An Evaluation of Eye-Foot Input for Target Acquisitions

The multimodal interaction technique that combines eye and foot input not only provides a great opportunity for the user with busy hands or hand disabilities to interaction with computer, but can also overcome the drawbacks when using dwell time to solve the “Midas Problem” of gaze input. However, the user’s capability of eye-foot coordination was still unclear. At the same time, the human performance in the basic task of target acquisitions by eye-foot input was also uncertain, and especially a proper performance model was lacking. Motivated by this situation, an eye pointing and foot tapping task experiment had been carried out to fill these gaps. A low-cost eye tracker and a USB foot pedal switcher were used as the input devices from different modalities. The experimental results indicated that the user was soon able to coordinate her/his foot with the eyes for target acquisitions, and that the user could respond fast to tap the foot pedal to finish a task trial in the level of 600 ms. The main performance measures of eye movement time (EMT) and eye pointing time (EPT) under the eye-foot multimodal input condition were significantly increased with the increase of saccadic amplitude A and/or the decrease of target width (size) W, and vice versa. Regression analysis shown that the \(ID_{eye}\) model was more suitable than the standard Fitts’ law to model the human performance in this multimodal interaction context.