Autonomous gait speed estimation using 24GHz FMCW radar technology

Gait speed is an important health indicator and therefore can be used as a way to understand functional capacity of older adults. Continuous gait speed measurement is required to detect deviations from normal speed, which may be a symptom of various cognitive or other problems. Among older adults, gait speed reduction from a normal level correlates with an increased risk of falls hence knowing about changes in gait speed could help support falls prevention. We present an autonomous gait speed estimation system (A-GSE) that employs a 24GHz radar sensor which works in frequency modulated continuous wave (FMCW) mode to track the gait speed of a person during walking. The A-GSE system was tested with 14 healthy young adult participants in a walking experiment with three levels of walking speed: low, normal and fast. A-GSE results were compared to ground truth values that were simultaneously captured by a GAITRite mat. Results show a difference between the low, normal and fast gait speed classes. Accuracy calculation shows the best performance for tracking fast gait speeds with an average accuracy of ~ 86%. Overall, this research suggests that 24GHz radar technology for autonomous gait speed measurement is a promising approach but requires further work to adapt it to home-type environments.