─ This research paper proposes a recently developed new variant of Particle Swarm Optimization (PSO) called Accelerated Particle Swarm Optimization (APSO) in speech enhancement application. Accelerated Particle Swarm Optimization technique is developed by Xin she Yang in 2010. APSO is simpler to implement and it has faster convergence when compared to the standard PSO (SPSO) algorithm. Hence as an alternative to SPSO based speech enhancement algorithm, APSO is introduced to speech enhancement in the present paper. The present study aims to analyze the performance of APSO and to compare it with existing standard PSO algorithm, in the context of dual channel speech enhancement. Objective evaluation of the proposed method is carried out by using three objective measures of speech quality SNR, Improved SNR, PESQ and one objective measure of speech intelligibility FAI. The performance of the algorithm is studied under babble and factory noise environments. Simulation result proves that APSO based speech enhancement algorithm is superior to the standard PSO based algorithm with an improved speech quality and intelligibility measures.

A new approach to dual channel speech enhancement is proposed based on a recently introduced metaheuristic optimization algorithm called hybrid PSOGSA. It is a novel algorithm which combines the ability of exploration in gravitational search algorithm (GSA) and the exploitation capability of particle swarm optimization (PSO) to offer a better local search process along with the social thinking. This paper aims to present such a hybrid combination as a promising and powerful technique to adaptive noise cancellation in speech enhancement and it is compared with the standard PSO (SPSO) and GSA based speech enhancement algorithms. Simulation results prove that the performance of PSOGSA is superior to SPSO and GSA algorithms, in the context of speech enhancement.

No abstract

No abstract