wearable inertial sensor for jump performance analysis · wearable inertial sensor for jump...
TRANSCRIPT
Wearable Inertial Sensor for Jump Performance Analysis
B. Milosevic, E. Farella
E3DA – FBK, Trento, Italy
WearSys’15 – Firenze, 18/05/15
With the support from:
Overview
• Jump performance is frequently used to monitor training progress in athletes or injured patients.
• Measurements are typically captured in clinic with accurate but expensive instrumentation.
• We propose the use of a versatile low-cost wearable device
– Equipped with inertial sensors
– On-board estimation of jump height
– Easily employed at home
Wearable Devices
Wearable devices: we all know they are having a huge diffusion in consumer applications
• Pros: low cost, ease of use, unobtrusive
• Cons: not highly accurate and not validated
http://www.getgrok.com/2013/01/a-comparative-review-28-days-with-the-fitbit-one-jawbone-up-nike-fuelband-and-bodymedia-link/
WD
Healthcare Applications
• Medical applications need high accuracy and clinical validation, thus they still suffer a gap in the inclusion of new technologies and wearable devices [Lee14].
• A new trend is the diffusion of at home and personalized therapy practices, which leverage the use of existing technologies (wearables, mobile) [Pantelopoulos10].
– e.g. Exergaming platforms based on wearable sensors or Kinect
• The research and development of such solutions is a challenging topic
– Low cost and accurate sensing solutions
– Patient interfaces and biofeedback
– Web interface for the clinician
HA
Jump Analysis
• Jumps are extensively used to evaluate the physical condition of patients and athletes [Kale09, Herbst15].
• Accurate and expensive (1k÷10k$) technology is used for jump evaluation in clinics:
– Multi-camera motion capture systems
– Force plates [Bosco83]
– Inertial-based systems, e.g the MyoTest [Nuzzo11, Choukou14]
• Alternative solutions have been studied using wearable inertial sensors [Picerno 11] or smartphones [Balsalobre14].
JA
Our Solution
Wearable device for the evaluation of jump performance
• Self-contained wearable device with on-board processing
• Low-cost (≈100$) and easy to use
• Analysis of two jump types:
– Counter-Movement Jump (CMJ) used for explosive force assessment
– PlyoMetric Jump (PMJ) used for reactivity assessment
• Validated against a commercial clinical device
System Description
Wearable low cost sensor node equipped with:
• ARM Cortex M3 MCU – STM32F1, 78MHz
• Inertial Measurement Unit (9-Axis IMU by Invensense, 300Hz)
• Bluetooth
• Button, LED, buzzer
Counter-Movement Jump (CMJ)
CMJ: one jump performed with a counter-movement starting from the upright still position
• 5 jump phases: counter-movement, take-off, flight, landing, recovery
• Jump height is the performance metric
CMJ Height Estimation
• Low pass filter: 20 sample mean filter
• Features: acceleration variance to estimate if the user is still/in motion
– To initialize the algorithm the user is required to stand still before a jump
• The orientation of the device is constantly updated
– Initialized when still with accelerometer
– Updated integrating the gyro during jump
• The vertical inertial acceleration aG is obtained by rotating the measured acceleration and subtracting g
• Jump phases are identified by thresholds: when in flight, aG is set to -g
• aG is double integrated to estimate the jump height
Algorithm output:
CMJ Height Estimation
PlyoMetric Jump (PMJ)
PMJ: sequence of 4 jumps performed in rapid succession
• The mean height of the last three jumps and the total contact time are used as performance metric
• Same segmentation and estimation algorithm as for CMJ
• At each landing interference acceleration peak due to impact on the floor– Difficult to filter out, hence we apply a correction step at each jump
PMJ Height EstimationAlgorithm output
Experimental Validation
• The proposed system was validated against the MyoTest Pro 2
– Clinically validated wearable device for jump assessment
• Dataset of jumps collected while wearing both devices:
– 40 healthy subjects (32/8 male/female)
– Different fitness levels (from sedentary to trained athletes)
– Each performed 3 CMJs and 2 PMJs
– Total: 120 CMJs and 80 PMJs
Results
CMJ:
• Height: mean difference 0.7cm, max: 1.6cm (2.6%)
Results
PMJ:
• Height: mean difference: 0.6cm, max: 1.5cm (1.9%)
• Contact time: mean difference 23ms, max: 33ms (9%).
Conclusion
• This work presented a wearable system for the evaluation of jump performance
– Low cost solution targeted for autonomous use at home
– CMJ and PMJ jumps analysis
– Validation against a clinical device on 200 jumps
• The results show that our system is accurate
– Mean error: CMJ = 0.7cm, PMJ = 0.6cm
• Future development:
– Integration in rehabilitation practices at home
– Evaluation and test for use at home
References[Lee14] J.-M. Lee et al., Validity of consumer-based physical activity monitors. Medicine and science in sports and exercise, 2014.
[Pantelopoulos 10] A. Pantelopoulos and N. Bourbakis. A survey on wearable sensor-based systems for health monitoring and prognosis. Systems, Man, and Cybernetics, Part C: Apps. and Reviews, 40(1):1–12, Jan 2010.
[Kale09] M. Kale et al., Relationships among jumping performances and sprint parameters during maximum speed phase in sprinters. The Journal of Strength & Conditioning Research, 23(8):2272–2279, 2009.
[Herbst15] E. Herbst et al., Functional assessments for decision-making regarding return to sports following acl reconstruction. part II: clinical application of a new test battery. Knee Surgery, Sports Traumatology, Arthroscopy, pages 1–9, 2015.
[Bosco83] C. Bosco et al., A simple method for measurement of mechanical power in jumping. European journal of applied physiology and occupational physiology, 50(2):273–282, 1983.
[Nuzzo11] J. L. Nuzzo et al., The reliability of three devices used for measuring vertical jump height. The Journal of Strength & Conditioning Research, 25(9):2580–2590, 2011.
[Choukou 14] M.-A. Choukou et al., Reliability and validity of an accelerometric system for assessing vertical jumping performance. Biology of Sport, 31(1):55, 2014.
[Picerno 11] P. Picerno et al., Countermovement jump performance assessment using a wearable 3d inertial measurement unit. Journal of sports sciences, 29(2):139–146, 2011.
[Balsalobre14] C. Balsalobre-Fernandez et al., The validity and reliability of an iPhone app for measuring vertical jump performance. Journal of sports sciences, pp.1–6, 2014