Abstract

In this paper, a Hidden Markov Model (HMM) speech recognition system which is based on Field Programmable Gate Array (FPGA) is designed. It introduces the principle of speech recognition algorithm and deduces the hardware frameworks accordingly. In terms of HMM recognition module, the conventional Viterbi algorithm has been improved and recognition speed has been increased. The core part of the hardware is EP2S60F1020C3 FPGA chip. The experimental result of this system shows that the speech recognition accuracy reaches 94% when ten numbers are being recognized, and the average recognition time is 0.669s.

This work is licensed under a Creative Commons Attribution 4.0 License.