Pig-rearing and data science, from SmartAHC

  • Sensors collect biometric data to identify optimal point of a sow’s oestrus cycle
  • No live pigs in the country was a challenge for Singapore-based precision farming startup
Pig-rearing and data science, from SmartAHC

 
AGRICULTURE has become an unappreciated industry in today’s digital world, with images of farmhands toiling for hours under the hot sun coming to the fore when the industry is mentioned.
 
Yet it remains an important industry, especially with increasing demand for protein. However, rising urbanisation has driven farms to remote locations, which brings more challenges.
 
Despite huge advances in automation and technology in amplifying farming efforts, the industry still relies heavily on manpower and experience to rear animals.
 
Smart Animal Husbandry Care (SmartAHC) is looking to change that, at least for the pig-rearing sector, where a labour crunch, lack of skilled farmhands, and increased demand for protein is making a potent combination of challenges.
 
Or business opportunity, depending on how you look at it.
 
Traditional farming requires a lot of manpower, and when it comes to animal husbandry, a certain degree of experience as well – for instance, in determining the oestrus or reproductive cycle of a sow, according to SmartAHC cofounder Howard Tang (pic above).
 
“The worker has to perform a series of tests to check if the sow is in heat,” he says, speaking to Digital News Asia (DNA) in Singapore.
 
These tests are subject to a worker’s judgement, with farms seeing a variance of results, Tang argues.
 
“It is the worker who judges whether the sow is in heat, and farms face the problem of this being subjective and based on the judgement of the worker,” he says.
 
This is where ‘precision farming’ startup SmartAHC comes in, bringing together sensors and data science to help establish the oestrus cycles of individual sows, to maximise production (or reproduction, if you prefer!).
 
“We help farm owners monitor their livestock in terms of health and oestrus cycle, trying to look for the best mating times,” says Tang.
 
Models and individuals
 
While data science can be used to generate a mathematical model to predict the optimal time for mating, the oestrus cycle in each individual animal is still dependent on factors which render these models inaccurate, according to Tang.
 
“The main problem with data analytics, we realise, is that it cannot be applied to everyone – as biological creatures, we are subject to a lot of biological differences,
 
“The traditional mathematical modelling method requires a sample – you conduct a test to gain a threshold, adjust the threshold for the best expected output, then apply it to the entire population.
 
“This is likely not accurate because there are simply too many compounding factors,” he says.
 
SmartAHC instead uses an algorithm which can remember the individual symptoms from individual cases in the study, Tang claims.
 
“The algorithm will automatically remember how the sow behaved during the last oestrus period, so the next time the algorithm sees the same conditions, it flags that the sow is in heat.
 
“Of course, it also cross-checks with the population – the artificial intelligence [behind the algorithm] can assign weightage on different sets of data, so we take the data from individual pigs and cross-check it against the population, and readjust the weightage to achieve the highest possible accuracy,” he adds.
 
Coupled with the world’s smallest Near Field Communication (NFC) temperature sensors (pic below) and other wireless sensors SmartAHC has built to be implanted into pigs, for low-cost and non-intrusive monitoring, data can be collected and sent to the cloud to be analysed.
 

Pig-rearing and data science, from SmartAHC

 
Challenges: No pigs, conservatism
 
What does one do when the sector you plan on improving is banned in the country you currently reside in?
 
That was the challenge SmartAHC faced. Pig farms have been outlawed in Singapore since 1989, hence no live pigs are available in the entire country.
 
“Because our product designing process was in Singapore, we had to buy a pig’s head to do product testing,” says Tang.
 
The company signed a collaboration deal with Sichuan Agricultural University in China to conduct tests on the pigs the university has, according to Tang.
 
“We couldn’t fly there quite often, so we sponsored two Master’s students to carry out our research, and we only got second-hand information on the testing,” he says.
 
This resulted in a longer product design cycle – it took an average of two months from doing a design to getting feedback on the new design.
 
“If we had a live pig next to us, we could keep modifying the design as we test it, and we could build an intermediate prototype instead of a brand new prototype every time,” laments Tang.
 
Agriculture is also a conservative market, and getting farms on board can be a challenge, according to Tang.
 
“We are dealing with a B2B (business-to-business) market. When we meet customers, the first question they ask is about our accuracy, which is between 85% and 95%; then they ask about our sample size, which is relatively small,” he says.
 
So SmartAHC has gone down the commercial testing route, where it collaborates with farms while enlarging its sample size and improving accuracy.
 
“They don’t have to pay for the device during this commercial testing period, but they have to pay a non-refundable deposit,” says Tang.
 
“If they are satisfied, they can subscribe to our product at an introductory price because they participated in our trial programme, which the farms are happy with,” he adds.
 
Plans and funding
 
SmartAHC plans to increase the accuracy of its solution to convince more customers to come on board with commercial trials, according to Tang.
 
“In the lab, we are achieving an accuracy rate of 95%; in the commercial tests, we are expecting an accuracy of a minimum 85% with a sample size of 3,000,” he says.
 
“Once we achieve that, we will launch into the market,” he adds.
 
The business model remains a subscription, based on the number of sows being monitored and other additional functions.
 
As for funding, SmartAHC is talking to potential investors in Singapore as well as overseas, according to Tang.
 
“It [response] has been quite positive, as we are bringing electronics into agriculture, lifting the eyebrows of investors in Singapore; we are also bringing in artificial intelligence to agriculture, which is lifting the eyebrows of overseas investors,” Tang declares.
 
“We are using AI in a neglected field, which is how we piqued the interest of investors,” he adds.
 
The startup has already secured undisclosed seed funding from the Small World Group incubator, according to Tang, and is looking to raise an undisclosed amount of funding.
 
“We are trying to raise a round as there is high pressure on the cash flow if there is any delay in the commercial tests.
 
“The investors called this a ‘mid-angel round,’ which sums out our situation in which we are not angel or seed, and not pre-Series A or Series A,” he adds.
 
Related Stories:
 
Elixir Capital lead investor in US$7mil agriculture big data play
 
Indonesia’s eFishery raises undisclosed pre-Series A funding
 
 
For more technology news and the latest updates, follow us on TwitterLinkedIn or Like us on Facebook.
 

 
Keyword(s) :
 
Author Name :
 
Download Digerati50 2020-2021 PDF

Digerati50 2020-2021

Get and download a digital copy of Digerati50 2020-2021