Smart farming is a concept of agricultural innovation that combines technological, social, economic and institutional changes. It employs novel practices of technologies and farm management at various levels (specifically with a focus on the system perspective) and scales of agricultural production, helping the industry meet the challenges stemming from immense food production demands, environmental impact mitigation and reductions in the workforce. Precision Livestock Farming (PLF) systems will help the industry meet consumer expectations for more environmentally and welfare-friendly production. However, the overwhelming majority of these new technologies originate from outside the farm sector. The adoption of new technologies is affected by the development, dissemination and application of new methodologies, technologies and regulations at the farm level, as well as quantified business models. Subsequently, the utilization of PLF in the pig and especially the poultry sectors should be advocated (the latter due to the foreseen increase in meat production). Therefore, more significant research efforts than those that currently exist are mainly required in the poultry industry.
1. Introduction
Smart farming is a concept in agricultural innovation that combines technological, social, economic and institutional changes [1]. It employs novel practices of farm management at various levels (specifically focused on the system level) and scales of agricultural production, helping the industry to meet the challenges stemming from the growing food production demands and reduction in the workforce [2]. The approach of Precision Livestock Farming (PLF) for a sustainable farming system refers to the continuous capacity of agriculture to contribute to overall human and animal welfare by using available information more effectively on farms. In turn, the better utilization of information enables farmers to provide sufficient goods and services in ways that are economically efficient and socially and environmentally responsible [3][4].
PLF uses smart farming technology, which includes the utilization of various types of sensors to collect data, which is thereafter usually transferred collectively by communication networks to servers using Information and Communications Technology (ICT). In this Internet of Things (IoT), by generally accepted definition, large amounts of data from interconnected devices are recorded and analyzed by management information systems, data analysis solutions
[5][6] and data analytics
[7] domains. The use of the data provided by smart farming potentially helps boost productivity and minimize waste by allowing the necessary actions to be carried out at the right time and in the right place
[8]. An FAO report
[9] highlighted the importance of ICT as a tool to help meet future food and feed requirements.
Digital technologies have been developed to continuously track real-time production performance and environmental conditions in various livestock facilities
[10]. In this sense, they facilitate an improved response to humans’ and animals’ needs by (a) maximizing production efficiency, (b) increasing product quality, (c) improving animal health and welfare, (d) reducing human occupational health and safety risk and (e) mitigating emissions from livestock. Policymakers can also benefit from increased information sharing, which allows them to gather a more complete overview of the situation at the national and regional levels. An additional major benefit connected with ICT use lies in the potential to reach all the layers of society. Moreover, recent technological developments in areas relevant to IoT facilitate an easier adoption of smart farming and its use by farmers
[11]. As farmers and their attending veterinarians, nutritionists and advisors become increasingly aware of the benefits of ICT, it will hopefully motivate them to upload data to central repositories on, for example, disease incidence, the number of live-born piglets in individual sows, feed intake and weather variables.
Processed data may collectively benefit animal production as patterns emerge or individually as perturbations in the individual animal or group of animals are detected in response to environmental variables. Furthermore, some of the world’s largest agricultural producers are promoting the use of IoT in smart farming by creating incentive programs and public policies to fund research and training
[12].
On the other hand, it was suggested that European farmers lack the knowledge to understand the benefits of ICT-based PLF
[13]. The acceptance of (new) IT technologies, such as big data, computer vision, artificial intelligence, blockchain and fuzzy logic in the smart agriculture field were evaluated in
[14]. A study of the consumer perceptions of PLF technologies showed that consumers expect that PLF technologies will enhance the health and welfare of farm animals while generating environmental improvements and increasing transparency in livestock farming
[15]. The researchers, however, also expressed the fear that PLF technologies will lead to more industrialization in livestock farming, that PLF technologies and data are vulnerable to misuse and cyber-crime and that PLF information may be inadequately communicated to consumers. Public opposition to the industrialization of livestock production is encouraging de-intensification by farmers, either to meet government standards or to capture higher product prices. However, less intensive livestock farming utilizes more land, a commodity in short supply with a growing world population and competition from carbon farming to offset increased emissions.
