Why are you hearing so much about sensors? According to Machine Design, sensors are going to be a critical part of the Internet of Things (IoT). Big Data, neural networks, smart machines, and artificial intelligence all require “source data.” In short, a great amount of the IoT discussion comes down to sensors.
Sensors of all types and sizes will be needed to generate the source data upon which the IoT’s intelligence will largely be built. Companies that make the most progress in the next 10 years in embedding and augmenting their hardware and electronics with data gathering and generating capabilities will likely be the market leaders in the following decade. What exists today is perhaps a tenth of what is coming.
Morgan Stanley echoes the significance of sensors in the near future. It points out that microcontrollers (MCUs), the tiny sensors that enable connectivity and control in all the “things” around us, have now become so inexpensive—at around $1 apiece—that they can be incorporated in just about everything, from industrial machinery and home appliances to wearable devices and even clothing.
In a recent survey of more than 100 key decision-makers who represented automotive, consumer electronics and industrial manufacturing companies, Morgan Stanley Research found that more than 90 percent of them are baking connectivity technologies into their designs. This marks a turning point for IoT, the next generation of personal computing, users and their environment.
Consumers and companies ultimately stand to reap the benefits of this new technology. In the near term, the semiconductor industry could be one of the biggest winners. Every 10 percent increase in connectivity using MCUs could add roughly $1 billion to the chip-making industry.
Given the preceding information, sensors clearly are expected to play a significant role in IoT which helps consumers, companies and industries. However, we should still look at them from an objective viewpoint, because they can mislead. Sensors, like the devices, machines, etc. that incorporate them, can make errors.
As Martin Willcox of Teradata points out, sensors sometimes lie. This often comes as a big surprise to business and IT people who tend to believe everything smart devices record can be assumed to be true, because smart devices never come to work hungover or distracted. Talk to the engineers who build and deploy sensor networks and you will rapidly discover that nothing could be further from the truth.
Willcox illustrates his point about sensor fallibility with heart-rate-monitor and desert-oil-field examples. More specifically, his heart-rate monitor falsely showed – in the middle of a 30-minute treadmill workout – that his heart rate was 70 bpm and stayed exactly that same way for several minutes. With desert oil field sensors, he explained that the extreme temperature range that the sensors endure often leads them to “drift” almost as soon as they are deployed and installed.
Consequently, sensors and sensor data do hold great promise, but they still require pragmatic oversight. Willcox emphasizes you can’t assume that machine-generated data are accurate, complete and consistent, just because they are produced by a machine. Furthermore, don’t assume that a low per-unit cost makes sensors easy to replace if they consistently make errors (i.e., fail). In some cases (e.g., desert oil field sensors), the cost of lost production time needed to replace them may not be so low.
(Photo: Arenas of IoT, Flickr)