Walking Turn Prediction from Upper Body Kinematics: A Systematic Review with Implications for Human-Robot Interaction

Abstract

Prediction of walking turns allows to improve human factors such as comfort and perceived safety in human-robot interaction. The current state-of-the-art suggests that upper body kinematics can be used for that purpose and contains evidence about the reliability and the quantitative anticipation that can be expected from different variables. However, the experimental methodology has not been consistent throughout the different works and the related data has not always been given in an explicit form, with different studies containing partial, complementary or even contradictory results. In this paper, with the purpose of providing a uniform view of the topic that can trigger new developments in the field, we performed a systematic review of the relevant literature addressing three main questions: (i) Which upper body kinematic variables permit to anticipate a walking turn? (ii) How long in advance can we anticipate the turn from them? (iii) What is the expected contribution of walking turn prediction systems from upper body kinematics for human-robot interaction? We have found that head yaw was the most reliable kinematical variable from the upper body to predict walking turns about 200ms. Trunk roll anticipates walking turns by a similar amount of time, but with less reliability. Both approaches may benefit human-robot interaction in close proximity, helping the robot to exhibit appropriate proxemic behavior interacting at intimate, personal or social distances. From the point of view of safety, they have to be considered with caution. Trunk yaw is not valid to anticipate turns. Gaze Yaw seems to be the earliest predictor, although existing evidence is still inconclusive.

Autonomous Bayesian escorting of a human integrating intention and obstacle avoidance

“…To address this, context cues have been studied and shown to improve prediction strategies but are often not used alone (Blaiotta, 2019;Schulz & Stiefelhagen, 2015;Unhelkar et al, 2015). For example, Lopez et al (2019)…”

“…Head orientation is the implicit indicator of motion observed because it has been found to redirect before a human changes their moving direction (Lopez et al, 2019; Patla & Vickers, 2003; Sreenivasa et al, 2008). The head detection method using a stereo camera proposed by Fischer et al (2018) is implemented as it can reliably detect the 3D orientation.…”

“…To address this, context cues have been studied and shown to improve prediction strategies but are often not used alone (Blaiotta, 2019; Schulz & Stiefelhagen, 2015; Unhelkar et al, 2015). For example, Lopez et al (2019) observed upper‐body kinematics including head yaw and trunk roll predict walking turns. In addition, Sreenivasa et al (2008) found the head turned into the direction of the turn at a constant distance before the obstacle ($$\unicode{x0007E}$$1.1 m) for turn angles up to $$13{5}^{\circ }$$.…”

“…The mathematical analysis validates the need for human intention inference in robotic escorting. For the human intention, head pose is used as an implicit indicator, which was proven in past literature (Lopez et al, 2019; Patla & Vickers, 2003; Sreenivasa et al, 2008). The future human pose is then predicted after (1) predicting the human pose independently of the human intention using physics‐based techniques and a particle filter (PF) and (2) fusing the independently predicted human poses using the Gaussian multiplication.…”

Real-Time Short-Term Pedestrian Trajectory Prediction Based on Gait Biomechanics

“…In [ 16 ], the short-term trajectory prediction is improved by tracking the user’s head in the context of virtual reality applications. Head orientation anticipates the trajectory, but it is also heavily influenced by other ambient-related impulses or distractions [ 17 ]. The authors suggest the use of eye trackers to overcome this problem, but this technology is too expensive to be used in more general intelligent environments.…”