Michael Spitzhirn — IMK automotive GmbH
Why collaborative robotics may answer occupational health issues and associated challenges

Human-robot-collaboration (HRC) is currently being redefined. Since only a few years, new robotic concepts are developed, which enable the human operator to directly cooperate with the robot without any physical barriers, such as cages and curtains like in the past. First industrial applications are implemented with the goal of improving production efficiency and reducing ergonomic strains for the operator. However, new questions and risks related to occupational safety arise and the need for a precise production planning with virtual methods becomes more important.

Introducing our workshop, this talk shows current challenges in occupational health across Europe, such as the aging workforce and the related increase of musculo-skeletal disorders among industrial workers. At the same time, companies have to improve their efficiency continuously due to the inexorable global competition. New collaborative robots may help coping with some of those challenges. However, robots need to be used for tasks they can really help with, not just for any task in the industry. And human-robot collaboration needs to be designed in an ergonomically way so that the efficiency of the entire work system is being increased and the safety of the human operator is guaranteed.

This talk is introducing the AnDy project, which aims to find new methods for optimizing dyadic human-robot-interaction. Besides the use of motion capturing technologies, which will be shown later in the workshop, the project also aims to enhance current software solutions like “EMA – Editor for Manual Work Activities” with specific functions for the virtual planning of human-robot-collaboration in 3D environments. A live demonstration of the EMA software and the AnyBody software for biomechanical modelling will show the current capabilities and explain further developments planned in the scope of the AnDy project. Furthermore, we will show some examples of successfully implemented work systems with optimized human-robot-interaction. Finally, we will discuss other interactive technologies that may support human operators in industrial work settings, such as exoskeletons and smart wearables.


Angelos Karatsidis — Xsens
Ergonomic Assessment in the Workplace Using Wearable Sensors

Ergonomics aims at optimizing design of systems, products, or processes by taking into account the way people interact with them. Therefore, accurate assessment of human body properties such as poses, movements, and interactions with the environment is fundamental for effective ergonomic analysis.
Traditional camera-based and force plates systems are proven as effective tools for this scope, but they are strongly limiting the analysis to a carefully set-up (mostly indoor) lab-like environment. Moreover, optical systems suffer from problems due to marker occlusions and light conditions, which may result in marker swap and noisy or missing data. For this reason, additional time-consuming post-processing steps are typically required. On the other hand, wearable technologies have demonstrated increased potential as an alternative solution for kinematic and kinetic analysis, especially for ambulatory assessment in the natural environment.

Inertial motion capture systems are typically composed of several inertial and magnetic measurement units (IMMU) firmly placed on a person’s body segments. IMMUs measure physical quantities, such as accelerations, angular velocities, and magnetic field, which are fed to a sensor fusion algorithm and combined with an underlying biomechanical model to estimate body poses and kinematics. This enables effective ambulatory motion tracking and effortless kinematic analysis in everyday-life conditions.

So far, due to the use of magnetometers, magnetic disturbances often resulted in degraded performance of IMMU measurements, especially in natural workplace environments due to the presence of ferro-magnetic objects such as chairs, tables, etc. However, latest advances in statistical state estimation and numerical optimization algorithms, in combination with use of background calibration routines, complex biomechanical models, and innovative sensor fusion and signal processing frameworks, can provide solutions to typical limitations of current technologies. In this way, highly accurate and consistent full body human motion tracking is possible in any environment using consumer-grade, low-power, inexpensive, accelerometer and gyroscope sensors.

Empowered by more consistent and accurate inertial motion tracking performance, recent studies performed full-body inverse dynamics to obtain estimates of joint forces and moments. In particular, recent works proposed approaches to estimate ground reaction forces and moments during gait only using inertial sensor inputs. Moreover, a workflow to perform musculoskeletal model simulations and estimate muscle, bone, and ligament forces has been already demonstrated using the Xsens MVN system. The further use of wearable force sensors, such as instrumented force shoes or pressure insoles, can provide additional increase in performance, consistency, and applicability in diverse use-scenarios of relevance.

In this session, we will show how wearable technologies can enable advanced kinematic and kinetic analysis for ergonomics, anytime and anywhere.


Gentiane Venture — Tokyo University of Agriculture and Technology
My model is not your model: personalized models for motion analysis

Models of the human body play a key role in human motion science. In particular, the dynamics relates the movement to the forces indispensable to achieve this motion. It also relates to the environment through interaction forces. Measuring this data is not always trivial. In the past decade we have developed solutions for the computation of the dynamic quantities and developed individual models. In this presentation I will present the state of the art and our latest advances in this area and show some examples of applications.


