Towards quantitative modeling of task confirmations in human-robot dialog

Abstract

Abstract-We present a technique for robust human-robot interaction taking into consideration uncertainty in input and task execution costs incurred by the robot. Specifically, this research aims to quantitatively model confirmation feedback, as required by a robot while communicating with a human operator to perform a particular task. Our goal is to model human-robot interaction from the perspective of risk minimization, taking into account errors in communication, risk involved in performing the required task, and task execution costs. Given an input modality with non-trivial uncertainty, we calculate the cost associated with performing the task specified by the user, and if deemed necessary, ask the user for confirmation. The estimated task cost and the uncertainty measure is given as input to a Decision Function, the output of which is then used to decide whether to execute the task, or request clarification from the user. In cases where the cost or uncertainty (or both) is estimated to be exceedingly high by the system, task execution is deferred until a significant reduction in the output of the Decision Function is achieved. We test our system through human-interface experiments, based on a framework custom designed for our family of amphibious robots, and demonstrate the utility of the framework in the presence of large task costs and uncertainties. We also present qualitative results of our algorithm from field trials of our robots in both open-and closed-water environments.

Ensuring safety in human-robot dialog — A cost-directed approach

“…To estimate task costs, we use two custom simulation models for each type of robot. For the underwater vehicle, we use a set of assessors that takes into account the operating contexts of an autonomous underwater vehicle, and use the simulator from [3]. The simulator has been designed to take into account the robot's velocity, maneuverability and propulsion characteristics to accurately and realistically simulate trajectories taken by the robot.…”

“…For the sake of brevity, we assume that a mechanism exists to only allow syntactically valid and semantically consistent commands to be sent to the robot. Details of the language structure can be found in [3], but for the sake of completeness a summary is given here.…”

“…In this work, human-robot interaction occurs using a structured language called RoboChat [1] [2], which has been previously used to program and communicate with an amphibious legged robot in both terrestrial and underwater environments. RoboChat enables a human partner to program a robot using an arbitrary communication medium, which may include but would not be limited to visual gestures, audio and more "traditional" input methods (such as mice and keyboards, or touch interfaces) Our previous work has looked at quantitatively modeling human-robot dialog, accounting for uncertainty in input and corresponding task cost [3]. In that work, after the human operator programs a robot to perform a set of tasks, the robot assesses command execution cost and uncertainty in the input, and uses a Decision Function to decide whether or not to confirm this task.…”

Human–robot planning and learning for marine data collection

“…There has been a recent push in environmental monitoring towards the development of Decision Support Systems that allow the human operator to seamlessly track the progress of autonomous vehicles and to issue commands on the fly (Das et al 2011;Li et al 2006;Sattar and Dudek 2011). Such systems are capable of monitoring the progress of autonomous vehicles operating in the ocean and in other unstructured environments by providing data to scientists in real time.…”

Trajectory learning for human-robot scientific data collection

“…There is a recent thrust in environmental monitoring towards the development of Decision Support Systems that allow the human operator to seamlessly track the progress of autonomous vehicles and to issue commands on the fly [19], [20], [21]. Such systems are capable of monitoring the progress of autonomous vehicles operating in the ocean and in other unstructured environments by providing data to scientists in real time.…”