Politecnico di Torino - Corso Duca degli Abruzzi, 24 - 10129 Torino, ITALY

+39 011 090 6100 info@tech-share.it

Procedure addressed to patients with motor disabilities to choose a command

Machine learningModel predictive controlMotor disabilitiesPersonalized assistance


The invention proposes an innovative solution for choosing a command through a graphic interface, with the specific aim of helping a patient during the movement of a cursor on a graphic interface in such a way as to enrich the current human-computer interface. This invention has valuable functions for patients suffering from neuro-degenerative diseases or injuries to the first or second motor neuron, suffering from paralysis or paresis.

Technical features

Due to some pathologies, the trajectories described by the cursor, in the absence of an aid solution, are not regular and this complicates the use of a graphic interface. To address this, the proposed patented solution provides two innovative features: its support action is based on a customisation process and is formulated in a predictive manner, i.e., it is based on the prediction of the target desired by the user while the user himself controls the cursor. Using automatic learning and control methodologies, this solution predicts which action the patient wants to undertake, assisting him/her in selecting, thus costing the patient as little effort as possible. To the knowledge of the inventors, none of the already existing solutions provide means of assisting patients with motor disabilities with a personalized method, based on innovative machine learning techniques, with an automatic control scheme, based on model predictive control and therefore such as to provide an action of predictive support, and with a similar degree of portability. The invention allows the patient to autonomously perform daily actions such as turning on the lights, adjusting the temperature in the room, etc.

Possible Applications

  • Home automation for people with motor disabilities who use human-machine interfaces


  • Methodology that can be used in integration with the software of various human-computer interface devices;
  • Customizable solution.