Cow Behavioural Activities in Extensive Farms: History
View Latest Version
Please note this is an old version of this entry, which may differ significantly from the current revision.
Contributor:
Dominga Mancuso
,
Giulia Castagnolo
,
Simona M. C. Porto

Animal welfare is becoming an increasingly important requirement in the livestock sector to improve, and therefore raise, the quality and healthiness of food production. By monitoring the behaviour of the animals, such as feeding, rumination, walking, and lying, it is possible to understand their physical and psychological status. Precision Livestock Farming (PLF) tools offer a good solution to assist the farmer in managing the herd, overcoming the limits of human control, and to react early in the case of animal health issues. The entry highlights a key concern that occurs in the design and validation of IoT-based systems created for monitoring grazing cows in extensive agricultural systems, since they have many more, and more complicated, problems than indoor farms. In this context, the most common concerns are related to the battery life of the devices, the sampling frequency to be used for data collection, the need for adequate service connection coverage and transmission range, the computational site, and the performance of the algorithm embedded in IoT-systems in terms of computational cost. 

  • precision livestock farming
  • Internet of Things
  • extensive farms

1. Introduction

“In recent years, animal welfare has become a central objective at a global level” (OIE Global Conference on Animal Welfare, 2004). Beyond the need for a common definition of animal welfare, the international scientific community has engaged in research aimed at improving the welfare of animals reared at various stages of production, from breeding to transport and slaughter, as well as developing and validating automatic systems for assessing animal welfare during the breeding process. In this regard, ICT-based monitoring systems have been created recently for the assessment of animal welfare at the farm level. These systems vary from one another in their specific task: to certify the level of well-being; evaluate the various housing systems; diagnose welfare problems on individual farms; and serve as a support tool for the breeder to find, prevent, or solve problems related to herd welfare. Consumers are also interested in achieving animal welfare because they value a proactive approach to managing animal health and welfare [1][2][3].
Animal behaviour is a clear sign of an animal’s physiological and physical state: cows’ major activities include feeding, rumination, lying, and walking, and their daily monitoring is crucial for farmers to evaluate cow welfare conditions. Operators can examine cows’ behavioural activities directly by visual examination, but it is a time-consuming and labour-intensive operation [4], especially on extensive farms. In the scientific field of precision livestock farming (PLF), ICT-based solutions are being developed and validated to increase the efficiency of livestock monitoring and management [5]. Such modern ICT-based solutions are becoming more and more efficient, ensuring the acquisition of a large amount of data that, when controlled by optimized algorithms, may be of tremendous assistance to farmers in monitoring the herd in an efficient and lucrative manner.
The fourth industrial revolution had a significant impact on industries and economic rules, as it allowed, through new technologies, the interconnection between machines, devices and people and laid the foundations for intelligent automation. This has led to a lesser presence of human help in the execution of repetitive actions, delegating these tasks to intelligent machines. Some of the pillars of the fourth industrial revolution are the Internet of Things (IoT), big data and analytics, autonomous robots, and cloud computing [6].
In recent years, solutions that integrate IoT systems with artificial intelligence techniques have been increasingly present. Advanced AI techniques have proved to be an efficient tool in the analysis of the large amount of data acquired from sensors producing new knowledge that cannot be obtained through traditional techniques [7][8].
The IoT is quickly evolving in the field of PLF. IoT-based systems connect computing devices, mechanical and digital equipment, items, animals, or humans to a network and transfer data without requiring human-to-human or human-to-computer contact [9].
The main elements of an IoT-based system are object identification, sensing, communication, computation, service, and semantics, as described in Figure 1.
/media/item_content/202305/645b55daed8a4sensors-23-03828-g001.png
Figure 1. Main elements of an IoT system [9].
Farmers’ animal management practices have clear limits in the context of cow breeding. Most existing solutions are time-consuming, labour-intensive, and hence costly. Many livestock producers rely on stockperson observations to discover health and welfare concerns, although many commercial facilities have high stockperson-to-animal ratios.
The three main hurdles to efficiently monitoring cow welfare are cost, validity, and timeliness of insights [10].
The use of IoT-based sensors enables the early diagnosis of cow sickness, allowing farmers to intervene earlier and optimize antibiotic administration, milk supply, and veterinary care costs [11].
As a result, the use of wearable sensors is becoming a critical tool for monitoring the health and well-being of cows in housed systems. In most situations, such technology consists of (i) a device (sensor) that measures certain parameters; and (ii) software that processes the sensor’s data, generating information, warnings, and suggestions for the breeder [5]. Monitoring changes in cow behaviour with IoT-based wearable sensors provides unique insights into the study of an animal’s condition and well-being. Such changes may be caused by health and welfare issues, as well as dangers and changes in their environment [12]. Farmers can now monitor vital indicators, including blood pressure, heart rate, and hormonal levels; animal behaviours, including feeding, standing, rumination, and walking [13]; abnormal food and water consumption behaviours, e.g., limited feed ingestion influenced by unappetizing pasture quality, ruminal inactivity, and excessive water consumption due to increased walking activity [14]; and other parameters, such as geolocation information that can be recorded and further analysed [15]. Obviously, by knowing the time spent by animals on each behaviour, it is possible to carry out assessments of their state of health.
Only a few research projects have focused on extensive livestock systems, although IoT-based solutions for monitoring cow behaviour and well-being have been created exclusively in intensive housing systems. In this latter situation, the grazing animal monitoring faces several issues connected to the expansion of the grazing area and the animals’ ability to show their natural behaviour. Furthermore, because there is less human oversight, it is difficult to monitor and analyse the reasons for any unusual behaviour.

