How Livestock Farmers are Using IoT for Animal Management

The precipitous rate at which the industrial revolution is prompting the pioneering and implementation of new technology is simply astounding. The transformation in the landscapes across nations and industries has reached unfathomable heights. Ideas that existed purely in science fiction movies like Minority Report (driverless cars) and Star Trek (3D printers) are being integrated into reality. No profession, sector, or part of the world is beyond the reach of these incredible changes.

And as we watch, the practice of agriculture is changing to become unrecognizable. It will not be more than a few decades before the current techniques will be recorded solely in the history books among the obsolete methods of our ancestors. From vertical farming to aquaponics being done in warehouses a tenth of the size of a small-scale farm, humans are now finding more sustainable and efficient methods to manage ourselves and our intake.

In this article, we will dive into the usage of IoT technology that is either currently available or in the process of development to improve the practice of commercial livestock farming. We will also discuss current platforms and processes which can be integrated to support IoT in the future.

What does IoT involve?

The Internet of Things is the interconnection and interaction of multiple devices and systems, independent of human intervention or stimulation.

Think of a livestock farm that is virtually self-automated. On such a farm, artificial intelligence will maintain regular processes, schedule bot or machine duties, check and identify flaws, and monitor all the animals, atmosphere, and elements for signs of deviation, disruption, or other malfunction. Sensors will collect data that will be requested by the AI at regular intervals. This data will be transferred, via a gateway to a central system, where it will be stored in a database. Subsequent refining and analysis will equip the AI to make calculated decisions. To implement those decisions, it will send commands to multiple machineries, pieces of equipment, and systems to begin or cease operation in a precise fashion. Inventory trackers will ping the central system when inventory is low and a new order will be placed with the suppliers. When the next batch of products is ready, the AI will hire transport or reach out to the distribution head to send another self-driven vehicle to collect the batch. Everything will be loaded, transported, managed, and tracked by intelligent systems which will report all information to a central overseeing system. This information will then be accumulated and reanalyzed in a database to identify potential improvements. Those improvements will be reviewed, designed, and executed by the central system, without the need for human involvement.

This entire process can be broken down into a few main functions: Data Collection, Transferal, Storage, Analysis, and Implementation. These self-management capabilities are not as futuristic as they may seem. Let’s discuss the current technology that has been and is being developed to support this capability in the future.

Cattle Data Collection and Monitoring

Precise results and good decision-making rely on the collection of consistent, high-quality data. The systems can identify physiological parameters which are indicative of certain conditions. These preliminary surveillance processes allow for data classification, orientation, and optimization.

The satellite-based monitoring of cattle has been gaining traction in the last 3 years, especially among large-scale Australian cattle ranchers. From basic data-collection systems such as ARGOS and GPS to advanced software developed by LoneStar and Moovement, tracking tags can now perform a variety of functions. Monitoring and analysis of special and temporal data on cattle location, movement, and interaction can provide benefits such as

• Checking their current physical condition
• Determining which animals are high-performing
• Measuring stress levels and how they can affect production and fertility
• Getting better financial and insurance options because of more accurate and reliable animal tracking.
• Faster data sampling: Precision Hawk’s agricultural drones can “gather data on 500 to 1,000 acres in less than a day.”

In many cases, particular animal movements can suggest diseases, medical conditions, and weakness. If the values collected exceed the regular parameters, this may reflect ill health or injury. In the case of lethargy, or minimal movement patterns, the system can alert animal caretakers to injuries, excessive weight, or the need for special attention. Furthermore, the GPS tag acquires information regarding the exact coordinates of the animal as well as the surrounding temperature, and stores it within the tag or collar. This is then downloaded with a wireless transceiver and transmitted to a central location.

Electronic detection systems installed at the milking stations can also recognize and monitor individual cows’ milk flow and yield. Heat detection systems such as Afimilk can even measure the electrical conductivity of milk. Based on this, they can alert the farmer of inflammation and potential mastitis. Pedometers are often integrated into foot tags to detect estrus, ill health, or weight gain. Infrared thermal imaging technology can be used as a screening technique to identify foot-and-mouth disease-infected animals. Digital decision support systems may also play an integral role in alerting farmers about suspected illnesses and advising them on response options.

