The Application Of VQ Using PSO Speech Recognition Based On DHMM

Speech recognition technology is a rapidly growing high-tech industry, which has been applied to many fields. It will become an important feature of future computer and its study has important significance.As a statistical model of the speech signal processing, Hidden markov model (HMM) is widely applied in many areas. In isolated word speech recognition, we must quantify the feature vector. Vector quantization (VQ) is a data compression technology, which can compress the information without distortion. Particle swarm optimization (PSO) algorithm is a global optimization algorithm which is easy to implement. Artificial immunity algorithm is a swarm intelligence algorithm, which comes from the simulation of nature biology community's immune system then achieves immune memory and self-regulation faction. It has become another research hotspot of artificial intelligence algorithms. In this paper, the in-depth research is done to the speech recognition based on DHMM by PSO. The main achievements are as follows:1,LBG algorithm trained by particle swarm optimization algorithm is proposed in order to overcome shortage of LBG algorithm which heavily depends on the initial code book, and easy to fall into local optimum, and has weak adaptive capacity. This algorithm is used to the isolated word speech recognition based on DHMM. Experimental results demonstrate that the algorithm has good performance, which is better than the speech recognition based on DHMM by LBG algorithm in the recognition rate.2,Point at the problem of premature convergence of PSO algorithm, it introduces artificial immune algorithm. In this paper, immune memory, self-regulation and immune vaccination mechanisms of immune system are involved into original particle swarm optimizer, and the particle swarm optimization algorithm with immunity are proposed. It is tested using four typical test functions. The results show that immune particle swarm optimization algorithm can improve the abilities of seeking the global excellent result and evolution speed.3,It combines immune particle swarm optimization algorithm and LBG algorithm for clustering. LBG algorithm trained by immune particle swarm optimization algorithm is proposed and used to the isolated word speech recognition based on DHMM. The experiment shows that the improved system has better recognition rate and robustness, compared with the result of the speech recognition based on DHMM by PSO algorithm.