Smart Precision Feeding System for Dairy Cows based on Amplitude Iterative Pruning Algorithm

This research work presents a smart precision feeding system for dairy cows based on an amplitude iterative pruning algorithm. Smart farms use information technology to optimize field management, increase yields, and reduce costs. The proposed system focuses on precision feeding, estimating individual characteristics of each cow, calculating precise feed formulas, and improving milk production and quality. The Amplitude Iterative Pruning algorithm, inspired by synaptic development in the human brain, reduces the complexity of Deep Convolutional Neural Networks (DCNNs) by removing unimportant connections, thereby increasing computational efficiency. The system incorporates an automatic identification system using low-frequency near-field passive radio frequency identification for efficient feeding control, and the hardware consists of six modules: recognition, feeding console, weighing and quality, feeding, behavioral data cache, and data storage. Aging tests and component sorting procedures are implemented to ensure static accuracy. The performance of the proposed algorithm is verified by experiments, with an average detection rate of 99.8%. The precision feeding decision system accounts for individual cow differences, promoting automation, reducing labor costs, and improving efficiency. Despite successful laboratory results, challenges in real farm environments include equipment stability, data accuracy, and real-time processing. Future research should focus on optimizing hardware and software for different farm conditions. The proposed system contributes to efficient and accurate precision feeding in the dairy industry, reflecting advances in smart agriculture and IoT applications.