Gateways

Put simply, IoT gateways connect all these above devices and systems and act as midpoints between the external hardware and the cloud (or other) network. This is useful as it allows farm managers to access and edit data from a single, central location while conveniently synchronizing the information. This is especially important as animal tags have a limited range in which to transmit information.

• Being able to connect to a gateway allows for maximum battery life as the tag no longer requires a great deal of additional storage space.
• Gateways facilitate data caching and streaming to heighten ease of access. Advanced versions can even perform edge computing which involves data optimization and pre-processing – summarizing, deduplicating, and cleansing of collected data – to improve its quality and functionality.

There are a variety of networks gateways can operate on, such as cellular, Wi-Fi, Bluetooth, ZigBee, or LPWAN, among others. While each of these has its benefits and drawbacks, it is important to look at which one is best suited for pastoral farms, which often span hundreds of square meters.

Bluetooth, Wi-Fi, and ZigBee all have a maximum network range of around 100 meters. In this, LPWAN acutely outpaces them by providing long-range communication over 10 – 40 km in rural areas. Cellular networks can also reach as far as 45 miles, although they require more power and consequently higher battery life. LPWANs are similar to WSNs (wireless sensor networks) as they both require little infrastructure and are scalable. However, while LPWANs are low-powered, WSNs are constrained in terms of power resources. They have a short lifespan because of the size of their battery. Despite using optical communication allows a lower SINR (signal-to-interference-plus-noise ratio) as compared to LoRa, WSN maintenance demands due to hardware constraints make them a less-than-ideal choice. LoRa (a subset of LPWAN) is a good choice as it adopts a star-shaped topology around devices and can be served by a single base station. Furthermore, it has the advantage of minimum investment and maintenance costs. The high MCL (maximum coupling loss) reached by LoRa and NB-IoT is of no consequence when applied in remote, open cattle ranches. It should be noted, however, that LoRa has this range because it utilizes unlicensed bands and its AES 128-bit encryption is much lower than the 256-bit 3GPP encryption that NB-IoT is built on. You can learn more information about IoT gateways here.

Storage

The primary and most widely-adopted server is the cloud. Private clouds are commonly used to store and manage corporate data. Using the cloud as a central system is beneficial for multiple reasons, one of which is the dual function of storage and analytics (which we will discuss further down).

• The data stored here is safely separate from the data within the system. This means that if the internal farm system is compromised, the cloud-encased data will be secure and vice versa.
• A farm is not the ideal location to house the massive infrastructure and hardware required for a data center. Data centers often warrant a great deal of energy consumption. The cloud server slashes all these unnecessary costs.
• A common issue that often arises from multiple entry points is the duplication or disharmony of data. All the raw data must be compiled, systemized, and synchronized so that it can be understood by unspecialized users.
• More specialized clouds even perform the advanced functions of a Database Management System (DBMS). This includes managing and presenting the information in a smooth and navigable format. Currently, DBMSs have the ability to self-improve – i.e., to review weaknesses, identify areas of improvement, and self-medicate to reduce the risk of cyberattacks, information masses, or other internal malfunctions.

Alongside cloud computing, edge computing is a viable and widely-adopted option.

What is Edge computing?

Edge Computing is a sub-set of IoT operations, which works alongside the Cloud. However, the bulk of the processing, usually done at the center, within the cloud analytics system, is transferred to the edge. The idea is for the computation to be performed as close to the source of the data as possible. This eliminates excess transmission load on the network and consequently reduces latency, allowing urgent insights to be deduced instantaneously. It also facilitates more accurate real-time responses since the lower network bandwidth often results in reduced image sizes and sample rates, or skipped frames in videos, when received by the centralized cloud. Edge Computing is deployable at remote locations with limited internet connectivity, such as in an open field, meadow, or cattle ranch, especially when having to transfer large datasets. It is ideal for cattle farms that gather in-depth information and make urgent, location-specific decisions.

