The department of Biomechanics, Kinesiology and Computer Science in Sport is partner in the project IntelliGait that is funded by Science Call 2015 of the NÖ Forschungs- und Bildungsges.m.b.H. The leading agency of this project is the FH ST. Pölten (Dr. Brian Horsak).

Instrumented gait analysis has become a crucial assessment tool in clinics, hospitals and rehabilitation centers to understand various (pathological) human movement patterns. Such three dimensional gait analysis techniques use motion capturing techniques and the measurement of ground reaction forces (GRF) via force plates to estimate joint kinematics and kinetics. The GRF represent the most commonly used biomechanical signals to analyze human gait, because the necessary equipment is affordable and the process of data collection is simple, therefore allowing for high patient throughput. However, apart from the simplicity of GRF data capturing, physical therapists and clinicians are often faced with a vast amount of GRF data and the need to interpret these data correctly. Due to the absence of automated analysis methods, the inspection of the data is performed visually, which is time-consuming and leads to subjective assessments.

Automatic analysis methods bear the potential to provide objective measurements and assessments of the signals. Recently, different methods have been introduced for the automatic classification of gait patterns. Existing methods, however, usually focus only on one specific functional deficit and are not applicable to the broad range of deficits that occur in clinical praxis. Furthermore, existing techniques are developed on rather small (artificial) datasets that do not reflect the complexity of data captured in clinical praxis. A still unsolved problem is the automatic generation of a generally applicable model for normal behavior that takes different parameters as walking speed, gender, body height, etc. into account and helps to distinguish normal behavior from abnormal behavior.

In this project a research partner maintains a large dataset of GRF data that contains measurements of patients with different ages, weights, and a broad range of functional deficits as well as detailed clinical diagnoses for each patient. This exceptional large-scale dataset bears the potential to develop novel powerful analysis techniques for gait patterns that fulfill the strong requirements of therapists and clinicians and are thus applicable in clinical daily routine. The novel methods should in future support the expert in detecting pathological or abnormal behavior, in making medical diagnoses and in the assessment of rehabilitation and training progress.