation tasks from online coactive feedback (2015, methodology): “Other recent works consider generating human-like trajectories [15, 14, 58].”
Effective robot teammate behaviors for supporting sequential manipulation tasks (2015, background): “Accordingly, recent work on collaboration has been published with a focus on motion planning that implicitly conveys intent [13], as well as motion planning that accommodates the expected motions of others within an interaction [14].”
Human-robot collaborative manipulation planning using early prediction of human motion (2013, methodology): “The recent work by Dragan and Srinivasa [14] proposes a legibility metric inspired by the psychology of action interpretation applied to robot motion planning using functional gradient optimization.”
Interpreting multimodal referring expressions in real time (2016, methodology): “Dragan and Srinivasa [5] created a framework for generating legible gesture, and we anticipate that enabling a robot to respond using gestures as in Holladay et al. [10] will further increase the efficacy of this system.”