Granular data like the respiratory rate, heart rate, grazing rumination, mobility, temperature, and milk quality can help farmers make decisions to improve resource allocation, recognize conservation opportunities, or prepare for upcoming disasters. With further capabilities, the farmer would soon be relieved of such decision-making burdens, as we will see in the analytics section.

Analytics

Weather patterns and climate disasters are becoming more and more unpredictable and unavoidable as a result of climate change. But the rate of human information processing is far outpaced by the learning ability of AI and deep neural networks (DNNs). While it takes time for predictions to disseminate among farming communities, AI might be able to accurately and keenly foresee the storm and the damage it will cause and take preventative measures before it even begins to form. This is because data collected and organized into datasets by database management systems are directly fed into the analytics software.

Let us delve into the applications of various modes of data analysis on pastoral farms. There are 4 main types of Data Analytics:

Descriptive Analytics

The process of:

• discovering similarities between symptoms to create a disease model
• identifying datasets to assist in tracking herd populations and the propagation of infectious agents
• separating abnormalities and ensuring a relatively heterogenous pastoral resource.
• finding correlations between variables
• identify the optimum feed time and release food accordingly, as is done by Tassal, a Tasmanian salmon producer.

Predictive Analytics

This involves leveraging artificial neural networks and machine learning techniques to:

• Categorize animals and herds by their future performance rates.
• Estimate the likelihood of disasters, opportunities, and the extent of their destruction or benefits given the farm’s current infrastructural condition.
• Uncover insights regarding vaccine and inoculation responses.

Prescriptive Analytics:

This is an advanced function whose focus is converting theoretical data into practical methods, such as

• The optimization of grazing rotation, transportation cycles, protocols for curbing the advance of outbreaks, etc.
• Minimize excesses and shortages or dispose of waste in an environmentally beneficial manner.
• Utilize by-products such as gelatin, leather, and internal organs in the most efficient and profitable manner.
• Finds ways to assuage delays in the supply chain by improving transparency and harmoniously managing inventory.
• Using the sensing system to solve problems that improve the agricultural ecosystem, as done by Australian agtech company, the Yield.

Implementation

As previously discussed, the most impressive and advanced aspect of IoT technology is its ability to act on the conclusions it has drawn and adaptively respond to rapidly evolving conditions. This would have to be done in conjunction with an AI system. What does this look like on farmland?

Corteva Agriscience’s fleet of drones can “offer immediate insights to diagnose agronomic, disease, and pest concerns.” Robotic animal herding systems are being developed to direct large groups of cattle over vast distances to designated locations such as shelters or barns during storms.

Upon discovery of the presence of contagious diseases, the central server will send instructions to command-based robots, which will quarantine infected animals. These command-based robots use actions defined by subsystems to execute instructions.

Re-fertilizing and cultivating grazed fields are massively important aspects of pastoralism. Exposed soil can be subject to soil erosion, which will impair its regenerative process and require more fertilizer in the long run. AI can act on grazing rotation forecasts by performing mulching as soon as herds move off a certain tract of land.

AI systems are also being developed to perform confirmatory diagnostic testing and create detailed disease maps. This means that without having to alert their human counterparts, AI would be capable of scheduling vet appointments and deciding on the best, most financially viable treatment option.

When it comes to decision-making, modeling of data and simulations can be performed to paint a clearer picture in the mind of the farmer. Scenarios may be simulated based on user-defined scenarios to reveal e.g., the average mass deviation of cows or the ideal design and feeding strategies.

To Wrap Up

It is clear that the growth of IoT technology has vast potential in the commercial agricultural and pastoral industries. There is so much more that we could further discuss, from self-sufficient dairy installations and feeder machinery to the predictive systems which can self-regulate farm operations. The bounds to which IoT can take farming production are practically limitless and only the future can unfurl this incredible capability to its full extent.

” Elianne Liong is a staff writer for Celeritas Digital.  She specializes in researching and publishing content related to a range of topics in the animal health and veterinary industry, including technology transformation, business processes, HR, data science, and advanced analytics. “