Pavel Galibarov — AnyBody Technology
Musculoskeletal Model-based Ergonomics for Collaborative Robotics

Workers on factory floor perform daily long, repetitive and high demanding tasks that lead to musculoskeletal disorders (MSDs) in the long term. Collaborative robotics has been brought up as a viable means to reduce MSDs. However, Robots lack understanding of human ergonomy that results in inefficient collaboration.

This work will demo the pipeline of ergonomics, where the data are coming from, how they are processed, and how they are deemed to be applied to the robots. The focus however will be on where in that workflow, musculoskeletal modeling will be placed, and how it can improve the ergonomics of human-robot collaboration.

Standard ergonomic scores (ES), such as EAWS (Ergonomic Assessment Worksheet), usually focus on basic static postures and static load cases. These standards also suffer from discontinuity of such cases so that the ES jumps from one number to another with sometimes a significant gap in between. Musculoskeletal models are well-established tools that access ensemble and individual muscle/joint dynamics, which could serve to improve standard ergonomic assessments.

The goal of this work is to harvest some of the many physiology-based variables of such models in real-life human-robot collaboration tasks. To this end, physiological metrics for musculoskeletal offline ergonomic assessment of tasks in factories will be introduced. Such metrics could include detailed muscle activity, joint load, metabolic energy consumption and fatigue. Ideas on how to standardize these quantitative measures to provide a collective value for ergonomic score benchmarking as well as improving commonly used ones will be introduced and showcased. Examples of ergonomics simulations will be showcased, and for each example, use of various musculoskeletal variables will be discussed.

Ideally these human-informed data could be fed back in real-time both to the robot and the human for more ergonomic collaboration, e.g. human posture correction and robot dynamic assistance level. This will potentially assist filling the gap of robots having a blind spot in current human-robot physical collaboration.


Frank Krause — Vrije Universiteit Amsterdam
Acceptance and effectivity of an upper extremity exoskeleton

Work-related shoulder injuries have been associated with overhead work, which is a frequent task conducted in industry (Shin et al. 2012). A passive exoskeleton can be assistive in overhead work by compensating gravity due to arm weight and carried load. Several exoskeleton systems have been designed to this purpose. Main challenge for the developers is to make these systems effective in reducing the risk for shoulder injury, while limiting the negative side effects which may stand in the way of acceptance.
In this paper, we review the state of the art of current upper extremity exoskeletons that are presented on the Internet or described in the literature, and the state of knowledge about their effect on shoulder load and acceptance. Additionally, we will present and discuss the effect of a specific upper arm exoskeleton, namely Skelex, on shoulder muscle activation and underlying factors of acceptance.

We studied the acceptance and effectivity of Skelex, both in a field and in a laboratory study. This field study took place in the aerospace sector and involved workers exposed to prolonged overhead assembly work. Main outcome measures included experienced shoulder load, mechanical torque provided by the system, and subjectively experienced productivity, usability and acceptance. In the laboratory, we measured the provided torque by the system, and the shoulder muscle activation in a range of selected (quasi-)static elevated arm postures in the sagittal and transversal plane and compared with vs. without exoskeleton conditions.

Results and conclusions
In the past three years passive and actuated upper extremity exoskeletons have been developed particularly in the past three years. However, reports on their acceptance and effectivity in shoulder load reduction are limited. Some study results (e.g. Huysamen et al. in press, Theurel et al. 2018) show potential effectivity in reduction of shoulder muscle activity and simultaneously a need for further development to increase potential acceptance (e.g. by reducing weight and size, by increasing motion freedom).
For Skelex, the results obtained regarding field acceptance and effectivity in terms of reductions of muscle activation, will be presented during the conference. On the basis of these results we will show the range of postures in which Skelex is particularly effective in reducing muscle activation.


Jonas Bornmann — Ottobock
Development and practical experience of industrial exoskeletons

The industrial exoskeletons has the goal to improve the ergonomic situation of humans at work by giving an adequate physical assistance. The demonstration session will contain the development and practical experience of an industrial exoskeleton. The main task of the presented industrial exoskeleton is the reduction of the stress in the upper limbs during overhead work. Extensive requirements were captured in a user-oriented design process with a following evaluation under real working conditions. The uniqueness of the exoskeleton is to preserve the flexibility of the human movement and safety at the working place. During the development attention was particularly payed to orthopedic technologies, user acceptance and biomechanical considerations.