2. Cow Behavioural Activities in Extensive Farms

Behavioural Activities Monitoring

Monitoring animal behaviour is critical for measuring animal welfare as well as successful herd management, particularly in extensive grazing systems. In this regard, ongoing automated behaviour analysis is a critical task since farmer-to-animal interaction is likely to be less frequent than in indoor breeding systems. As a result, today’s extensive farms are broader than in the past. It is not always easy to monitor animal behaviour by direct visual inspection by farmers.

Devices

Sensors have played an important role in improving agricultural conditions and, more broadly, farm management since the introduction of PLF. Furthermore, with the advancement of increasingly effective IoT-based technologies, it has become feasible to employ sensor networks even in hostile settings such as barns, which are characterized by dust, a lack of energy, and a lack of an internet network. It showed that there are various ICT-based monitoring systems developed for cows kept in indoor systems, but relatively only limited applications in extensive grazing farms. This is most likely due to the difficulties of employing ICT-based monitoring systems in rural locations where telecommunication network coverage is typically poor. Furthermore, the use of wearable sensors powered by batteries may result in management and maintenance expenses for farmers if the ICT-based systems are not optimized for energy savings [16].
Wearable sensors can collect a significant volume of data, as well as evaluate the raw data and inform farmers if the cattle’s behaviour is odd within a certain range [11].
In animal husbandry, several biometric and biological available sensors are often used, and they can be classified as non-invasive or invasive, as reported in Neethirajan et al. [10].
Invasive sensors are generally ingested or implanted in the animal’s body for tracking physiological measurements such as internal temperature. Non-invasive sensors are generally applied to an animal’s body by using collars or other attachment system to monitor livestock behaviour; moreover, non-invasive sensors are often installed in the breeding environment to monitor environmental parameters such as air temperature, relative humidity, and ventilation.
Since invasive sensors can directly measure animal health factors, they generate more accurate data than non-invasive sensors. However, non-invasive sensors are the most used macro-category because they are easily worn, have a lower cost, can be reused, and, most importantly, they cause less stress to animals than invasive ones.
Table 1 contains the major non-invasive sensors utilized in PLF applications, as well as the animal aspects monitored. Cameras and accelerometers are the most frequent non-invasive sensors used to monitor cow behaviour in indoor environments. Video-recording systems are low-cost solutions to observe the behavioural activities of several animals at the same time with a small number of cameras. An issue to consider when employing such non-invasive systems is animal identification, which is challenging and not always possible, even with the most advanced computer vision-based methods. Obviously, in extensive farms, it is not possible to install efficient video surveillance systems, due to both the extent of the grazing areas and unavailability of a constant and reliable energy source.
Table 1. Non- invasive sensors in PLF applications.
Sensor/Device Aspect of Animal Disease/Used for
GPRS, GPS and transponders, accelerometers Cattle’s position, inside the barn or outside the barn Grazing, feeding, lying, behaviour and welfare monitoring
Motion changes Lameness, oestrus
Pressure - Feeding and drinking monitoring
Microphone, UHF sensors Monitor sound levels in barns Mooing, pain and welfare conditions, rumination, breathing disease
Temperature Temperature monitor Fever, ovarian cysts, pneumonia, retained placenta, mastitis
Thermal infrared camera, 2D cameras, 3D cameras - Behaviour monitoring, lameness, oestrus
Load sensor Weight distribution Lameness
Gas sensor Breathe ketones, methane emission Displaced abomasum, ketosis
Radio-frequency identification Identification Behaviour and welfare monitoring
However, in recent years, with the introduction of increasingly efficient unmanned aerial vehicles (UAVs) equipped with cameras, some animal monitoring systems in extensive pastures based on the use of UAVs have been proposed [17][18]. The UAV-based systems just mentioned still needs improvement and development, as the limitation, to date, is in the very short battery life.
The non-invasive sensors most used in the field of cow monitoring in extensive farms are GPS (Global Positioning System) and accelerometers.
Accelerometer-based systems are extremely flexible and inexpensive. They can be installed in an animal’s leg or neck to monitor behavioural activity. Pedometers are often attached to barn animals in the distal portion of the left hind limb and monitor the number of steps made by the cow; collars are attached to the neck and monitor head movements.
GPS sensors are used to locate grazing animals in extensive breeding systems; as in the case of accelerometers, they, too, can be worn by the animal by using collars. GPS is beneficial in those situations when the territorial extent of the grazing grounds does not allow for frequent and exact management of the herd.

