Optimizing Contextual Ergonomics Models in Human-Robot Interaction

Abstract

Current ergonomic assessment procedures require observation and manual annotation of postures by an expert, after which ergonomic scores are inferred from these annotations. Our aim is to automate this procedure, and to enable robots to optimize their behavior with respect to such scores. A particular challenge is that ergonomic scoring requires accurate biomechanical simulations which are computationally too expensive to use in robot control loops or optimization. To address this, we learn Contextual Ergonomics Models, which are Gaussian Process Latent Variable Models that have been trained with full musculoskeletal simulations for specific tasks contexts. Contextual Ergonomics Models enable search in a low-dimensional latent space, whilst the cost function can be defined in terms of the full high-dimensional musculoskeletal model, which can be quickly reconstructed from the latent space. We demonstrate how optimizing Contextual Ergonomics Models leads to significantly reduced muscle activation in an experiment with eight subjects performing a drilling task.

Operations management issues in design and control of hybrid human-robot collaborative manufacturing systems: a survey

“…To prevent injuries, there exist tools able to detect dangerous situations and send a warning to workers. Several danger prediction technics exist, such as monitoring and detecting collisions (Wang et al, 2013;Schmidt & Wang, 2014), human motions in an assembly line (Liu & Wang, 2017;Coupeté et al, 2019;Unhelkar et al, 2018;Marin et al, 2018;Gopinathan et al, 2018), workers' presence and intentions (Koch et al, 2017).…”

Predicting and Optimizing Ergonomics in Physical Human-Robot Cooperation Tasks

“…The first trend is the task of holding a workpiece in the optimal position in space while a human works on it. For example, in tasks like drilling in which a human applies a tool to an object, the position of the object held by a robot has been ergonomically optimized to minimize joint torques [17], [18], muscular effort [19] or the RULA score [20], [21].…”

Caring About the Human Operator: Haptic Shared Control for Enhanced User Comfort in Robotic Telemanipulation

“…They derived a continuous cost function based on the Rapid Entire Body Assessment score (REBA) and used it to choose the robot position optimizing the human joint angles (and thus the ergonomic comfort). Marin et al [24] optimized the ergonomics of an HRI task where subjects are asked to drill on a board carried by the robot. They relied on musculoskeletal simulations to train a Contextual Ergonomics Model through a probabilistic supervised learning Gaussian process.…”