On Vision: Cutting Vision’s Cord
Sometimes in the vision system world, the real-world challenges are what lead to the best solutions. Our business has shifted over the last few years to serving more supply chain applications than industrial applications, as that’s where it seems the bigger problems are. As a solution provider, our business model is to solve problems. In serving that market, solutions have become more out-of-the-box and standardized at much lower price points than industrial systems.
For the supply chain, once your product is fairly standardized, the remaining pain point has often been running power and running Ethernet to put the system on the network. For some places, this has been painful. However in the customer’s defense, getting power and Ethernet run to more than 40 dock doors when you’re having trouble finding labor for your core business functions can be tricky. Nobody wants to tell the operations manager that employees didn’t load all the trucks today, but they did run some cables for some new system.
Like all good problems, this gets you thinking. The device we most often sell into the logistics market is a drive-through barcode scanner. It’s a standard vision system from a hardware perspective that runs an industrial PC and a camera. In total it’s about 60 W plus whatever monitor is plugged in. Your average solar panel has gotten to 400 W, so it could easily power the hardware when the sun shines. A lithium-ion or LFP battery that can run it for more than 12 hours costs ~$500 to run it when the sun goes down. There are amorphous solar panels that work on indoor light (think of a calculator) so you could just run 24/7 from indoor light. If you cut the computer from the equation, a lot of the CMOS cameras measure their power consumption in Milliwatts not Watts. If you want a computer at the edge for low latency, next generation edge device computers are more often 20 to 30 W or you could purchase a smart camera where some of the smaller models are now beneath 10 W. I’m not sure we will ever buy a 100-W or greater computer again. Many LCD monitors have gotten to be well below 10 W.
Similarly, 5G cell service is in many cases faster than USB 3.0. Obviously, it depends on traffic in the area and line of sight. But, the point is that wireless is getting to be faster than older generations of wired communications. So, you could just mount a camera using mW power and send images to a server independent of any local network connection. There would be no smart camera or computer needed.
When does the cord just get cut? Power tools used to all have cords. I remember when you had to run extension cords to wherever the weekend project was taking place. Now, they still exist, but I can’t recall the last time I’ve used one. There is still one in the house I grew up in. It works just fine. The hardware store still sells them—if you really look.
There are many clear benefits to cutting the cord. Cable routing and conduit take significant time to install. From a reliability perspective, challenges include securing the cables so they can’t wobble their way loose, protecting them from being pinched or cut, unused cable ports, and the ingress problems they create for dust and water. Additionally, remember the support hours spent trying to figure out which cable isn’t plugged in.
What if the future end user vision system was just a camera and 5G transmitter and small amorphous solar cell built into the device—not a bulky separate panel. It could be fully sealed from the factory so it would be fully water and dustproof. It would have one button basically for on-off. It would probably weigh 100 g, maybe 500 g (1lb.), so it could mount to the wall with sticky mount or a hook-and-loop mount. For the traditionalists, it could allow you to mount it with a drill—even a corded drill. It is true that cordless drills can’t produce quite the torque that the corded ones can, but we see which way the world has gone. Vision might be headed the same way. Industrial systems will have electrical panels and plenty of cords, but to take a picture, process it, and return a result—the core functions of a vision system—I’d argue no cord is needed. Free from cables, vision may break into all kinds of new applications.