Artificial Raindrop Algorithm for Optimal Parameter Preference in Digital IIR Filters

Yuansheng Huang, Ying Qiao
2017 Advances in Modelling and Analysis C  
The system identification of digital Infinite Impulse Response (IIR) filter, as a key knowledge domain, is an important research subject in the automatic control field. However, the error surface of digital IIR filter is usually nonlinear and multimodal, which makes the cost function rather difficult to minimize. Whilst some global optimization techniques such as metaheuristic algorithms are essential for avoiding local minima encountered in conventional IIR modeling mechanisms. In this paper,
more » ... rtificial Raindrop Algorithm (ARA), a metaheuristic approach recently developed as a member of the family of nonlinear optimization, is applied to identify the unknown parameters in the design of digital IIR filter. The ARA is inspired by the phenomenon of natural rainfall, whose components include the generation of raindrop, the descent of raindrop, the collision of raindrop, the flowing of raindrop and the updating of vapor. The paper studies algorithm's performance by a comparative law, aiming at eight primal intelligence optimization algorithms, as some state-ofthe-art models, and eight improved metaheuristic algorithms. The experimental results show that ARA can more accurately identify the parameters as most of chosen and widely used cases and 115 may become a promising candidate for digital IIR filter. Keywords Design of digital IIR filter, Global optimization, Artificial raindrop algorithm, System identification. 2 1 N Vapor J Vapor J Vapor J Vapor J  = by the Eq.(11);
doi:10.18280/ama_c.720202 fatcat:ture6bakk5ctvjqggn2v3nx2ju