The application of the Internet of Things (IoT) to the world of agriculture has seen something of an upsurge in the last few years. There have been projects and companies offering to connect eels, vineyards, bees, and trees. Two years ago at Mobile World Congress in Barcelona both Deutsche Telekom and NTT Docomo showed off connected cow services, and it’s hard to forget a company called Findmysheep.
Nevertheless we have been impressed by a new offering from Israeli start-up MOOnitor. MOOnitor’s solution is a combination of: a cow collar with an accelerometer, solar panels and communications modules; localised software, running on a low power microcontroller to process and compress the data, to save on data traffic; a cloud platform for further analytics; and a presentation layer for a smartphone app and the web. Communications involve a one-directional satellite link that uses the Globalstar system and a bi-directional LoRa based RF link. Apart from a low power accelerometer, the only local sensor used for the application is a GPS receiver. There is a temperature sensor too, but that is only used to monitor the health of the equipment and plays no part in providing the application’s intelligence. LoRa is used, for collar maintenance when necessary, and the Globalstar link is used to send data to the cloud.
Much of the solution’s cleverness consists in its ability to conserve scarce resources. Rather than simply gather data and then send it to the cloud for heavy-duty processing, much of the work is done locally. Accelerometer measurements are converted, using an algorithm running on the local microcontroller, into numeric values corresponding the energy used by the cow during a short time period. The algorithm results are further categorized into cow activities: Resting, Grazing, and Walking.
The high resolution temporal data is summed on a daily basis, representing the amount of time that the cow has spent grazing, walking and resting. Once a day, shortly after midnight (local time, synced by the GPS), the collar sends a 36-byte message to the cloud platform. This message is derived from analysis of ~50MB of accelerometer data and three time-stamped GPS records, taken at set periods.
The solution conserves battery life as well as data transmission. Hardware that is not being used goes to sleep, so that it consumes little power. It can work for seven days on its battery alone, and the solar panels on the collar are sufficient to completely recharge the batteries in two hours. The collar itself is designed to last for ten years in harsh outdoors conditions. The cloud platform is able to make sense of the activity data to infer information about each cow. The individual cow’s measurements are compared to its own history and to those from other members of the herd. A sick cow will spend fewer hours grazing; peaks in walking activity indicate a cow in oestrus. A high proportion of repeated oestrus events indicates that cows have not conceived, and can therefore be an indication that of mis-performing breeding bulls. Cows’ activities can even be used to estimate pasture quality and herd energy balance.
A future upgrade to the collar’s software may enable the detection of heat stress in the cow. Neck movement during panting has a distinctive periodic nature. This distinct inertial pattern associated with panting can be detected locally after transformation to the frequency domain.
The ranchers app, and the web-based portal, allow farmers to see the key metrics for their herds, including the quality of the pasture, the location of individual cows (which can be located in real time via a local LoRa dongle), and which cows are pregnant, lactating or in oestrus. A key metric is the energy balance of the individual cows and the herd as a whole. Positive energy balance is a strong indicator of the cow’s welfare. Energy balance is key to the cow’s successful reproduction cycle from conceiving to calving and healthy calf at weaning. The most critical period is right after calving, when the newly born is lactating while the cow should get pregnant. . Attending to the cows’ needs will eventually increase the number of weaned calves.
MOOnitor is distinguished from its competitors by the very clear focus on a specific segment of cattle farming – the production of weaner calves for beef, on highly extensive farms where direct human monitoring of the herd is impractical. Although there are other ‘internet of cows’ solutions available, most of these are oriented towards dairy farming. Instead MOOnitor focuses on extensive cattle producers in Australia, the US and South America. It says that competitors are either from the dairy industry, which currently uses sensors in conjunction with monitoring milk quality, or from animal tracking companies, which have emphasised position reporting and crude measures of activity.
MOOnitor offers beef cows producers a straightforward business case. Herd performance is measured in terms of yield percentage – the proportion of cows that produce a weaned calf each year. An average extensive herd can deliver yields of 50-60%, but one that is better managed can reach yields of 85-90%. Better management, though, usually involves lots of hands-on human intervention, which is expensive; so an IoT data collection and analytics platform is obviously of interest, because it can help the farmer deliver more calves per year more profitably.
The worked example it provides is for a herd of 250 cattle, with one in five cows equipped with a monitoring collar. MOOnitor says that beef herds are for the most part managed as a collective, so monitoring some cows provides statistically reliable data while keeping costs down. This approach makes for a one time cost of USD180 per collar, and USD9,000 for the 50 cows. The farmer pays a monthly per collar fee of USD10, making the opex for the year USD6,000. Since each calf is worth USD850 at the point of sale, a 10% increase in yield (that is, 25 extra calves from the 250 cows) produces an extra income of USD21,250 – which compares well to the first year system costs of USD15,000.
At the moment MOOnitor, founded in August 2014, is a very small start-up, only employing its (admittedly impressive) three founders and a very few other part time employees , most of whom are rewarded with options rather than salaries, and with subcontracting relationships in place with another Israeli company for the small-scale manufacture of the collars. It has raised USD300,000 from angel investors to date, and is looking to increase this to USD1m through a first round funding exercise currently under way. It has a road map in place for further developments of its product, and it has a small-scale pilot in the north of Israel.
A research version of the collars, which provide higher resolution temporal activity and location data is used by the Department of Animal Production and Pastures at the School of Agronomy in the Universidad de la República Uruguay. Another research system will be used in the Department of Plant and Soil Sciences at the University of Kentucky.
MOOnitor aims to find a global distribution partner that already has reach into its chosen sector, and can provide the professional services and training that users of its services will require. The company is now exploring possibilities for an international distribution network as preparation for first commercial sales during 2016.