Sampling Rate and Data Collection

In GPS-based monitoring systems, the time acquisition interval affects the precision of the distance travelled by cows, the battery life, and how quickly farmers can respond to theft and trespassing [19]. If the device sends the positions at very long intervals, the risk is that any data loss could negatively affect the monitoring, so as to make the device inefficient. The sampling intervals normally used for GPS-based devices are between 1 and 60 min.
The system developed by Porto et al. [20] guaranteed long-term monitoring of the animals by allowing a collection of waypoints, such as latitude and longitude, of the cows selected in the study, the date and time of the survey, and the distance travelled by each animal. The time interval of data acquisition was set at 20 min to both ensure a long battery life and make possible further analyses carried out in a Geographical Information Systems (GIS) environment; i.e., the application of Kernel Density Estimation (KDE) algorithms. After receiving position information, the device sent it to a cloud server by using the Sigfox telecommunication network.
The sampling rate of the accelerometers in devices that monitor animal behavioural activities affects not only battery life but also the ability to correctly determine some behaviours. High sampling rates make it possible to acquire a lot of information, and therefore have more samples; however, they negatively affect battery life. On the other hand, low sampling rates help to preserve battery life but do not always allow the acquisition of good-quality data for classification purposes, as proved in Benaissa et al. [21].

Data Analysis

The data collected by GPS sensors are mainly analysed by using statistical and geo-spatial tools, such as GIS tools.
In Porto et al. [20], all information was transmitted to a custom AppWeb operating on mobile devices or a personal computer. The data were then imported and processed by using statistical and geospatial analysis. In detail, to better understand the interaction between livestock activities and the environment, spatial analyses were carried out by using the QGIS software, which allowed data processing and visualization at the territorial level. Analyses on the location of each animal equipped with the created devices were performed by using the KDE tool. In particular, the KDE analyses allowed the computation of the home range of the species and offered an estimate of the territorial areas mostly occupied by the animals.
Gonzàlez L. et al. [22] developed an algorithm which classified data acquired from collars into five behavioural activities. The aim of the study was to obtain the proportion of the daily time that individual animals spent on each activity. They employed two different datasets in the experimental trial, where data from accelerometers and GPS were first aggregated by computing the mean and standard deviation (SD) over 10 s time intervals. The first dataset, which included a subset of data in which behavioural activities were identified based on visual observations, was used to determine differences between activities from sensor data values; to inspect frequency distributions (histograms) of data with different activities; to select variables suitable for decision trees; and to construct conceptual decision trees. The second dataset, containing all data related to unknown behavioural activities, was used to fit probability density functions in mixture models that determined threshold values to separate populations of data points.

