Skip to main content
Research topic

Gait Analysis and Footstep Recognition Research

Research using floor sensors and machine learning to study gait, walking patterns, and footstep recognition for healthcare and biometric-security applications.

Reading movement through floor sensors

This research studies walking as a spatio-temporal signal, a signal that varies across the space under the foot and across time, recorded by a floor-based footprint sensor. The work examines how changes in gait and individual footstep patterns can be represented and classified directly from sensor measurements.

The University of Manchester thesis connects this sensing work to two research settings: changes in gait during cognitively demanding tasks and biometric verification based on footstep patterns. The healthcare study included dozens of participants, and the security study included over a hundred.

Different models answer different questions

Temporal gait studies compare signal features and machine-learning methods for recognizing walking activities. Footstep-recognition work instead evaluates whether models can distinguish people from their recorded steps. Convolutional and deep residual networks explore learned spatial and temporal representations for these tasks.

From the 2016 poster at the EuroScience Open Forum (ESOF) through later publications, the doctoral thesis, and research seminars, this work progressed from a smart-carpet demonstration to detailed gait classification and biometric studies. Each publication page describes its methods and results.

Research applications and significance

Gait sensing supports research on mobility and healthcare, while footstep recognition advances biometric-security research. These applications bring distinct objectives and evaluation approaches to movement analysis.

Because the floor itself does the sensing, nobody has to wear a device, which suits both unobtrusive health monitoring and biometric verification. The thesis and related studies bring together participants, sensors, signals, and evaluation procedures across gait analysis and footstep recognition.

Topics