To ensure that agriculture supplies secure and nutritious food while minimizing environmental threats, farmers need specific economic incentives, help with incorporating innovation into their enterprise and knowledge exchange to encourage the use of advanced and smart technologies. Coherent agricultural, environmental, trade and R&D policies must be presented by the government. It is also vital to base policy decisions on robust, well-established scientific criteria so that the decisions are justified and can be explained to all stakeholders. The EU has been fostering PLF through funding and investment since the FP7 program. The EU CORDIS service (cordis.europa.eu, accessed on 31 May 2023) provides details on 77 forefront projects dealing with animal production systems and animal health, which have received an EU contribution of € 508 million under Horizon 2020 and Horizon Europe programs
[16][17].
2. Technologies in Livestock Farming
Generally, sensors such as thermal imagery, microphones, GPS and others are used in PLF to collect real-time data
[11]. Due to the significant amount of raw data collected, algorithms are often applied to aid analysis. The data can either be directly processed or immediately relayed to the farmer, or it can be transferred to the server of a service provider company where it is analyzed, and the feedback is sent to the farmer. ICT can promote learning, which in turn can facilitate technology adoption among farmers, and it has the potential to revolutionize early warning systems through better quality data and data analysis. However, the information relayed by ICT should be properly targeted and relevant if it is to affect farmers’ production decisions.
The manner by which information is delivered is also a crucial determinant of effectiveness. ICT encompasses many different technologies, from computers and the Internet to radio and television to mobile phones. Their impact varies widely depending on which specific technology is used but also on farmers’ level of technological literacy. A growing body of evidence suggests that in many circumstances, mobile phones can increase access to both information and capacity-building opportunities for rural populations in developing countries
[18]. Farmers can get access to timely and high-quality information on products and inputs, as well as on environmental and market conditions. Short message services (SMS), voice messages, short video trainings, audio messages, social media interventions and virtual extension platforms that can improve peer networks (through online platforms/websites) can effectively enable farmer-to-farmer and farmer-to-experts information sharing.
Within the framework of the AutoPlayPig project
[19], funded by the EU’s Horizon 2020 program under the Marie Skłodowska-Curie grant, a comprehensive review was published on information technologies (ITs) developed for welfare monitoring within the pig production chain, evaluating the ITs developmental stage and how these ITs can be related to the Welfare Quality
® (WQ) assessment protocol
[20]. Of the 101 publications included in the systematic literature analysis, 49% used camera technology, 18% used microphones and 15% animal attached sensors, including accelerometers and radio frequency identification (RFID) tags. The sensor technology used to measure environmental biomarkers included thermometers, an Environmental Monitoring Kit, an anemometer, an air-speed transmitter and a weather station. Most publications investigated feature variables on individual or pen levels of behavioral animal biomarkers.
All of the data generated by the aforementioned sources need to be exploited to validate and further develop useful algorithms; however, this requires the availability of advanced infrastructure
[21]. As such, big data generated from technological sources require advanced analytics for effective exploitation. Advanced infrastructure is also needed for the timely and efficient execution of these big-data-enabled algorithms prior to delivery to the farmer. The recently completed EU project CYBELE
[22], funded under the Horizon 2020 Programme, aimed to introduce to all stakeholders along the agri-food value chain an ambitious and holistic large-scale High-Performance Computing (HPC) platform, offering services in data discovery, processing, combination and visualization and solving computationally-intensive challenges requiring very high computing power and capable of actually generating value and extracting insights from the data
[21].
3. Operational Concepts and Technological Solutions in the Pig Sector
Various areas of research are reflected in the European studies. Among them, several areas are prominent.
