A Vibration Processing Analysis Approach to the Development of Fish Feeding Regime System

Abstract

Feeding is important in life cycle of fish, whereas one of the major challenges facing aquaculture industry is the inefficient management of the feeding. Excess feeding results in left over feed which eventually produces negative environmental effect. Feeding fish based on their demand will reduce feed wastage to the bearest minimum and effectively address the problem of water contamination. This paper proposes a feed dispensing algorithm using vibration processing analysis. The intent is to develop a feeding regime system that can identify fish behavioral activities to determine when fish is hungry, and provide them with precise amount of food as demanded. The dataset adopted in this study was obtained using data logger incorporating triaxial gyroscope, accelerometer and magnetometer. The gyroscope used to determine the changes in angular velocity with respect to x-, y- and z-direction, while accelerometer and magnetometer employed to determine the changes in acceleration in x-, y- and z-direction. These patterns were further process using sliding window analysis. The mechanical differences between these patterns is proposed to be used for classify when fish are hungry from other activities. It is expected that the proposed system will has potential of reducing overfeeding and increase fish productivity.