By Kenneth Wong, originally published in Digital Engineering

If you think of IoT as an opportunity to sell more widgets and gadgets, your approach might be too simplistic, perhaps even fatally flawed. Speaking at the recent Sensors Expo and Conference in San Jose, California (June 22-23), Christopher Rommel, executive VP of IoT and embedded technology for VDC Research, pointed to the history of the Guinness Surger as a cautionary tale.

The Surger is a base unit that uses ultrasound waves to create the smooth, creamy Guinness foam. It debut around 2006 for £16.99, with the promise to let you enjoy bar-style draft taste at home. But to reap the Surger’s stated benefit, you’d have to also buy Surger-suitable cans, keep them chilled, fill the device with water, power it from an outlet, pour the beer into the right-sized glass at the right speed, and be patient while it activates the foam. If any of the steps is incorrectly executed, you don’t get the foam.

The gadget news and review site Pocket-lint summed up, “You have to ask yourself whether you are a dedicated enough Guinness follower to want to go through all this just to get a pint of the black stuff … When we decide to have a beer at home or even a cheeky one on a Friday night at the office, we don’t want to have to go through a 5-minute process every time.”

Compare that to solutions that monitor beer kegs with sensors, like SteadyServ. “Someone thought of a way to instrument beer kegs with sensors to figure out how full they are so [bartenders] know when to switch a keg,” said Rommel. And distributors love it, because they have a way to tell when [the bars] are about to run out.”

VDC’s 2016 IoT Survey

In VDC Research’s 2016  IoT Survey, 31% of the respondents said the initial catalyst for IoT initiatives came from the desire to “reduce cost for end customers.” Other leading motivations were to “enable end user deployment of sensor networks” (46%) and to “provide them with cloud-based monitoring and support” (38%).

Among the participants, 30% indicated they’re integrating more sensors of different types in their current projects, and 36% indicated they plan to do this in the next three years.

Rommel and his colleagues deduced, “predictive maintenance will drive need for traditional sensor types” and “new use case will drive need for machine vision solutions.” They believe “access to relevant expertise” is a critical factor to success — something businesses tend to underestimate.

“Security challenges and concerns” are rated as the top inhibitor to IoT, followed by “end user demand and readiness.” Even so, Rommel said companies “by and large don’t delay or postpone IoT projects due to security concerns.” One reason may be the cost of the fix. “If the fix is at a component level, the cost may be too prohibitive,” Rommel observed. Another is time-to-market and competitive pressure. “People are always fighting to meet deadlines; they’re already late. They also know their competition is rapidly making progress,” he said.

Machine Vision, AI Features

Sensors are generally perceived as devices that measure or detect pressure, heat, flow, chemical, sound, vibration, and other phenomenons. But the proposed use of depth-sensing cameras for input and response expands this traditional view.

Exhibiting in Texas Instruments’ booth, China-based INMOTION SCV pitches its 3D depth camera line as “high performance robot navigation and positioning solutions.” The SCV in the company’s name stands for “sensor-controlled vehicles.” Its Segway-like unicycles and personal mopeds have been “praised and used worldwide by politicians, celebrities, pop stars, and sports car enthusiasts,” according to the company.

At the most basic level, a depth-sensing camera could be deployed to measure distance. But products like NaviPack, INMOTION pointed out, can be used for indoor positioning, automatic obstacle perception and avoidance, and autonomous path planning. The centerpiece is an algorithm processor.

With a double-decker demo trailer parked right inside the exhibit hall, semiconductor maker NXP shows how its technology is powering smart homes, smart appliances, gyroscopes, and 3D printing. But Mike Stanley, who heads NXP’s algorithms and systems engineering sector, wants you to know you can train sensor-equipped system to do a whole lot more than sensing with machine learning.

In a recent blog post, he explores sensor data analytics with algorithms. “The starting point for this is often raw vibration data from an accelerometer,” he said. “The machine condition monitoring industry has been utilizing vibration data for many years to predict machine failure before it occurs … For cost reasons, machine monitoring was historically used only for very expensive machines that cannot tolerate unscheduled downtime. MEMS technology has now reduced the sensor cost to negligible levels, and the only thing standing in the way of further adoption is expertise and software availability.”

Machine learning and AI are also the reasons a software developer like Reality Analytics is at the sensor expo, hoping to convince attendees that sensing devices capable of some logic (in a manner of speaking) are much better than passive sensing devices that capture and spit out data but can’t do much. The company writes, “Our products use advanced, patented artificial intelligence techniques to detect real-world events in sensor and signal data, so apps and devices can take action.”

Looking Beyond Hardware

IoT is fueled by embedded systems, circuits, and sensors — a lot of hardware; yet, the value of IoT is not mindless sensor integration into products that may not benefit from it. It rests elsewhere, in business transformation, predictive maintenance, service offerings, and valuable insights into consumer behavior.

Some VDC Survey participants are looking at connected devices as a way to “bundle additional professional services.” Others are exploring “fees based on product usage.”

Rommel said, “[IoT] fundamentally changes how you go about monetizing your product, how you derive value.”