Implementation of speech recognition on MCS51 microcontroller for controlling wheelchair

This paper describes about implementation of speech recognition on microcontroller. The microcontroller used in this system is ATMEL AT89C51RC microcontroller which is one of the MCS51 family microcontrollers. Speech recognition system is implemented to recognize the word used as the command for controlling movement of a wheelchair. There are two approaches used to recognize the speech signal. The first approach is linear predictive coding combined with Euclidean squared distance. LPC is used as the feature extraction method and Euclidean squared distance is used as the recognition method. The second approach is hidden Markov model, which is used to build reference model of the words and also used as the recognition method. Feature extraction method used in the second approach is a simple segmentation and centroid value. Both approaches work on time domain. Two DC motors are used as the actuator for driving the wheelchair. Both DC motors are controlled by ATMEL AT89C52 microcontroller and using a simple open loop control system. Experiments were done to analyze performance of both approaches. Each approach has advantages and disadvantages. The highest average recognition rate that can be achieved using LPC-Euclidean squared distance approach was 78.57%. The highest average recognition rate that can be achieved using HMM-segmentation and centroid approach was only 32.86%.