Adjustable Pedals by Gesture Recognition: A Novel Approach for User Interface in Automotive
Muhammad Haroon Yousaf, Hafiz Adnan Habib · 2012
This paper proposes the enhancement in the design of adjustable pedals in automotive by incorporating gesture recognition methodologies. The design addresses the three primary controls: Accelerator, Brake and Clutch. Mechanical input of driver's feet is deduced at location variant adjustable pedals in terms of feet gestures. Research work aims vision based technology for the feet gesture recognition. Machine vision algorithms are developed to recognize feet gestures. Low level features based upon background subtraction, connected component analysis, optical flow and motion trajectories are computed to infer spatial, symbolic, pathic and emotional information of feet gestures. Classification of motion trajectories is done by logic based analysis. Accurate classification of feet gestures proves a workable adjustable pedals framework based upon feet gestures. Research work recommends the substitution of conventional fixed pedals with user-defined gesture controlled adjustable pedals facilitating the drivers in terms of ease, flexibility, ergonomics and cost-effectiveness.