Filtrado Adaptativo de Componentes Involuntarias en Marcha Asistida por Andador para Detección de Intenciones

Abstract

The advances in technology have allowed the use of sensors, processors and actuators in the context of the mobility assistive devices for disabled people, improving their usability and safety. Considering walkers, the introduction of such technologies has created a group of devices called Smart Walkers. Among other devices, in the Bioengineering Group at CSIC, a smart walker was developed in the framework of the Simbiosis Project. This work presents a method of adaptive filtering designed for the suppression of involuntary force components from the interaction forces between user’s upper limbs and walker. This process is based on the selective attenuation of components related to the user’s trunk oscillations in assisted gait. For that purpose, gait cadence is estimated in real-time from a ultrasonic measurement system using the Weighted-Frequency Fourier Linear Combiner (WFLC). The cadence is, then, used to adjust a adaptive notch filter buld upon the Fourier Linear Combiner (FLC) algorithm, that filters in real-time the force data acquired by the sensors installed on the device’s handles. The proposed methodology offered a real-time cancelation of about 80% of the forces components’ energy at the desired frequencies. The output of the presented algorithm will be used as inputs for the controller that will be designed to command the motorized walker.