Recognition of Navigation Commands for a Smart Walker Through Force Sensors

Smart Walkers are robotic devices that may be used to improve physical stability and sensorial support for people with lower-limb weakness or poor balance. Even though such devices may offer support to people who cannot safely use conventional walkers, their interaction strategy with the user still needs to be improved. In this context, this work presents a strategy to obtain navigation commands of a smart walker based on multi-axial force sensors. It also shows how to obtain the user's motion intention from the interaction between his/her arms and the robotic walker. An admittance controller and an adaptive filter are used to obtain the user's cadence. To validate the proposed strategy, a lemniscate curve marked on the floor is used to be followed by both the user and the smart walker. The parameters of the admittance controller are adjusted to find the more suitable values that allow a natural locomotion. As a result, the user can command the smart walker and establish a natural speed for his/her locomotion (around 0.3 m/s) using the admittance control.