Dataset Information

Would a thermal sensor improve arm motion classification accuracy of a single wrist-mounted inertial device?

ABSTRACT:

Background

Inertial Measurement Unit (IMU)-based wearable sensors have found common use to track arm activity in daily life. However, classifying a high number of arm motions with single IMU-based systems still remains a challenging task. This paper explores the possibility to increase the classification accuracy of these systems by incorporating a thermal sensor. Increasing the number of arm motions that can be classified is relevant to increasing applicability of single-device wearable systems for a variety of applications, including activity monitoring for athletes, gesture control for video games, and motion classification for physical rehabilitation patients. This study explores whether a thermal sensor can increase the classification accuracy of a single-device motion classification system when evaluated with healthy participants. The motions performed are reproductions of exercises described in established rehabilitation protocols.

Methods

A single wrist-mounted device was built with an inertial sensor and a thermal sensor. This device was worn on the wrist, was battery powered, and transmitted data over Bluetooth to computer during recording. A LabVIEW Graphical User Interface (GUI) instructed the user to complete 24 different arm motions in a pre-randomized order. The received data were pre-processed, and secondary features were calculated on these data. These features were processed with Principal Component Analysis (PCA) for dimensionality reduction and then several machine learning models were applied to select the optimal model based on speed and accuracy. To test the effectiveness of the scheme, 11 healthy subjects participated in the trials.

Results

Average personalized classification model accuracies of 93.55% were obtained for 11 healthy participants. Generalized model accuracies of 82.5% indicated that the device can classify arm motions on a user without prior training. The addition of a thermal sensor significantly increased classification accuracy of a single wrist-mounted inertial device, from 75 to 93.55%, (F(1,20)?=?90.53, p?=?7.25e-09).

Conclusion

This study found that the addition of the thermal sensor improved the classification accuracy of 24 arm motions from 75 to 93.55% for a single-device system. Our results provide evidence that a single device can be used to classify a relatively large number of arm motions from arm rehabilitation protocols. While this study provides a conceptual proof-of-concept with a healthy population, additional investigation is required to evaluate the performance of this system for specific applications, such as activity classification for physically affected stroke survivors undergoing home-based rehabilitation.

SUBMITTER: Lui J

PROVIDER: S-EPMC6505300 | biostudies-literature | 2019 May

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundElbow and wrist chronic conditions are very common among musculoskeletal problems. These painful conditions affect muscle function, which ultimately leads to a decrease in the joint's Range Of Motion (ROM). Due to their portability and ease of use, goniometers are still the most widespread tool for measuring ROM. Inertial sensors are emerging as a digital, low-cost and accurate alternative. However, whereas inertial sensors are commonly used in research studies, due to the lack of information about their validity and reliability, they are not widely used in the clinical practice. The goal of this study is to assess the validity and intra-inter-rater reliability of inertial sensors for measuring active ROM of the elbow and wrist.Materials and methodsMeasures were taken simultaneously with inertial sensors (Werium™ system) and a universal goniometer. The process involved two physiotherapists ("rater A" and "rater B") and an engineer responsible for the technical issues. Twenty-nine asymptomatic subjects were assessed individually in two sessions separated by 48 h. The procedure was repeated by rater A followed by rater B with random order. Three repetitions of each active movement (elbow flexion, pronation, and supination; and wrist flexion, extension, radial deviation and ulnar deviation) were executed starting from the neutral position until the ROM end-feel; that is, until ROM reached its maximum due to be stopped by the anatomy. The coefficient of determination (r 2) and the Intraclass Correlation Coefficient (ICC) were calculated to assess the intra-rater and inter-rater reliability. The Standard Error of the Measurement and the Minimum Detectable Change and a Bland-Altman plots were also calculated.ResultsSimilar ROM values when measured with both instruments were obtained for the elbow (maximum difference of 3° for all the movements) and wrist (maximum difference of 1° for all the movements). These values were within the normal range when compared to literature studies. The concurrent validity analysis for all the movements yielded ICC values ≥0.78 for the elbow and ≥0.95 for the wrist. Concerning reliability, the ICC values denoted a high reliability of inertial sensors for all the different movements. In the case of the elbow, intra-rater and inter-rater reliability ICC values range from 0.83 to 0.96 and from 0.94 to 0.97, respectively. Intra-rater analysis of the wrist yielded ICC values between 0.81 and 0.93, while the ICC values for the inter-rater analysis range from 0.93 to 0.99.ConclusionsInertial sensors are a valid and reliable tool for measuring elbow and wrist active ROM. Particularly noteworthy is their high inter-rater reliability, often questioned in measurement tools. The lowest reliability is observed in elbow prono-supination, probably due to skin artifacts. Based on these results and their advantages, inertial sensors can be considered a valid assessment tool for wrist and elbow ROM.

Project description:Ascertaining the accuracy of the pedicle screw (PS) trajectories is important as PS malpositioning can cause critical complications. We aimed to determine the angle range over which estimation is unreliable; build a low-cost PS placement support system that uses an inertial measurement unit (IMU) to enable the monitoring of surgical tools and PS trajectories, and determine the situations where IMU support would be most beneficial. In PS insertion experiments, we used cadaver samples that included lumbar porcine spines. Computed tomography images obtained before and after PS insertion were viewed. Offsets between the planned and implanted PS trajectories in the freehand and IMU-assisted groups were analyzed. The PS cortical bone breaches were classified according to the Gertzbein and Robbins criteria (GRC). Added head-down tilted sample experiments were repeated wherein we expected a decreased rostro-caudal rotational accuracy of the PS according to the angle estimation ability results. Evaluation of the PS trajectory accuracy revealed no significant advantage of IMU-assisted rostro-caudal rotational accuracy versus freehand accuracy. According to the GRC, IMU assistance significantly increased the rate of clinically acceptable PS positions (RoCA) than the freehand technique. In the head-down tilted sample experiments, IMU assist provided increased accuracies with both rostro-caudal and medial rotational techniques when compared with the freehand technique. In the freehand group, RoCA was significantly decreased in samples with rostral tilting relative to that in the samples without. However, In the IMU-assisted group, no significant difference in RoCA between the samples with and without head-down tilting was observed. Even when the planned PS medial and/or rostro-caudal rotational angle was relatively large and difficult to reproduce manually, IMU-support helped maintain the PS trajectory accuracy and positioning safety. IMU assist in PS placement was more beneficial, especially for larger rostro-caudal and/or medial rotational pedicle angles.