In this paper, the suitability of a specifically designed and developed inertial sensor-based system for the real-time detection of dairy cow behaviours was assessed. The hardware was composed of an accelerometer sensor integrated into a Bluetooth Low Energy (BLE) device and a Raspberry Pi equipped with a USB BLE. The software components included a Python script for the Raspberry Pi platform and a Java Graphical User Interface. Experimental tests were carried out in a free-stall barn located in Southern Italy. Three dairy cows were equipped with sensors at the forelimbs. Data were reorganised and filtered twice by applying low pass Butterworth filter and data analyses were performed by designing an algorithm in MATLAB® that utilises statistical parameters of the three accelerometer components. With the aim of recognizing cow posture in order to compute the lying behaviour duration, the mean value of the acceleration on the X axis was compared with a 0.5g threshold. Furthermore, motion analysis was performed to identify walking activity. To this aim, standard deviation of each acceleration component was computed to find out whether the forelimb was in motion, subsequently data variances were utilized to discriminate between a low or a rapid motion. The system detections were validated by using information from a video-recording system. Good results were achieved in posture recognition and detection of behavioural activity change. In detail, changes in posture were always recognised whereas small forefoot movements without changes in position were not detected. The misclassification rate was equal to about 24%. However, further experiments are needed to obtain the actual value since disconnections between the sensor units and the computing unit frequently occurred during the experiments. Although some improvements are needed, the experimental tests showed that this system is promising since it is based on a low-cost technology with a low energy consumption.

Feasibility study on the development of a real-time inertial sensor-based system for dairy cow behaviour detection in free stall barns

ARCIDIACONO, Claudia
;
CASCONE, Giovanni;CATANIA, Vincenzo;R. Di Natale;A. R. Intilisano;M. Mancino;PORTO, SIMONA MARIA
2015-01-01

Abstract

In this paper, the suitability of a specifically designed and developed inertial sensor-based system for the real-time detection of dairy cow behaviours was assessed. The hardware was composed of an accelerometer sensor integrated into a Bluetooth Low Energy (BLE) device and a Raspberry Pi equipped with a USB BLE. The software components included a Python script for the Raspberry Pi platform and a Java Graphical User Interface. Experimental tests were carried out in a free-stall barn located in Southern Italy. Three dairy cows were equipped with sensors at the forelimbs. Data were reorganised and filtered twice by applying low pass Butterworth filter and data analyses were performed by designing an algorithm in MATLAB® that utilises statistical parameters of the three accelerometer components. With the aim of recognizing cow posture in order to compute the lying behaviour duration, the mean value of the acceleration on the X axis was compared with a 0.5g threshold. Furthermore, motion analysis was performed to identify walking activity. To this aim, standard deviation of each acceleration component was computed to find out whether the forelimb was in motion, subsequently data variances were utilized to discriminate between a low or a rapid motion. The system detections were validated by using information from a video-recording system. Good results were achieved in posture recognition and detection of behavioural activity change. In detail, changes in posture were always recognised whereas small forefoot movements without changes in position were not detected. The misclassification rate was equal to about 24%. However, further experiments are needed to obtain the actual value since disconnections between the sensor units and the computing unit frequently occurred during the experiments. Although some improvements are needed, the experimental tests showed that this system is promising since it is based on a low-cost technology with a low energy consumption.
2015
978-889097532-5
inertial sensors, behaviour analysis, accelerometer, real-time analysis
File in questo prodotto:
File Dimensione Formato  
Feasibility study on the development of a real-time inertial sensor-based system for dairy cow behaviour detection in free stall barns.pdf

solo gestori archivio

Tipologia: Versione Editoriale (PDF)
Dimensione 647.29 kB
Formato Adobe PDF
647.29 kB Adobe PDF   Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.11769/80776
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 3
  • ???jsp.display-item.citation.isi??? ND
social impact