11 / 2025-03-26 11:19:01

Dynamic estimation the individual feed intake of housing cattle using a noseband pressure sensor

dynamic estimation,feed intake,housing cattle,pressure sensor

3. 智慧牧场与动物福利 > （2）智慧养殖的大数据、机器学习和物联网应用

Individual feed intake can be estimated through the feeding and rumination behavior of cattle. The current research on estimating individual feed intake in cattle is based on empirical models that rely on the fixed linear or nonlinear relationships using the behavioral data (feeding time, chewing frequency, etc.) monitored by devices. However, these empirical models demonstrate limited adaptability to dynamic changes in physiological and forage states. This study proposes a algorithm framework for estimating individual feed intake in cattle by dynamically analyzing jaw movement variations to quantify feeding rates, the amount of feed consumed per minute. Feeding rates are shaped by multiple interacting factors, including diverse hunger levels and varying feed pile volumes. Feeding behaviors and feeding rates will be monitored by the wearable noseband pressure sensor and analyzed under varying conditions include four different hunger levels and three different feed pile volumes. A behavioral classification model will be initially developed using the collected data to differentiate behavior patterns including feeding, ruminating, and chewing. A regression prediction model will be constructed to estimate feeding rate by feature extraction from the collected data classified as feeding behavior. And the key behavioral parameters including feeding duration, rumination duration, and chewing frequency will be quantified and further used for calibration the feed intake estimated through the feeding rates and thereby optimizing the estimation precision. The results of this study would provide an innovation method to monitor individual feed intake of cattle by wearable devices and enable their individual precision feeding management and performance evaluation.