SELF-DRIVING CAR CONTROL MODEL EXTENSION WITH VOICE COMMANDS CONTROL

The developed experimental models of self-driving car demonstrate high accuracy (about 99%), but there is still a need to improve the overall safety of real traffic on the roads. Especially when there are people in the car, if the autonomous vehicle loses control on the road, the quick intervention to prevent a possible crash or a more serious road accident can be only through voice com-mands between the person and the execution control devices of self-driving car. The existing self-driving car control models are mainly based on incoming from mounted on the autonomous vehicle video cameras information, processed from deep learning neural networks and artificial intelligence. This paper proposes to extend these models with voice commands recognition, spoken by a person in the car, in order to correct the self-driving car movement, to prevent possi-ble traffic accidents, and therefore to increase traffic safety. For this purpose deep learning neural network with artificial intelligence is developed to recog-nize the spoken by the person voice commands, which can be interpreted by the executive control devices of the autonomous vehicle. The presented results from simulation tests show the ability of the proposed extended self-driving car control model to correct with voice commands the self-driving car motion on the road leading to essential increase of the safety of traffic.