In my role as Technology Officer for the Detron ICT Group I have the privilege of visiting and talking to other companies on an ongoing basis. One company I recently visited was Fujitsu where they showed us, and we talked about their cloud and IoT portfolio. Very interesting and exciting stuff to say the least. While the so-called Connected Cow was only one example of what they do when it comes to combining IoT and clout technologies, it’s a special one for sure and it definitely got my attention.
Before going into the bits and bytes let me first briefly explain the problem a Connected Cow tackles.
How it began
The story starts in Japan, a couple of years ago where a farmer was struggling with cattle (cows in this case) production. You need to know that for a cow to get pregnant the mating needs to be precisely timed. Every 21 days (on average that is, it can and will fluctuate) a cow is receptive to a male (a period known as ‘Estrus’) or ready to mate in order to get pregnant.
But there’s more. During the Estrus cycle, every 21 days on average there is a time window of somewhere between 12 to 18 hours in which the mating ritual would need to take place for the cow to get pregnant. Studies have shown that between 10PM and 8AM achieves the highest success rate.
Knowing this you can probably imagine how hard it can be for a farmer to monitor all of his/her cows. The average milk farm in the U.S. alone has over 10.000 cows, imagine that, 10, 15, or 20+ thousand cows (I’ve read that they are planning a monster farm with over 100.000 cows) who all need to be monitored somehow. And not only for mating purposes, but for potential health issues as well.
Of course, things like the artificial insemination (AI) of cows help a great deal, but only when the timing is right, as just explained. One of the main reasons why this technology has only been relatively successful. Historically the AI of cows is around 50% accurate, regarding the earlier mentioned Estrus cycle that is, which leads up to a 35% of pregnancy rate. Not bad, but no great either.
When Estrus approaches
This is where things get interesting. When a cow is ready to mate, and thus the Estrus phase approaches they start to show very different movement patterns. Let me explain. With a movement pattern, I mean the amount of steps a cow takes, or small jump even.
By constantly monitoring the cows, in the barns as well as while they are in the meadows a difference in movement patterns will be easy to spot, especially since the amount of steps/jumps increases are significant around the Estrus phase. This method has shown to be successful in 95% of all cases when it comes to detecting the mating cycles of, in this case cows, put it can apply to other farm animals as well.
Pedometers
Is what they use to measure the amount of steps/movements made by the (Connected) cows. This makes it extremely easy to know when AI is needed, when desired by the farmer of course.
Fujitsu has manufactured a Pedometer, specifically designed for just this purpose. It has a battery life of at least 5 years and is connected through an antenna/receiver (with a range of up to 300 meters) over to a 3/4G router, which is connected to the main network (base station) from where the data is send over to the Azure cloud. It is here where all data will be analysed (big data). As soon as an abnormality in the cow movement patterns has been detected an alert will be send to the (smart) phone, or PC/laptop etc. of the farmer in question. Next it will be up to him/her what to do with the received information.
As easy and simple as it is efficient.
The use of these pedometers has successfully increased cattle production up to 70%. The earlier mentioned 50% accuracy around determining the upcoming Estrus cycle went up to 95%, and the success rate regarding pregnancies rose from 35% up to 65+%. Impressive numbers.
But it gets better even. Over time they also discovered that through the change in cow movement patterns certain health issues can also be detected. Around 8 to 10 different deceases can be proactively detected.
The icing on the cake.
While studying the cow movement patterns in combination with the Estrus cycles they found that there is a certain time window surrounding the optimal time a cow should receive AI, which comes around 16 hours after the first changes in movement have been detected. If they inseminate the cow in the first half of this window, in 70% of all cases she will get pregnant with a female (mostly used for milk production). If the seamen is inseminated during the second half of that window she will get pregnant with a male (mostly used for meat production) again in 70% of all cases. Isn’t that helpful?
And there you go, a very interesting use case for IoT combined with the intelligence and big data analytical capabilities of the cloud, one of the real/true added advantages of cloud computing if you ask me.
I’ll continue to post more of these use-cases as I come across them, and I’m expecting to see plenty going forward.