The objective of this research was to design, implement, use and evaluate a human-machine touch interface for driving an electric wheelchair. This new interface allows control of the electric wheelchair with touch screen smartphone technologies (android, IOS or Windows phone). In addition, an intelligent calibration algorithm, based on the neural network (NN), is implemented in this interface to overcome manoeuvring problems, namely the inability to move correctly in one or more directions.

Our work offers a touch screen human-machine interface for controlling the electric wheelchair. Three patients aged 15 to 66 years participated in our experiment. They were asked to control the electric wheelchair using two types of interface (the standard mechanical joystick and the intelligent touch screen joystick we offer) according to the established protocol. This allowed us to extract the benefits of using the calibrated touch interface to move the wheelchair. The advantage of this control is that it allows the patient to drive the wheelchair with a precise variable speed like a traditional motorized joystick with less effort.

We have set up a two-way communication between the tablet and the control system. This allows us not only to control the wheelchair, but also to detect and treat communication errors as a sudden disconnection. In this case, the wheelchair stops and waits for the user to take further action. The use of Wi-Fi has many advantages since it is a lighting solution and the user can even park the wheelchair in his room when he is lying in bed.

We can notice that users are faster with the intelligent touch joystick than with the mechanical joystick. This is because less muscular effort is required on the first than on a mechanical joystick, the user only uses his finger to move the chair. Moreover, the advantage of using the neural network ensures increased speed and stability of movement even if the user has the impression that his finger is in the wrong direction on the interface and that his speed is reduced because of his pathology. In fact, one of the expected results was that the tablet maintained constant speed longer than the standard mechanical joystick. This can be explained by the fact that it is easier with the tablet to maintain a maximum speed while correcting the trajectory deviation with a minimum effort.

In this article, we explored an intelligent touchscreen solution that will address the problems faced by our target users. We conducted a thorough research on the types of existing control interfaces. We found that most of these interfaces use a discrete control of speed and direction. After testing it with a group of disabled users, we noticed the importance of analogic control. As part of our iterative design approach, we designed a prototype that provides this analogic control and in which we have implemented a calibration algorithm based on the neural networks algorithm.