Development of Foot Displacement Detection Algorithm for Power Wheelchair Footplate Pressure and Positioning.

ABSTRACT

Inadvertent lower extremity displacement (ILED) puts the feet of power wheelchair (PWC) users at great risk of traumatic injury. Because disabled individuals may not be aware of a mis-positioned foot, a real-time system for notification can reduce the risk of injury. To test this concept, we developed a prototype system called FootSafe, capable of real-time detection and classification of foot position. The FootSafe system used an array of force-sensing resistors to monitor foot pressures on the PWC footplate. Data were transmitted via Bluetooth to an iOS app which ran a classifier algorithm to notify the user of ILED. In a pilot trial, FootSafe was tested with seven participants seated in a PWC. Data collected from this trial were used to test the accuracy of classification algorithms. A custom figure of merit (FOM) was created to balance the risk of missed positive and false positive. While a machine-learning algorithm (K nearest neighbors, FOM=0.78) outperformed simpler methods, the simplest algorithm, mean footplate pressure, performed similarly (FOM=0.62). In a real-time classification task, these results suggest that foot position can be estimated using relatively few force sensors and simple algorithms running on mobile hardware.Clinical Relevance- Foot collisions or dragging are severe or life-threatening injuries for people with spinal cord injuries. The FootSafe sensor, iOS app, and classifier algorithm can warn the user of a mis-positioned foot to reduce the incidence of injury.