Paintballing Pigs, Business Analytics and Real-World IoT Applications
Feb 12, 2019
For pork producers, economic survival hinges on efficient nutrition. At about 11 weeks of age, a pig enters a 16-week growth cycle when it consumes 6 to 10 pounds of feed a day. During that period, a farmer aims to get the average per animal weight of the pen to the “market ready” level of 280 pounds as cost effectively as possible.
Achieving that goal is more complicated than you might think. One problem is the cause/effect calculation of what a pig eats versus how quickly it grows. Mixing field scraps such as corn stalks together with purchased grain can save money, but are those savings offset by slower growth? Conversely, does investing in more nutritious (and more expensive) grain feed pay off in terms of faster growth?
Measuring growth by weighing animals within a feeding area can help refine the formula, but even a few particularly large or small animals can skew the average. Moreover, different hogs have different feeding patterns. Some like three square meals a day, while others are snackers who frequently visit the feeding area. This too adds complexity to the equation.
While expensive analytical tools can parse these issues, the economics of hog farming preclude investment in such technology. As a result, farmers have traditionally struggled with juggling these variables.
Enter the Internet of Things (IoT) and, more specifically, intelligent edge devices that enable insightful analytics and better business outcomes.
With inexpensive, easy-to-deploy weight sensors at feeding stations, a hog farmer can precisely track the weight and growth rate of different pens, as well as outlier (super skinny and super chubby) animals within a pen. By monitoring total weight gain as well as number of visits to the feeding stations, the sensors address the complication of varying feeding patterns. For example, if 10 hogs that each weigh 200 pounds all eat once a day, the sensors register a total weight of 2000 pounds for the day. If 5 of the hogs eat 3 times a day and the other 5 eat 10 times a day, calculating average weight against number of visits still yields a total of 2000 pounds a day. This would hold true even if someone were to add marijuana to the feed, turning all 10 pigs into stoners who visited the feed stall 20 times a day to munch out.
Tagging individual animals also presents a challenge – RFID is too expensive, and trials of retina scans have been ineffective. An alternative is to weigh the animals using a scale similar to the one in your bathroom. An IoT actuator deployed above the scale can apply logic in the edge gateway to fire a paintball to mark the outliers to be separated from the pen. (An ancillary application – entertainment for farm workers.)
So, let’s say a pork producer has used sensors to collect detailed aggregate data on what hogs eat and how fast they gain weight. What’s next? By linking sensors and actuators to an Internet connection via a gateway, (typically Ethernet, WiFi or cellular), the sensors can transmit small amounts of data to an intelligent platform for analysis. Getting the data to the cloud is imperative, so a robust edge capability can fall back to low-cost cellular when all else fails.
In the cloud, pattern recognition and machine learning capabilities glean needles of insight from haystacks of information. Specifically, a farmer can identify the precise mix of feed that delivers the optimal economic outcome, i.e., the fastest weight gain at the lowest cost to the most animals within the operation. Data analytics accounts for confounding variables such as unusually large or small animals and different types of feed. The result is greater precision in defining feed formulas and a higher ROI. And as more and more data are collected and analyzed, farmers can explore increasingly detailed and nuanced questions to drive ongoing refinement and improvement.
While few of us are hog farmers, the lesson here is broad and worthy of note. The connection of low-cost, data-collecting and data-transmitting sensors to intelligent analytical platforms is becoming increasingly viable – and valuable. This opens vast opportunities to redefine and rethink how a wide range of businesses operate – how restaurants are managed
, for example, how auto insurers track driver safety and how transportation and energy companies monitor assets in remote locations.
Discussions of the IoT tend to focus on ultra-sophisticated uses such as driverless cars and self-repairing jet engines. But more mundane applications are where the technology’s impact will perhaps be most transformational.