Weighing optimization–The completed European project ALL-SMART-PIGS
[23], funded by Horizon FP7, was one of the first EU projects to showcase commercialization as a main focus. The Weight-Detect
TM application (PLF Agritech, Toowoomba, Australia) is an innovative video image analysis system that determines the group average weight of a pen of animals by a video observation system. It enables farmers to determine growth and any weight-based indexes without physically weighing the animals
[24]. Pig weighing optimization was selected for the evaluation and technical validation of a platform in the aforementioned CYBELE project
[22]. The tool is a convolutional neural network that takes images and captures videos above the pens of fattening pigs throughout their weight gain and encodes these images into a latent vector representation. Together with additional relevant information, it estimates the mean ± SD live weight of the pigs in the pen. Body weight recording was the subject of the ClearFarm project
[25]. The automated estimation of body weight was conducted by a depth camera (iDOL65, dol-sensors a/s, Aarhus, Denmark)
[26] placed above the individual feeding station or three-partitioned feeder, which worked in combination with an RFID system installed in the feeding stations. The performance of the depth camera and its underlying algorithm was satisfactory at both installations; however, a lack of frequent maintenance, changes in pens’ uniformity and dietary shifts may compromise image sampling and body weight estimation. Similar results were reported by
[27].
Play behavior–In general, the scientific literature supports the use of play behavior as an indicator of good animal welfare and affective states that are valanced
[28][29][30][31]. The AutoPlayPig project
[19] aimed at taking the first steps in developing a system for automatic detection of play behavior in young pigs as an indicator for welfare assessment. This was accomplished by developing an algorithm to extract heart rate (in beats per minute) from raw video data of an anesthetized and resting pig wearing an electrocardiography (ECG) monitoring system, thus combining ethology and computer science into one field of Computational Ethology (CE)
[32]. Play behavior frequency over the process of weaning piglets was investigated in the ClearFarm project
[25] by analyzing the effects of two weaning methods [conventional weaning: two litters mixed in a weaner pen of different size and design vs. litter staying in the farrowing pen after removing the sow] and two genetic hybrids [DanBred Yorkshire × Landrace vs. Topigs Norsvin TN70 Yorkshire × Landrace]
[33]. The results showed that weaning stress in pigs may be reduced both by using a genetic hybrid pig breed with higher birth and weaning weights and by keeping litters intact in a familiar environment after weaning.
Tail biting–In intensive piggeries, tail biting is common and is considered an indicator of negative welfare
[34]. This issue is addressed in the on-going European project Code Re-farm
[35][36] using Duroc × (Landrace × Yorkshire) piglets in free-farrowing pens. In conclusion, the study’s proposed method detected tail-biting behavior from video sequences of entire pig pens, claiming its CNN-LSTM model to be superior to the CNN-CNN model.
Virus detection–A novel and affordable field diagnostic device, based on advanced, proven, bio-sensing technologies to tackle six important swine diseases has recently been developed within the Horizon 2020 SWINOSTICS project
[37]. The diagnostic device allows threat assessment at the farm level, with the analytical quality of commercial laboratories. The SWINOSTICS mobile device can simultaneously analyze four samples to detect six of the most important swine viral pathogens: Porcine Parvovirus (PPV), Porcine Circovirus 2 (PCV−2), Classical Swine Fever Virus (CSFV), Porcine Respiratory and Reproductive Syndrome (PRRSV), Swine Influenza Virus (SIV) and African Swine Fever Virus (ASFV)
[38][39][40].
Additional research areas–Two smart farming applications ready for commercialization on European pig farms were evaluated within the ALL-SMART-PIGS project
[23]: a feed intake measurement device (Feed-Detect
TM, PLF Agritech, Toowoomba, Australia) and an environmental monitoring (Enviro-Detect
TM, PLF Agritech, Toowoomba, Australia) device
[10][41][42].
4. Operational Concepts and Technological Solutions Used in the Poultry Sector
PLF development in the EU has most commonly focused on broiler farming, followed by laying hens. Modern broiler strains in intensive production systems reach their target weight in just 5–6 weeks or less
[43]. This short life span means that it is difficult to maintain a balance between production objectives and bird welfare. A review of the trends in PLF in the broiler production industry, supported by the Irish Innovation Partnership Pathway,
[44] elaborated that while the use of electro-chemical sensors in precision farming is quite common, the use of state sensors measuring physical properties such as temperature, acceleration or location is still at a preliminary state of deployment. As the cost of wearable sensors decreases, the option of fitting a large number of birds with these physical state sensors seems more and more feasible.
