The Internet of Things or IoT is the technology of devices connected through the internet. At Agersens, it is the technology that allows eShepherd™ to enable virtual fencing. It is early days for the IoT industry. Only recently costs in hardware and improvements in connectivity have allowed IoT to be deployed in agritech.
As eShepherd is a world first system, there have been many challenges to address: here are three issues overcome to create a practical virtual fencing system for livestock.
Many IoT devices use mains power, while others contain a battery that must be recharged and, in some circumstances, replaced. The eShepherd collar is designed to overcome these shortcomings. It operates autonomously in the remote areas of Australia where the animal may be seen once every six months. An autonomous power supply is critical to success in Agtech. After extensive research, the Agersens power systems employs two solar panels together with a rechargeable lithium ion battery. This approach ensures the collar runs without a battery replacement.
Reading GPS signals is a power intensive activity. It consumes a majority of battery charge on a daily basis. The challenge lies in developing an extensive algorithm and strategies to manage the power levels of the battery based on a cow’s position and behaviour.
GPS sensor power draw has been a primary barrier to the development of a practical GPS based virtual fencing technology for livestock.
Agersens is at the forefront of developing on device data analytics and machine learning. Nifty algorithms detect animal behaviour along with position and velocity to reduce the need for frequent GPS readings. The system knows when the animal is not moving or exhibiting slow moving activities such as grazing.
eShepherd is power autonomous so that no battery changes or recharging is necessary. The farmer simply fits the collar and releases the animal. Agersens continues to develop on collar data analytics to continually increase the capability and utility of the system.
Range of Signal
Virtual fencing needs to work both across a spectrum of farm sizes: smaller farms such as dairies to large cattle stations where land is measured in square kilometres. To communicate across this spectrum of ranges, the LoRa (low range) bandwidth spectrum was chosen. LoRa is increasingly being used in IoT devices for its ability to carry a signal over longer distances compared to bluetooth or WiFi. The base station must handle hundreds of connections to the collars. The challenge was to integrate the three separate signals of satellite, base station and internet while ensuring a fast, stable connection throughout the system.
Thorough assessment of the property an important step to ensure line of sight connection to the base station to over a range of 20+ km. Agersens has tested several different configurations of base station to determine the optimal setup for each cattle operation. These locations can be from the desert conditions of Central Australia to lush mountainous areas of New Zealand.
While commercial trials continue with early adopters, good connectivity management has become one the strengths of the virtual fencing system from Agersens.
The connectivity challenge
Wirelessly connecting many devices to the Internet is still one of the biggest challenges in IoT. For virtual fencing, each collar needs to receive GPS signals from a satellite. Periodically, it needs to communicate position and other data to the base station. There may be hundreds of collars connecting to a base station.
The Agersens approach avoids some of the issues found with other IoT providers in the agricultural industry. The system is designed to operate without relying on mobile or cellular coverage in the area. Secondly, each unit does not rely on constant feeding of data via the Internet to a data cloud. Only the base station needs to be connected to the Internet and short term interruptions to internet connectivity of the base station do not impact operation of the virtual fence. This approach resolves many of the issues of IoT systems as mobile coverage is not necessary and short term loss of internet connectivity does not affect the system.
Once the eShepherd collar has received the virtual grazing area for its animal the collar is autonomous. It does not need constant connection to the base station. If the internet connection to the base station is temporarily lost, the system continues to operate effectively. Each collar “shepherds” its animal to stay within the virtual grazing area. When internet connectivity returns to the base station, the collar will once again check in with cloud system. With this approach, the collar is not dependent on constant internet, mobile phone services, or satellite connections to operate effectively.
The system must also handle huge data input from devices reporting in from around the world. It must be stored reliably and processed to provide real time reporting. The eShepherd platform was designed to be scalable and distributed across continents to handle high volume data feeds globally. Amazon Web Services was chosen to provide many of the online data services for Agersens. It is a data company that can handle the enormous amounts of data while providing scalability as well as redundancy within the platform.
As supporting IoT technologies continue to develop at break neck speed, Agersens is watching. We constantly test new technologies, materials and devices to increase power, capability, speed, and longevity. This in turn, enables us to deliver new and improved services to our customers.