Assessing the Accuracy of an Algorithm for the Estimation of Spatial Gait Parameters Using Inertial Measurement Units

Abstract

We have reviewed and assessed the reliability of a dead reckoning and drift correction algorithm for the estimation of spatial gait parameters using Inertial Measurement Units (IMUs). In particular, we are interested in obtaining accurate stride lengths measurements in order to assess the effects of a wearable haptic cueing device designed to assist people with neurological health conditions during gait rehabilitation. To assess the accuracy of the stride lengths estimates, we compared the output of the algorithm with measurements obtained using a high-end marker-based motion capture system, here adopted as a gold standard. In addition, we introduce an alternative method for detecting initial impact events (i.e. the instants at which one foot contacts the ground, here used for delimiting strides) using accelerometer data. Our method, based on a kinematic feature we named 'jerkage', has proved more robust than detecting peaks on raw accelerometer data. We argue that the resulting measurements of stride lengths are accurate enough to provide trend data needed to support worthwhile gait rehabilitation applications. This approach has potential to assist physiotherapists and patients without access to fully-equipped movement labs. More specifically, it has applications for collecting data to guide and assess gait rehabilitation both outdoors and at home.

Gait Analysis in a Box: A System Based on Magnetometer-Free IMUs or Clusters of Optical Markers with Automatic Event Detection

“…One is the exploitation of kinematic constraints; for example, using the knowledge that the elbow joint does not permit abduction/adduction, or that the shoulder cannot attain more than 180 degrees of abduction or flexion [45,46,[49][50][51][52]. Another approach could be to use zero-velocity updates [54] or dead reckoning [55] algorithms that reset the errors when detecting zero-velocity periods during the gait. If the capture time were prolonged, both the MH-IMU and the MH-OPT configurations could be used for the for rehabilitation itself.…”

“…The most straightforward approach for resolving the drift errors is to limit the capture time, which is appropriate when only a short period is required to execute the movements being investigated [38]. For long-term captures, one approach is to use the aforementioned kinematic constraints, since they limit the drift artefact [45,46,[49][50][51][52]; another possible method is to use zero-velocity updates [54] or dead reckoning [55] algorithms, which reset integration and acceleration errors when detecting zero-velocity periods during the gait.…”

Haptics for the Development of Fundamental Rhythm Skills, Including Multi-limb Coordination

“…Theoretical insights came from research in music perception by Bamberger [12], Lerhahl and Jackendoff [13], and others, as well as from work in ethnomusicology by Arom [14]. Once various prototype versions of the bracelets were built [2,3,11,15,16], research from cognitive science, particularly theories of human biological entrainment and neural resonance theory, proved invaluable in understanding key aspects of how humans interact with the bracelets.…”

“…2 In the present chapter, the term "vibrotactile" is often used as a noun to mean "vibrotactile actuator". 3 A now discontinued Wi-Fi solution.…”

“…Interestingly, the Haptic Bracelets have also found applications in the digital health domain, particularly in rehabilitation for sufferers from a range of movementrelated neurological conditions including stroke, Parkinson's, Huntingdon's and brain trauma [1][2][3][4]. However, this is mostly outside of the scope of this chapter.…”