Biomechanics is the science related to the study of the internal and external forces acting on the human body and the effects produced by them. In particular, biomechanics is the study of the locomotion of living organisms and of the forces causing those movements. This analysis is frequently used in both clinical and sporting practice by clinicians and can play a crucial role in athletes’ performance enhancement, injury prevention and effective patient rehabilitation. More specifically, in the latter case, it is essential to track patient progress and, consequently, to tailor patient-oriented rehabilitation programmes, through the accurate assessment of human motion during the performance of clinically defined tasks. A common example of the technology regarded as the ‘gold standard’ of quantitative movement-based analysis is given by camera-based motion analysis systems (such as the ones provided by Vicon, Optitrack, or Codamotion) which, during formal gait analysis by rehabilitation professionals, can help to ascertain measurements of temporal (time) and spatial characteristics associated with gait parameters. This enables clinicians to identify gait deviations in paediatric and amputee populations, screening elderly people for risk of falling, to objectively monitor patient’s progress, and to help determine the efficacy of surgical and therapy interventions. [caption id="attachment_34300" align="alignright" width="277"]Fig 1: Tyndall Wireless Inertial Measurement Unit (WIMU) [3]. Fig 1: Tyndall Wireless Inertial Measurement Unit (WIMU) [3]. (CLICK TO ENLARGE)[/caption]Another approach, which has been explored as a more practical and viable solution to biomechanical motion capture and monitoring in sporting and patient groups, is through the use of small-size low-cost wearable inertial sensors [1]. Inertial sensors are mainly used in devices to measure velocity, orientation and gravitational force. They are massively diffused in a great number of applications, such as industry quality control, robotics, navigation systems, sports, augmented reality systems and so on [2]. Biomechanics, in particular, has achieved significant progress from the adoption of this technology. More specifically, with regards to gait analysis, accelerometers and gyroscopes have been used. They have been worn on the lower limbs to obtain gait parameters, which can be derived by the integration of angular acceleration or angular velocity after the correct identification of the beginning and the end of each gait cycle.

Research aims and methodology

[caption id="attachment_34301" align="alignright" width="300"]Fig 2: Walking gait phases Fig 2: Walking gait phases (CLICK TO ENLARGE)[/caption] The aim of the present work is, therefore, to implement a wireless, portable, easy-to-use system, with two sensors per leg, suitable for free-living environments and able to provide a complete biomechanics assessment (generated on a report) without the constraints of a laboratory. The system evaluates both gait and joint range of motion for several scripted activities that are typically performed during the rehabilitation process and also in a daily routine. The system, tested with both healthy and impaired athletes, provides the possibility to evaluate gait during rehabilitation and to identify gait abnormalities. The derived outcome can be analysed by clinicians and sport scientists to study the patient’s overall condition and provide accurate medical feedback. Validation for the lower limbs using state of the art, camera-based motion capture has been also carried out. [caption id="attachment_34302" align="alignright" width="300"]Fig 3: Gait characteristic instants from gyroscope signal on the sagittal plane Fig 3: Gait characteristic instants from gyroscope signal on the sagittal plane (CLICK TO ENLARGE)[/caption] In terms of methodology, the parameters selected for assessment were:
  • Temporal events: toe-offs, heel-strikes, mid-stance;
  • Temporal intervals: gait cycle duration, stance phase, swing phase, single and double support, cadence (or step rate), number of cycles, swing symmetry;
  • Spatial parameters: stride length, stride velocity (or speed), peak angular velocity, shank clearance; and,
  • Knee joint range of motion.
For each of those parameters, it is possible to calculate minimum, maximum, mean and median standard-deviation values and extrapolate the related dimensionless variability (or coefficient of variation, e.g. the ratio between the standard deviation and the mean-value). Figures 2 and 3 give an example of some of the temporal parameters assessed and angular rate signal collected on the sagittal plane of the shank.