3. Conclusions

As previously mentioned, the combined systems, which have GPS and accelerometers, allow monitoring of the activities of grazing animals in a more complete way, compared to systems that use only one type of sensor. To date, it has emerged that in the state-of-the-art applications, various prototypes of GPS collars have been proposed which send the positions of the animals in real time, while in the case of accelerometers, the prototypes which send the data in real time are numerically smaller. In many studies, the data acquired with the accelerometer were downloaded at the end of the tests, as the objective of the studies was to propose methodologies for behaviour detection. Therefore, the data were not sent in real time. Sending raw data from real-time accelerometers requires a telecommunications network capable of transferring large amounts of data. If, instead, the computation is performed on-site, it is necessary to deal with the computational cost of the processing and the relative performance of the batteries.
In addition, as far as the combined systems are concerned, in this case, the devices that jointly process the data acquired by the accelerometer and GPS are very few. In this case, many of the acquired data are processed subsequently, and, moreover, not always jointly; i.e., two different processes are carried out based on the nature of the data.
In the future, it is hoped that researchers will focus on real-time behavioural analysis, using joint position and movement data, to respond to farmers’ needs. What has been said involves, in part, low-power networks, battery optimization techniques, and low-computational-cost firmware.

This entry is adapted from the peer-reviewed paper 10.3390/s23083828

References

  1. Clark, B.; Panzone, L.A.; Stewart, G.B.; Kyriazakis, I.; Niemi, J.K.; Latvala, T.; Tranter, R.; Jones, P.; Frewer, L.J. Consumer Attitudes towards Production Diseases in Intensive Production Systems. PLoS ONE 2019, 14, e0210432.
  2. Brusselles: European Commision Attitudes of EU Citizens towards Animal Welfare, Report; Special Eurobarometer 442; 2016. Available online: https://europa.eu/eurobarometer/surveys/detail/2096 (accessed on 5 April 2023).
  3. European Commission Online Consultation on the Future of Europe; Second Interim Report; 2019. Available online: https://commission.europa.eu/about-european-commission/get-involved/past-initiatives/citizens-dialogues/list-citizens-dialogues-events-2015-2019/progress-reports-citizens-dialogues_en (accessed on 5 April 2023).
  4. Riaboff, L.; Aubin, S.; Bédère, N.; Couvreur, S.; Madouasse, A.; Goumand, E.; Chauvin, A.; Plantier, G. Evaluation of Pre-Processing Methods for the Prediction of Cattle Behaviour from Accelerometer Data. Comput. Electron. Agric. 2019, 165, 104961.
  5. Rutten, C.; Steeneveld, W.; Vernooij, J.; Huijps, K.; Nielen, M.; Hogeveen, H. A Prognostic Model to Predict the Success of Artificial Insemination in Dairy Cows Based on Readily Available Data. J. Dairy Sci. 2016, 99, 6764–6779.
  6. Vaidya, S.; Ambad, P.; Bhosle, S. Industry 4.0—A Glimpse. Procedia Manuf. 2018, 20, 233–238.
  7. Neethirajan, S. The Role of Sensors, Big Data and Machine Learning in Modern Animal Farming. Sens. Bio-Sens. Res. 2020, 29, 100367.
  8. Almalki, F.; Soufiene, B.; Alsamhi, S.; Sakli, H. A Low-Cost Platform for Environmental Smart Farming Monitoring System Based on IoT and UAVs. Sustainability 2021, 13, 5908.
