Human Productivity in a Workspace Shared with a Safe Robotic Manipulator

McGhan, Catharine L. R.; Atkins, Ella M.

doi:10.2514/1.54993

“…The state has three mission goals of sate hunger (g i 1 ), sate thirst (g i 2 ), and complete a general mission-oriented work effort (g i 3 ), as well as one high-priority goal of button inactive (f i 1 ) that indicates that a button needs to be pushed (corresponding to a safety-critical mission task that might need to be completed). We have conducted previous human subject experiments of these tasks with a safe robotic manipulator arm that confirms the feasibility of a shared workspace [12].…”

Section: Case Studysupporting

confidence: 54%

“…The human-aware robot decision-making process we recommend uses linked but separate MDPs: one for human intent prediction and one for robot action choice. In RAC, the robot's state and actions must be modeled as well as the human's state and actions; the decoupling we propose is only possible because of our assumption that the human will not be influenced by the robot's actions [12]. Suppose the robot is executing a simple set of button-pressing tasks.…”

Section: Case Studymentioning

confidence: 99%

Human Intent Prediction Using Markov Decision Processes

McGhan

¹

,

Nasir

²

,

Atkins

³

2015

Journal of Aerospace Information Systems

Self Cite

View full text Add to dashboard Cite

This paper describes a modeling method for predicting a human's task-level intent through the use of Markov Decision Processes. Intent prediction can be used by a robot to improve decision-making when human and robot operate in a shared physical space. This work presumes human and robot goals are independent such that the robot seeks to avoid interfering with the human rather than directly assisting the human. The proposed human intent prediction system transforms goal sequences the human is expected to complete, a limited past action history, and a correlation of observed behaviors with actions into a prediction of the in-progress or next action the humans is most likely to take. An intra-vehicle activity space robotics application example is presented.

show abstract

“…The state has three mission goals of sate hunger (g i 1 ), sate thirst (g i 2 ), and complete a general mission-oriented work effort (g i 3 ), as well as one high-priority goal of button inactive (f i 1 ) that indicates that a button needs to be pushed (corresponding to a safety-critical mission task that might need to be completed). We have conducted previous human subject experiments of these tasks with a safe robotic manipulator arm that confirms the feasibility of a shared workspace [12]. Equation ( 5) gives our domain representation:…”

Section: Case Studymentioning

confidence: 77%

Human Intent Prediction Using Markov Decision Processes

McGhan

¹

,

Nasir

²

,

Atkins

³

2012

Infotech@Aerospace 2012

View full text Add to dashboard Cite

This paper describes a modeling method for predicting a human's task-level intent through the use of Markov Decision Processes. Intent prediction can be used by a robot to improve decision-making when human and robot operate in a shared physical space. This work presumes human and robot goals are independent such that the robot seeks to avoid interfering with the human rather than directly assisting the human. The proposed human intent prediction system transforms goal sequences the human is expected to complete, a limited past action history, and a correlation of observed behaviors with actions into a prediction of the in-progress or next action the humans is most likely to take. An intra-vehicle activity space robotics application example is presented.

show abstract

“…From previous human subject testing, we have determined that, in our ground-based scenario, the inclusion of the robot in shared-workspace operations without explicit communications did not significantly reduce human productivity, even when only minimal conflict-avoidance algorithms were used for 'intelligent' task-selection. 14,16 This implies that so long as the robot causes only minimal interference or conflict with the human, the human would be expected to have similar productivity as if the robot was not there, so any additional goals accomplished by the robot would improve overall team productivity.…”

Section: The Use Of Predicted Companion Intent Results In Improved Rementioning

confidence: 99%

“…Our representations below correspond to a simple example scenario that is discussed in more detail in our previous work. 14 We propose a two-step decision process that allows a robot to determine the locally-optimal action-choice for overall human-robot team efficiency and productivity with constraints imposed to maintain a minimum level of safety, where safety translates to collision avoidance in this work. We describe the information transfer between architectural modules and specifics of the MDP models required for human intent prediction (HIP) and action choice.…”

Section: Introductionmentioning

confidence: 99%

Towards Guaranteeing Safe and Efficient Human-Robot Collaboration Using Human Intent Prediction

McGhan

¹

,

Atkins²

2012

AIAA SPACE 2012 Conference &Amp; Exposition

0

View full text Add to dashboard Cite

This paper describes an autonomous framework for determining a robotic manipulator's optimal actions in real-time when interacting in close physical proximity to a human in a shared workspace environment. This framework allows the robot to purposefully choose to avoid physical and mental conflicts with a human companion while each agent performs tasks to complete their respective, separately-assigned goals. We pose scenarios in which the human does not need to divert attention to internally model the robot's behavior, or track or acknowledge the robot's actions during operations. The robot is meant to unobtrusively 'work around' the human rather than directly collaborate on task completion. The distinction of this work is in its use of human intent prediction (HIP) as a key factor in robot action selection for task-level planning. We choose to model HIP with a Markov Decision Process (MDP). Human state data is input into the HIP MDP policy that then outputs the predicted human intent, which we define as the best-matched or most-likely in-progress and future action-choice(s) that the human is or will be pursuing to complete mission goals. Predicted human intent is then used by a second MDP to determine the optimal policy with respect to the robot's action-choice. We present an autonomous framework that integrates the HIP MDP and robot action-choice (RAC) MDP to support autonomous close-proximity operations and propose offline and online (scaled) formulations of the two MDPs. During real-time policy execution, once the optimal action for the robot to take is determined, it is passed to the robot's path planner to be translated from a task-level command to a trajectory and motion primitives, which are then given to a low-level controller to enact. We evaluate our HIP MDP in simulation, and find that the policy output from our system is consistent and smooth across small changes in parameter values.

show abstract

Human Productivity in a Workspace Shared with a Safe Robotic Manipulator

Cited by 5 publications

References 24 publications

Human Intent Prediction Using Markov Decision Processes

Human Intent Prediction Using Markov Decision Processes

Human Intent Prediction Using Markov Decision Processes

Towards Guaranteeing Safe and Efficient Human-Robot Collaboration Using Human Intent Prediction

Contact Info

Product

Resources

About