PARAMETERS ESTIMATION OF DOUBLE EXPONENTIAL SMOOTHING FOR HAND JITTER REDUCTION USING GENETIC ALGORITHM
Nor Farizan Zakaria, Mohd Asyraf Zulkifley, Mohd Marzuki Mustafa · Jurnal Teknologi · 2015
Hand jitter is a natural tremor that has become a concern in many areas such as microsurgery, collaborative environment and tele-tutoring as it can cause imprecision, inaccuracy and misleading pointer information. Over the recent years, many researches have been done to reduce hand jitter but they are either too complex or too time consuming. Hence, to overcome the limitations, Double Exponential Smoothing (DES) has been used as an alternative, which is a simple and fast prediction algorithm. However, estimating the parameter values of DES is a difficult process that requires juggling between several criteria. In this paper, an optimal parameter estimation technique of DES using Genetic Algorithm was developed to find the optimal parameter values. Thorough comparisons have been made with previous methods to prove the magnitude of improvement. Our study found that the proposed method is able to reduce the hand jitter by 52% compared to the benchmarked methods. Hence, DES is suitable to be implemented in many applications that require precise and accurate hand-based pointing system.