Effective Human-Robot Collaboration in near symmetry collision scenarios

Abstract

Recent works in the domain of Human-Robot Motion (HRM) attempted to plan collision avoidance behavior that accounts for cooperation between agents. Cooperative collision avoidance between humans and robots should be conducted under several factors such as speed, heading and also human attention and intention. Based on some of these factors, people decide their crossing order during collision avoidance. However, whenever situations arise in which the choice crossing order is not consistent for people, the robot is forced to account for the possibility that both agents will assume the same role i.e. a decision detrimental to collision avoidance. In our work we evaluate the boundary that separates the decision to avoid collision as first or last crosser. Approximating the uncertainty around this boundary allows our collision avoidance strategy to address this problem based on the insight that the robot should plan its collision avoidance motion in such a way that, even if agents, at first, incorrectly choose the same crossing order, they would be able to unambiguously perceive their crossing order on their following collision avoidance action.

Waist rotation angle as an indicator of probable human collision avoidance direction for autonomous mobile robots

“…Silva et al proposed an effective measure based on robotics that focused on boundary dividing collision avoidance into "first" or "subsequent" between humans and robots (Silva et al, 2019).…”

“…However, they must ensure the safety of pedestrians, even under complex situations (Meerhoff et al, 2018). Thus, several studies in the field of robotics engineering have utilized human-like behaviors to develop robots that are safer and friendlier to use (Kamezaki, 2019; Kruse et al, 2013; Silva et al, 2019; Tamura et al, 2010; Tomizawa & Shibata, 2016). For instance, Kamezaki verified control methods whereby a robot guided or made way for a human depending on the situation, based on the estimation of the positional relationship between the human and the robot (Kamezaki, 2019).…”

Robust collaborative collision avoidance between robots with nearly symmetric crossing trajectories

“…Accepting that roles are sometimes reversed reveals a subtle issue [12]: whenever in any pairwise interaction only a single agent reverses their role ineffective collaboration is produced. A recent approach to this issue [12] considered that only a person can mistake its role. In our work a general model is proposed based on a sensor-specific error model where any robot can be at fault.…”

“…More precisely, the decision of agents r and i of passing each other on right or on the left side will influence the sign of the integral over time of their derivative of the bearing angle Effective collision avoidance requires that agents attempt to pass each other on the same side, that is, make the same homotopy class decision. In experiments with people, the derivative of the bearing angle has been shown to reliably predict the homotopy class decision of an agent [12] The derivative of the bearing angle over time in four distinct head-on collision scenarios. Robots circumscribed radius is 0.8m and move at 1.2 m/s.…”

“…any collision avoidance action has been made. Based on this, a simplified predictor of homotopy class decision before collision avoidance has started, presented in [12], can be used as below…”