As the EU population is getting older, the number of Europeans with disabilities is rising significantly. Among them, a consistent percentage is represented by persons with severe mobility impairment, making them dependent of wheelchairs. This paper proposes a human wheelchair command interface based on head movement tracking. The head movements are tracked using an artificial vision-based system. The wheelchair user can select between two control modes in a first control mode, simple commands (as left, right, forward, stop) are given; in the second control mode, the wheelchair movements are correlated with both amplitude and speed of head movement. For the both control modes, the wheelchair movement commands are assisted by a 3D cartography analysis system, having the role to prevent the collision with near-by obstacles.