Evaluating Virtual Fencing Technology at Field Scale Today

Project Title

Virtual Fencing


Chris Foster (Two Story Robot) chris.foster@twostoryrobot.com

Jonathan Bowers, Caleb Sharp, Courtney Milligan, Bronwen Evans (Two Story Robot Labs), Bruce Miller, Craig Gauld (OneOak Design) Xavier Dumouli, Alex Naylor (OneOak Design), Chris Solecki (Tatalrose Ranch)

Status Project Code
Completed February, 2023 FRG.11.21


Fencing is costly, running between $1,560 to $4,660/km in AB and $17,270 to $20,000/km in BC. This cost
limits the adoption of both optimal grazing (e. intensive rotational grazing) and environmental (e.g., protection of sensitive habitats and riparian areas) management practices. Fences are prone to cutting and fire and moose and need maintenance. The lack of fencing is also a big deterrent to using cattle to graze crop residues, etc. on neighbouring operations.

Virtual fencing can solve these issues. The user defines GPS boundaries, accurate to within two meters, to fence cattle in or out of an area. The cattle wear a collar fitted with a battery, GPS antennae, LTE radio, and a “bump” to conduct electricity. When the antenna senses the cattle approaching a defined boundary, the collar emits audible electrical signals to discourage the cattle from approaching. The closer they get to the boundary, the stronger the signals get. The system works off LTE cellular technology when animals are in cell range, and GPS when they’re not. Pasture moves can be done remotely by re-defining the boundaries. The technology also indicates where the cattle are to enable ranchers to keep tabs on their animals.

We previously developed the user interface to visualize the virtual fencing; this is a desktop application with basic tablet functionality. We tested the software with ranchers in the field and received positive feedback.


  • Development of a second generation collar built using learnings from previous trials that address the hardware issues encountered with the previous version of the collar by making gen 2 more robust with improved functionality.

What they Did

We designed and developed 20 second generation collars. These collars are a brand new design, started completely from scratch, that offer significantly improved physical robustness and functionality over our previous prototype.

Our team went through a rigorous process that included the development of a thorough power model, design and testing of multiple product configurations, hardware design and development, manufacturing, firmware programming, API development, and experiment design. The collars are now lighter than our previous design with a smaller surface area and more comfortable strap. We additionally anticipate that the collars will be more affordable. Crucially, these second generation collars will resolve the hardware failures we encountered with our first generation collars.

We have performed a promising fit test in Kamloops, BC and will next be performing additional field testing in the Burns Lake area to inform further development.

What They Learned

An experienced rancher in the community performed a fit test with one of the new collars while our team monitored for support and to collect feedback. The fit test was recorded and photographs were taken. We identified some opportunities for improvement around the stability of the collar, as the collar could slide down to the side when the head was shaken aggressively. The stimulus terminals remained in contact with the animal and the animal was comfortable with the collar after a few minutes.

We have designed and planned a complete field test for April 2023 which will confirm that we have resolved the hardware issues we encountered with our first prototypes. Our goal is to confirm that cattle respond to the stimuli at the chosen parameters and then learn the association between the stimuli and the geofence. Once the field test is successful, rancher feedback about the product’s usability and technical data from the collars will inform on our priorities for future product development towards commercialization.

What It Means

We believe the next stage is a commercial generation of collars. After performing the primary field testing in April and processing the results, we will develop a concrete plan for development over the summer of 2023. Depending on the primary field test results, actual hardware modifications between the second and third generation may be either minimal or more significant, but we anticipate hardware modifications will be minimal and development will be largely focused on the control software.

We anticipate that development over the summer will lead to another small scale test in the Fall, followed by production of at least 100 but less than 1,000 collars. These collars would be tested for a complete season with multiple ranchers over the 2024 grazing season. If our 2024 season test was successful, the hardware would be ready for a commercial launch.