The opportunity exists in the poultry sector to monitor ammonia concentrations using multiple sensors feeding data into a central console. However, developing an adequate ammonia sensor is still a challenge that has not been resolved in a satisfactory way. One of the main outputs of the European project EU-PLF
[45], funded by Horizon FP7, was an embryonic blueprint for commercializing PLF type technologies. Within EU-PLF, broiler activity was defined as a key indicator for welfare and health. The remote camera detection of broiler behavior enabled the development of an early warning system to alert managers to unexpected broiler behavior with 95% true positive events
[46].
Alerting farmers to welfare problems in real-time, especially during winter nights when ventilation is low, allows for fast and targeted interventions, which will immediately benefit the flock compared to traditional welfare assessments that have usually occurred on the next morning
[47]. Ammonia concentrations are often higher during the daytime in livestock buildings due to increased evaporation via higher temperatures, greater movements of birds and increased airflow
[48].
Research and development in poultry disease identification and control should be prospective and incorporate new technologies and should pay special attention to zoonotic diseases. Such a perspective was demonstrated by
[49], in the framework of two completed Horizon 2020 projects–SMARTDIAGNOS
[50] and VIVALDI
[51]–with the study of two
Campylobacter species–
C. jejuni and
C. coli–in poultry flocks. These two species account for most human campylobacteriosis
[52], and poultry and poultry products are considered to be the main sources of disease transmission
[53]. To tackle this, a simple and rapid Loop-Mediated Isothermal Amplification (LAMP) assay was used to detect
C. jejuni and
C. coli in chicken feces.
Broiler production systems must be optimized to enhance their energy/resource efficiency, minimize carbon footprint and create sustainable supply chains by developing the necessary infrastructure across all stages of production, including breeding, hatching, rearing, processing and distribution to consumers. Collaborative research and advanced technologies can help tie together the different components of the system and their relationships. The consequences of not supporting farmers in implementing new technologies may result in the loss of social licence and even threaten the poultry industry’s premier position in the global marketplace and the ability of the industry to provide safe and nutritious poultry products to consumers worldwide
[54]. The lack of collaboration between the private and public sectors and the lack of innovative ways to articulate concerns from producers and consumers to policymakers remain barriers to technological adoption
[5].
5. Farmer’s Attitudes and Obstacles in Acceptance of Technological Solutions in the Livestock Sector
Qualitative and quantitative assessments of the attitudes and barriers to PLF technology adoption have shown the manifold factors that influence farmers’ technology decisions and highlighted the economic, socio-demographic, ethical, legal, technological and institutional aspects that need to be considered for widespread technology acceptance
[55][56][57]. They also showed that “innovation uncertainty” has led to a rather slow uptake of precision technologies by farmers thus far
[58].
The most reported aspect in almost any region is the fear of high investment costs that are needed to enable PLF
[57][59][60][61][62][63][64]. This barrier is particularly prominent for smaller production sites, as the expected investment returns are more limited compared to big farms
[60][63][65].
Aack of trust in the technological capabilities and robustness of the technology was another frequently reported factor that affected technology adoption
[56][57][64][66]. As trust has many different notions, there are several associated aspects that directly or indirectly influence the confidence in specific technologies. Farmers that are in close proximity to other farms and are part of a wider network tend to adopt novel technologies quicker and more optimistically
[57]. Trust in technology is higher if colleagues have used the technology effectively before
[67]. This networking effect was also shown by
[59], who found that 68% of farmers make adoption decisions based on information obtained from colleagues. Trust may also be a relevant factor in the sense of security and privacy. As modern PLF technologies are embedded in an IoT environment and are often accompanied with decentralized data storage (e.g., cloud or edge devices), concerns about data safety may arise. Privacy and security concerns are one of the most prominent barriers that inhibit digital technology adoption by farmers in Wisconsin
[56].
Some other factors, such as technological relevance or lack of awareness have been identified by individual studies
[56][57][64]. These are believed to be of minor importance, and in practice, most attitudes are closely linked to the already mentioned farmers’ characteristics and associated barriers. This also highlights the potential of positive side-effects if one addresses the individual fears and needs of farmers in terms of technology adoption.
