The paper focuses on the field problems encountered in voice control of the motorized wheelchairs and goes a step further to propose the solution that uses voice control in co-ordination with the 2D space mapped control scheme, called the voice-cum-auto steer control scheme. The paper describes various design aspects of the wheelchair including the proximity sensing for maneuverability and situational awareness, automated guideway control and voice-cum-auto steer control. The voice control is through the feature based, language independent but speaker dependent, isolated word recognition system. The proximity sensing for collision avoidance is based on the principle of transmitting coded IR rays and analyzing the reflected signal from the obstacle. The output of these sensors, along with position/velocity sensors, are the inputs to the fuzzy control algorithm. A 2D map of the space is fed in bitmap form that acts as the reference to all the decisions taken by the motor controller. A method to generate these maps by a walk-through procedure during the training of the wheelchair is also mentioned.
展开▼
