Multi-task versatile control development and physical interaction assessment for occupational back-support exoskeletons

The leading cause of workplace absenteeism in Europe is musculoskeletal injuries, which constitute 60% of work-related health problems. The back is the body section most affected, accounting for 43% of the incidence. Reducing work-related musculoskeletal disorders (WMSDs) would have a significant impact on both the long-term health of workers and the economy of companies and societies by decreasing absenteeism, thereby saving money and preserving the workforce. WMSDs are caused by the prolonged performance of strenuous work, particularly manual handling of loads and long-term maintenance of incongruous postures. Different solutions have been explored to support the backs of industrial workers with passive exoskeletons (i.e., relying on spring or elastic-based actuation units). In contrast, active devices (i.e., relying on electric motors or battery-powered actuators) should be further investigated, particularly in the field. Although there is a growing interest in researching occupational exoskeletons, their adoption by industry remains limited. Reasons accredited by the scientific community include the lack of customized development for targeted tasks and industries, and the limited number of assessments in real work scenarios. The lack of customized developments includes the design and implementation of task-specific control strategies that suitably support the worker throughout the activity execution. On the other hand, evaluation in real work environments is a fundamental step in determining the actual effectiveness and applicability of these developments. The following PhD research aims to fill this gap by proposing solutions and presenting evidence to facilitate the deployment of occupational exoskeletons in workplaces. The research focuses on developing control strategies deployed on an active exoskeleton (StreamEXO) to support the trunk and hip joints of construction workers during heavy-duty tasks, such as manual material handling (MMH). StreamEXO features an advanced kinematic design that enhances comfort, stability, and adjustability and has been customized to meet the requirements of the construction sector. Thus, the requirements of end-users have been carefully and fully evaluated to define tailored solutions that maximize the effectiveness of the control strategies. After the requirements analysis, this work is structured following two iterative phases: (i) Task-specific control strategies and high-level control architecture development; (ii) Evaluation of the developments with a focus on the ergonomic risk assessment and end-users’ acceptance. The requirement analysis focused on investigating occupational exoskeletons in industrial settings and WMSDs. Evidence shows that the construction sector ranks fourth in Europe for WMSDs affecting the back, making it an excellent use case. Moreover, only 18\% of back-support exoskeletons tested in industrial tasks are active. The effects and performance of active, passive, task-specific, and generic-purpose exoskeletons were compared during multiple bending tasks. The results showed that active and versatile exoskeletons (i.e., exoskeletons that can change the assistive strategy according to the activity performed) are more suitable for supporting workers, reducing the self-reported fatigue and the activity of agonistic muscles. Thus, versatile exoskeletons were selected for future developments. In the development phase (i), task-specific control strategies were designed and implemented to actively support the worker during static and dynamic bending and locomotion. In fact, static and dynamic bending are very different and require specific control strategies to be assisted. On the other hand, the hauling of materials in MMH involves the carrying of loads. Therefore, a dedicated assistive strategy is required to fully support this task. However, during MMH, workers are unable to manually select the desired assistive strategy with a user interface, especially during hauling. Thus, the device should autonomously recognize the activity performed by the user and change the assistive strategy accordingly, to provide adequate assistance without hindrance. For this scope, a high-level control architecture capable of recognizing the main activities associated with MMH was developed. In the assessment phase (ii), the control performance and the impact of the exoskeleton on reducing ergonomic risk in complex work activities were evaluated. Through a systematic analysis of ergonomics that combines classic and innovative methods, it was verified that the use of the exoskeleton improves postures and reduces the perceived load lifted, thereby decreasing the overall ergonomic risk. On-site tests were conducted with construction workers performing hauling activities. Feedback on the performance and acceptance of the exoskeleton was collected through questionnaires. Results on human factors show an 87\% acceptance rate and a 42\% reduction in physical exertion perceived on the back. Overall, this research contributes to advancing the scientific literature on active occupational exoskeletons by providing new methods to evaluate performance in complex tasks, assessing the impact on ergonomic risk, and developing task-specific control strategies and architectures for versatile assistance. The results suggest that these developments on the exoskeleton can pave the way for increasing the adoption of occupational exoskeletons in construction, potentially decreasing WMSDs, and bridge the gap between academia and industry to promote the large-scale adoption of occupational exoskeletons, making workplaces safer.