Emergency tractor shut-off using a voice command system

Being isolated and far from aid, farmers need better emergency response options in case of an accident. Currently, farmers can be left in the field for hours after an incident before they are discovered. The primary objective of this project was to develop a prototype system that will respond to voice commands to stop the tractor in the case of an emergency. The challenge is to make voice commands recognizable from a distance while tractor background noise is present. Specifically, I developed a speech recognition system using MATLAB and the neural network toolbox and audio recordings of tractor and voice commands. Recognized commands will be used to cut power to the tractor. The system will be tested on a tractor and when a voice command of STOP is issued, the tractor will be shut-off. The results of this research will lead to future studies to increase the complexity of the interface to accomplish tasks such as automatically sending a help signal when the tractor operator is not responding or has not returned to their seat within a specific period of time. The system may also be the basis for a more comprehensive system for voice commands that perform other safety related tasks that include the need for emergency stops and response.