Short-term fibre intake estimation in goats using surface electromyography of the masseter muscle

• Published in: Biosystems engineering 2019-07, Vol.183, p.209-220
• Main Authors: Campos, Daniel P., Abatti, Paulo J., Bertotti, Fábio L., De Paula Vieira, Andreia, Hill, João A.G., da Silveira, André L.F.
• Format: Article
• Language: English
• Subjects: Chewing pattern; Electronic instrumentation; Intake prediction; Precision livestock farming (PLF); sEMG; Signal processing
• ISSN: 1537-5110; 1537-5129
• DOI: 10.1016/j.biosystemseng.2019.04.021

The demand for quality control and environmentally acceptable animal production are important requirements in livestock monitoring and control. Monitoring the ingestive behaviour of grazing ruminants has applications in measuring animal welfare, detection of animal diseases and modelling the animal production system. In this work, a non-invasive method for fibre intake estimation on ruminants is presented using a surface electromyography (sEMG) based sensor system. In order to acquire bite/chewing sEMG signals, superficial disposable electrodes were placed on three uncastrated male goats’ masseter muscle, housed in individual pens, and data was sampled during eating using an analogue-to-digital converter. Feed samples and-left-overs were weighed before and after the experiment, respectively, to estimate the total intake. Electromyographic preprocessed data was sent to a computer, where seven signal features were extracted. Feed intake was modelled by fitting sEMG features as a predictor by means of a linear model. Results indicate that fibre intake could be successfully predicted in goats eating several forages (Tifton 68 and Tifton 85 grass hay, bad quality Tifton hay, and forage oat hay) with coefficients of determination (R2) higher than 0.867, using the signal feature called Slope Sign Change (SSC) as a predictor, which expresses signal frequency characteristics and indicates physiological aspects of bite/chewing. The present work demonstrates that sEMG based systems can be used as a non-invasive feed fibre intake estimation tool. •Novel electronic method for monitoring the feeding behavior.•Method based on signal processing of the masseter muscle surface electromyography.•Extracted sEMG features highly correlated to fibre intake for different feeds.•Feed intake prediction model does not require known feed.