IMU-based motion capture system for rehabilitation applications: A systematic review, discusses the specific applications of IMU-based Mocap systems in the field of rehabilitation therapy from three aspects.
Abstract
In recent years, the application of Motion Capture (Mocap) systems based on Inertial Measurement Unit (IMU) in rehabilitation therapy has rapidly expanded. This article discusses the specific applications of IMU-based Mocap systems in the field of rehabilitation therapy from three aspects: rehabilitation applications, target populations, and system deployment and measurement parameters. Additionally, the paper explores the direction of development for Mocap systems applied to upper limb rehabilitation under sensor constraints. Literature analysis indicates that studies on rehabilitation assessment, training, and those encompassing both constitute 82%, 12%, and 6% respectively. The majority of research primarily targets stroke patients. Furthermore, general rehabilitation methods not confined to any particular pathological disease have also been extensively studied, with the paper especially emphasizing the importance of lower limb gait analysis research.
Rehabilitation application
Rehabilitation applications primarily fall into two categories: rehabilitation assessment and rehabilitation training. Rehabilitation assessment encompasses evaluating the patient’s condition, rehabilitation objectives, monitoring their movements, and analyzing relevant parameters, which aids in determining the severity of the patient’s condition and the effectiveness of the rehabilitation plan. Rehabilitation training involves executing specific tasks or plans, typically carried out post-assessment, to provide intervention measures or feedback information for the patient’s rehabilitation. Literature analysis shows that the majority of IMU applications are for rehabilitation assessment (82%), with fewer in rehabilitation training (12%), and a minority of studies addressing both rehabilitation assessment and training (6%), as illustrated in Figure 2.
Target population
Target populations are typically categorized into three types: neurological rehabilitation, musculoskeletal injury, and general rehabilitation. Within the type of neurological rehabilitation, stroke patients are the most researched group, with about 14 articles focusing on stroke rehabilitation, while the research proportion for other neurological diseases is relatively low. A total of 16 articles were included in the analysis of musculoskeletal injuries, with most studies concentrating on osteoarthritis. Additionally, 25 articles employed IMU-based motion capture technology for general rehabilitation.
These findings showcase the multifunctionality of IMU-based motion capture technology. It can be utilized not only for specific diseases but also for broader applications, such as diagnostics, disease prevention, and daily monitoring.
System deployment and measurement
Figure 3 illustrates the proportion of IMU studies deployed on different segments of the human body. Accordingly, for the upper-body analysis, two IMU sensors are usually selected and placed on the upper arm and forearm. For the lower-limb analysis, two IMU sensors are selected and placed on each of the legs for gait analysis and the thigh and shank for knee joint analysis. It is apparent that the majority of IMU studies are conducted on the lower limbs, with fewer studies on the upper limbs. Regarding human kinematics, joint orientation is most frequently measured due to its convenience and clear representation. Followed by specific gait parameters, spatiotemporal parameter measurement is the most commonly used method in gait analysis.
Conclusion
This article investigates the motion capture systems based on IMUs in the field of rehabilitation, with a particular focus on three aspects: rehabilitation applications, target populations, and system deployment and measurement.
The most extensively researched application is rehabilitation assessment, with relatively lesser focus on rehabilitation training. Patients with neurological and musculoskeletal disorders constitute the primary target populations, among which stroke is the most frequently analyzed condition. Additionally, non-pathological rehabilitation, particularly gait analysis, has been widely studied. Regarding system deployment and measurement parameters, statistical data indicates that employing two IMUs for lower limb gait analysis to measure spatiotemporal parameters is the most commonly researched method.
However, the limitations posed by sensor drift and low battery longevity have resulted in a lack of precise analysis of upper limb motion and long-term training applications in existing studies. To address these issues, we propose potential solutions from two perspectives: 1) to tackle low battery longevity, the development of low-power chips or low-energy transmission methods is suggested; 2) to resolve the sensor drift problem, utilizing sensor fusion to create multimodal systems is recommended, to minimize sensor errors and provide additional information.
Reference
[1] C. Gu, W. Lin, X. He, L. Zhang, and M. Zhang, “IMU-based motion capture system for rehabilitation applications: A systematic review”, Biomimetic Intelligence and Robotics, vol. 3, no. 2, p. 100097, 2023, doi: 10.1016/j.birob.2023.100097.
