The objective of this proposal is to address core scientific challenges related to sensing, actuation and control of cyber-physically assistive clothing (CPAC). CPAC is a kind of Human-in-the-loop Cyber-Physical System (HCPS), in which actuated clothing is coordinated in unison with human body movement to enhance safety and health. We propose addressing key HCPS challenges within the context of using CPAC to reduce societal incidence of low back pain, by preventing lumbar (spine) overloading and overuse injuries. Low back pain is targeted because it is one of the leading causes of physical disability and missed work. High and/or repetitive forces on lumbar muscles and discs can occur during daily tasks, and are known to be major risk factors that can lead to back pain and injury. The long-term vision is to create smart clothing that can monitor lumbar loading, train safe movement patterns, and directly assist wearers to reduce the musculoskeletal forces that cause pain and injury. This proposed transformation of clothing is similar to how wristwatches have transformed from timepieces into health monitors; however, CPAC is even more exciting because it combines the form-factor of clothing with the assistance benefits of an exoskeleton to reduce biological tissue loading for a broad range of individuals, occupations and tasks. Thrust 1 will adapt machine learning techniques in order to monitor lumbar loading and detect excessive spine forces via portable, wearable sensors, such that timely feedback/intervention can be provided. This thrust will result in the creation of a publicly shared data set that contains synchronized, multimodal (lab-based and wearable) sensor data collected from >500 actions per subject, the largest such corpus for machine learning in this domain. Thrust 2 will model the dynamics of cyber, physical and human components of CPAC in order to develop optimal control and learning strategies. Thrust 3 will integrate sensors, fusion algorithms and portable actuation into a complete wearable prototype. A human subject experiment will be performed to objectively evaluate the function of CPAC. At the focus of this proposal is the human body; monitored, analyzed and assisted by multidisciplinary CPS technologies. The project integrates expertise in biomechanics, machine learning, sensor fusion, soft robotics, wearable assistive technology, and clinical management of low back pain to transform clothing from materials that cover the body into wearable systems that can track and protect low back health. The key scientific HCPS challenges that need to be overcome, and which are addressed in this proposed research, in order to realize the broad societal benefits of CPAC are: (1) real-time sensing and assistive control of the HCPS and its co-adaptation to different subjects and diverse environments, (2) system design and verification ensuring safe operation and that no harm is done to human subjects through unanticipated feedback, (3) selection and placement of low cost sensors aiding affordable and realistic manufacturing of CPAC, (4) integration of wearable sensors and actuators into a reliable and effective HCPS.
CPAC has the potential for broad societal impact given the high prevalence of low back pain. CPAC provides a unique and potentially paradigm-shifting opportunity because it combines the assistive benefits of an exoskeleton with the lightweight and unobtrusive form-factor of everyday clothing. CPAC is expected to (1) fit seamlessly into daily life, (2) connect individuals with health info to empower them to modify their physical activity, (3) generate public health data to inform evidence-based clinical and workplace practices, (4) augment lifting biomechanics to reduce lumbar loading, (5) reduce incidence of low back pain, and costs due to health care and missed work.