Procedure and results

The system consists of two Tyndall Wireless Inertial Measurement Units (WIMUs) [3] per leg with 3D accelerometer/gyro (@ 250 Hz) and BLE (or SD cards), as shown in Figure 1. The developed technology is currently at the sixth level (technology demonstration) of the Technology Readiness Level (TRL) scale, where the engineering-scale system has been tested in a relevant environment and is ready for demonstrations. [caption id="attachment_34303" align="alignright" width="300"]Fig 4 Fig 4: WIMU and camera reference comparison (CLICK TO ENLARGE)[/caption] The system has been tested with both healthy and impaired athletes. WIMUs have been attached to the anterior tibia, 10cm below the tibial tuberosity, and to the lateral thigh, 15cm above the tibial tuberosity using sticky tape. Algorithms are implemented in Matlab and the scenarios considered (walking, half squat, hamstring curl and lunges) simulate a free-living environment and exercises performed in a rehabilitation procedure. In the walking scenario, the subjects stand on a treadmill, which is operated at different speeds (3-4-6 km/h) for approximately one minute per test. In the half-squat scenario, the subjects stand with the feet at shoulder’s distance apart and arms crossed on the chest. Keeping the chest lifted, the hips are lowered about 10 inches, planting the weight in the heels. The body is then brought back up to standing by pushing through the heels. In the hamstring curl scenario, the subjects stand and bend the affected knee (or the knee in the dominant leg, as per healthy subjects) raising the heel toward the ceiling as far as possible without pain, relaxing the leg after each repetition. [caption id="attachment_34304" align="alignright" width="300"]Fig 5 Fig 5: WIMU vs camera error estimation (CLICK TO ENLARGE)[/caption] In the lunge scenario, the subjects stand with feet at shoulder's width apart, spine long and straight, shoulders back, gaze forward, and step forward with the affected leg into a wide stance (about one leg's distance between feet) while maintaining spine alignment. The hips are lowered until both knees are bent at approximately a 90-degree angle. Finally, the subject pushes back up to starting position by keeping the weight in the heels. A high-speed Basler camera (@ 100 Hz) [4], adopted in conjunction with active markers, has been used as a reference for the validation. The developed system has been firstly tested against the video reference (Basler camera). Knee joint angles have been estimated with both technologies for all the scenarios (an example for the hamstring curl is shown in Figure 4). Finally, all the gait spatio-temporal parameters mentioned in the previous section have been calculated for the walking scenario (at 3km/h, 4km/h and 6 km/h) for both legs of the impaired and unimpaired subjects by adopting the inertial data. Results are summarised in Figure 6.

Wearable inertial system

[caption id="attachment_34305" align="alignright" width="300"]Fig 6: Gait spatio-temporal parameters estimated for impaired/unimpaired subjects during walking scenario (3km/h, 4km/h and 6km/h) Fig 6: Gait spatio-temporal parameters estimated for impaired/unimpaired subjects during walking scenario (3km/h, 4km/h and 6km/h) (CLICK TO ENLARGE)[/caption] This work presents a wearable inertial system for the implementation of a complete portable wireless lower-limbs analysis system. Overall, results present good repeatability and the accuracy is comparable with the state of the art. Moreover, detection of atypical movement characteristics was possible by comparing performance and differences in the two legs. Athletes/patients can be monitored throughout their complete rehabilitation in order to gather response to the therapeutic treatment. This work was a phase-one feasibility study designed to experimentally validate the system implemented in Tyndall for lower-limbs motion capture. In order to further validate the drawn conclusions in statistical terms, additional clinical trials, with larger and homogeneous populations, are needed. This feasibility study represent a prerequisite to a larger cross-sectional/cohort trial that will assess sport-performance analysis and movement pattern alterations detection in people following knee injuries pathologies through wearable inertial sensing technology. As reported in the Engineers Journal [5], through the gateone project, Tyndall is collaborating with German-Serbian software SME, Nissatech Innovation Center, and trials are currently starting in a Budapest hospital. However, the present feasibility study has already proved that inertial sensors can be used for a quantitative assessment of knee-joint mobility, and gait mechanics in ambulatory or free-living environments during the rehabilitation programme of injured athletes.

Acknowledgements:

This work has emanated from research supported in part by a research grant from Science Foundation Ireland and is co-funded under the European Regional Development Fund under Grant Number 13/RC/2077 – CONNECT. The authors also acknowledge the support of the EU H2020 program through funding of the gateone project. Authors: [caption id="attachment_33878" align="alignright" width="300"]tyndall-smart-knee Thomas Healy, commercial and research project manager and John Barton, staff research scientist at Tyndall National Institute[/caption] John Barton, Salvatore Tedesco, Thomas Healy, Brendan O’Flynn References:
  1. D.T.P. Fong, Y.Y. Chan, ‘The use of wearable inertial motion sensors in human lower limb biomechanics studies: A systematic review,’ Sensors, vol. 10, no. 12, pp. 11556-11565, 2010.
  2. N. Ahmad, R. Ariffin, N.M. Khairi, V. Kasi, ‘Reviews on various inertial measurement unit (IMU) sensor applications,’ Int J Signal Processing Systems, vol. 1, no. 2, pp. 256-262, 2013.
  3. [1] S. Tedesco, A. Urru, J. Peckitt, B. O’Flynn, ‘Wearable inertial sensors as a tool for quantitative assessment of progress during rehabilitation,’ Int Conf on Global Health Challenges, October 8-13, 2016, Venice, Italy, pp. 56-59.
  4. http://www.baslerweb.com/en/products
  5. http://www.engineersjournal.ie/2016/12/19/tyndall-national-institute-smart-knee-device/