  9. Al-Fuqaha, A.; Guizani, M.; Mohammadi, M.; Aledhari, M.; Ayyash, M. Internet of Things: A Survey on Enabling Technologies, Protocols, and Applications. IEEE Commun. Surv. Tutorials 2015, 17, 2347–2376.
  10. Neethirajan, S.; Kemp, B. Digital Livestock Farming. Sens. Bio-Sens. Res. 2021, 32, 100408.
  11. Jorquera-Chavez, M.; Fuentes, S.; Dunshea, F.R.; Jongman, E.C.; Warner, R. Computer Vision and Remote Sensing to Assess Physiological Responses of Cattle to Pre-Slaughter Stress, and Its Impact on Beef Quality: A review. Meat Sci. 2019, 156, 11–22.
  12. Doulgerakis, V.; Kalyvas, D.; Bocaj, E.; Giannousis, C.; Feidakis, M.; Laliotis, G.P.; Patrikakis, C.; Bizelis, I. An Animal Welfare Platform for Extensive Livestock Production Systems. CEUR Workshop Proc. 2019, 2492, 1–7.
  13. Arcidiacono, C.; Porto, S.; Mancino, M.; Cascone, G. Development of a Threshold-Based Classifier for Real-Time Recognition of Cow Feeding and Standing Behavioural Activities from Accelerometer Data. Comput. Electron. Agric. 2017, 134, 124–134.
  14. Oudshoorn, F.; Cornou, C.; Hellwing, A.; Hansen, H.; Munksgaard, L.; Lund, P.; Kristensen, T. Estimation of Grass Intake on Pasture for Dairy Cows Using Tightly and Loosely Mounted Di- and Tri-Axial Accelerometers Combined with Bite Count. Comput. Electron. Agric. 2013, 99, 227–235.
  15. Brennan, J.; Johnson, P.; Olson, K. Classifying Season Long Livestock Grazing Behavior with the Use of a Low-Cost GPS and Accelerometer. Comput. Electron. Agric. 2021, 181, 105957.
  16. Tamura, T.; Okubo, Y.; Deguchi, Y.; Koshikawa, S.; Takahashi, M.; Chida, Y.; Okada, K. Dairy Cattle Behavior Classifications Based on Decision Tree Learning Using 3-Axis Neck-Mounted Accelerometers. Anim. Sci. J. 2019, 90, 589–596.
  17. Marcos, J.T.C.; Utete, S.W. Animal Tracking within a Formation of Drones. In Proceedings of the 2021 IEEE 24th International Conference on Information Fusion (FUSION), Sun City, South Africa, 1–4 November 2021; pp. 1–8.
  18. Zhou, M.; Elmore, J.A.; Samiappan, S.; Evans, K.O.; Pfeiffer, M.B.; Blackwell, B.F.; Iglay, R.B. Improving Animal Monitoring Using Small Unmanned Aircraft Systems (SUAS) and Deep Learning Networks. Sensors 2021, 21, 5697.
  19. McGavin, S.L.; Bishop-Hurley, G.J.; Charmley, E.; Greenwood, P.L.; Callaghan, M.J. Effect of GPS Sample Interval and Paddock Size on Estimates of Distance Travelled by Grazing Cattle in Rangeland, Australia. Rangel. J. 2018, 40, 55.
  20. Porto, S.M.C.; Castagnolo, G.; Valenti, F.; Cascone, G. Kernel Density Estimation Analyses Based on a Low Power-Global Positioning System for Monitoring Environmental Issues of Grazing Cattle. J. Agric. Eng. 2022, 53, 1323.
  21. Benaissa, S.; Tuyttens, F.A.; Plets, D.; Cattrysse, H.; Martens, L.; Vandaele, L.; Joseph, W.; Sonck, B. Classification of Ingestive-Related Cow Behaviours Using RumiWatch Halter and Neck-Mounted Accelerometers. Appl. Anim. Behav. Sci. 2018, 211, 9–16.
  22. González, L.; Bishop-Hurley, G.; Handcock, R.; Crossman, C. Behavioral Classification of Data from Collars Containing Motion Sensors in Grazing Cattle. Comput. Electron. Agric. 2014, 110, 91–102.
More
© Text is available under the terms and conditions of the Creative Commons Attribution (CC BY) license; additional terms may apply. By using this site, you agree to the Terms and Conditions and Privacy Policy.
This entry is offline, you can click here to edit this entry!
ScholarVision Creations
Feedback