Interviews and surveys constitute the main methodology through which the voices of farmers are heard, but it is always important to consider their accuracy, especially in relation to sensitive animal welfare concerns. In the EU-PLF project
[45], farmers voiced their hesitation to purchase PLF technologies, unless its derived benefits are clear and unequivocal, and they also had concerns about maintenance. The issues and importance of training on-site, providing professional on-demand and continuous support, especially concerning animal welfare assessment
[68], and establishing demonstration farms were stressed.
Farmers from the pig and poultry sectors in the UK and Spain interviewed in the EU project Feed-a-Gene
[69] stressed the importance of providing accurate and complete information to farmers and the need for a detailed evaluation of novel technologies in a commercial setting before more widespread adoption. Interviewees from the pig sector favored the concept of precision feeding and the resultant improvement in feed conversion efficiency and improved animal welfare; however, farmers from the poultry sector (in Spain) were largely unenthusiastic. In the pig sector, the benefits seemed clearer for gestating sows than for breeding pigs.
The expected high costs for investment have led to scepticism about whether gains in feed efficiency necessary to justify the investment would be realized. This would particularly apply if existing buildings, infrastructure and feeding systems could not be simply adapted, as most farmers believed to be the case. Concerns were raised about the necessary skill level of operating such precision feeding systems, as it would require skilled labor, which is expensive and could increase labor costs.
One of the targets of the SusPigSys European initiative
[70] (part of the ERA-Net Cofund activity SUSAN) is to promote farmers’ wellbeing. Farmers of pig production systems in seven EU member states (Austria, Germany, Finland, Italy, Poland, the Netherlands and the United Kingdom) participated in national workshops with stakeholders, where the important social implications for farmers themselves were discussed. The participants from Germany, Finland and the UK stressed the importance of consumers’ power along the supply chain and societal acceptance of the public image of pig production and the farming profession, highlighting the disconnect between the industry and the consumers.
To counter this and other negative associations with PLF technologies, the “LivestockSense” project
[71] was implemented in seven different EU countries to encourage PLF technology adoption and increase the general understanding of these technologies. An online quantitative survey was undertaken, and follow-up interviews, as well as focus group discussions (FGDs), were organized to obtain qualitative results. The quantitative questionnaire results demonstrated that the existing level of automatization on the farms, the average age of the livestock buildings (and associated production technologies) and the availability of internet connectivity were clear indicators of livestock producers’ “readiness levels” to adopt PLF technologies.
6. Conclusions
PLF systems can help to increase production efficiency and meet consumer expectations of more environmentally and welfare-friendly production at a time when there is extreme pressure on land availability for food production that deters farmers from reducing the intensity of their operations. Several EU-funded projects have helped to identify and develop PLF technologies that could benefit the livestock industries, particularly in the pig and poultry sectors. The large volume of research in the pig industry is welcomed, but attention must be paid to the global and European trends of an increase in broiler-meat production, which is growing faster than any other meat type, including pork production. Therefore, greater research efforts are required in the poultry industry, with particular recommendation for the development of enhanced research infrastructures in the sectors of laying hens, turkeys and quails, etc., on the basis of their being underrepresented in the plethora of active research projects in Europe.
Considerable obstacles to widespread PLF adoption have been identified and must be addressed. High investment costs, a lack of trust in the technology, uncertainty in the future market for their products and the usability of the technologies have all been identified as impediments to adoption. Fifty-eight percent of European farm managers are 55≤ years of age, and of them 33%, are over 65 years. Most of them work on small-size farms, mainly family farms, which constituted a staggering 94.8% of EU farms in 2020. These data illustrate the magnitude of the challenge in embedding and implementing PLF in contemporary livestock agriculture in Europe. It can be concluded that there is enormous importance in the integration and involvement of stakeholders from the fields of social sciences in order to mediate the farmer–technology interface.
This entry is adapted from the peer-reviewed paper 10.3390/ani13182868