Optimal Linear Phase FIR Filter Design Using Multivariant Optimization Process Memorise Algorithm

The digital filter is one of the important components of digital signal processing, which is widely used in control system, audio and video processing system, communication system, image processing system, etc. Based on the network structure, the digital filter is mainly divided into the Finite Impulse Response Filter (FIR) and the Infinite Impulse Response Filter (IIR). Because it is easy to be designed and implemented and with high stability and strict linear phase characteristic, FIR filter always be the prior choice. There are a lot of design and implementation ways for the digital FIR filter. The optimal design method can design out both good damping characteristics and good edge frequency of digital FIR filters. The optimum design of the digital FIR filter is a process that parameters of the FIR filter are optimized under certain optimization criterion using certain optimization algorithm. In recent years, there are many literatures about design of the optimal digital FIR filter. Using the swarm intelligence optimization algorithm for design of the optimal FIR filter which have achieved fruitful results is a very important research direction.Multivariant Optimization process memorise Algorithm (MOA) is a new kind of generalized swarm intelligence optimization algorithm. With constructing a special MOA structure which makes full use of the advantage of modern computer hardware resources, MOA overcomes such faults as being limited by computer memory capacity and processing speed which have limited some early swarm intelligence optimization algorithm such as Particle Swarm Optimization Algorithm, Ant Colony Optimization Algorithm. The main idea of MOA is to use MOA structure to memory search process and share search information. The MOA structure makes it possible to realize the global and local search alternatively, which achieves global optimization. This paper attempts to use MOA to design the optimal digital FIR filter. The significance of the study is two-sided. On the one hand, this paper can find a new method and provide meaningful reference for the design of the optimal FIR filter. On the other hand, it can rich the field that MOA being applied.Firstly, the characteristics of the FIR filter have been discussed and the optimal design criterion of the FIR filter has been introduced. Then the limitations that using windows function and system identification to design FIR filters have been discussed. Secondly, MOA algorithm to design the optimal FIR filter has been introduced. Then, to find out the relationship between the neighborhood radius and the algorithm performance repeated experiments and detailed discussions have been made. In the end, MOA is applied to design four types of the FIR filter and satisfactory design results have been achieved. With comparing the genetic algorithm and particle swarm algorithm, it is verified that Using MOA to design optimal FIR filters is feasible, stable and full of advantages.