Someone asked me about automation controls, and where Nvidia and Qualcomm’s “Edge Computing” technology fit:
Industrial Automation Controls is dominated by two types: Industrial Controllers (Programmable Logic Controllers/ PLCs) and Industrial Processors (Programmable Automation Controllers/ Industrial PCs).
You can think of industrial controllers as really robust Arduino’s, and industrial processors as really robust Raspberry Pi’s.
The main difference is controllers primary function is basic mechanical controls processing with 12-24V inputs and analog signals and motion controls. They are very easy to program (primarily ladder logic, as well as structured text similar to C) and maintain and troubleshoot, and can run for a decade without problems. This is mostly “Factory Automation”. The top PLC player is Allen Bradley (also known as Rockwell Automation). Or Siemens.
Industrial Processors main function is processing lots of data. They have much greater processing power, and are typically doing “process automation”. Think any type of refining process requiring fine tuning and measurement, such as refining and mixing etc. These controls systems tend to be less stable and more difficult to trouble shoot, because their underlying programming architecture is more complex to achieve speed. For example, a window’s update may cause serious problems for the application you’re running. Many of these company’s have their own Window’s or Linux version that’s locked down to prevent these issues. They can also be a bit more difficult to maintain and troubleshoot. I don’t specialized in these types of controls, so I’m sure there are exceptions to everything said here. The top or leading industrial processor player is Beckhoff Automation. Or Siemens.
As the “Internet of Things” and “Smart Technology” become more ubiquitous in the manufacturing environment, we’re seeing more and more need for data processing, as a result of collecting data about the process for analysis and intelligent automated decision making. Whether that’s collecting sensor data, tracking and traceability data from barcode’s or RFID, recipe data, production information, process data, etc.
“Edge Computing” is a way to do data processing on a field level to make real time decisions, rather than passing it up stream or to a master controller/PC or server to be processed. This is a direct outgrowth of the whole IOT movement, where AI and “intelligent algorithms” are being applied to machine/robot process and sensor data to make intelligent decisions. Edge Computing requires high processing power, and hence a new type of controls solution with greater data processing capability.
The closest product we have to Edge Computing is our compact multi-function PLC. This allows you to have a dedicated controller in a field or sub-system to do controls processing, rather than have a centralized controller doing all the work. However, the FP0H is not doing typical Edge Computing. It’s just doing relatively simple field controls with minimal data processing, but it’s a step in that direction.
As Machine Vision and Advanced Sensing Technology continue to permeate manufacturing and commercial applications, and work with automated machine systems and robots to do intelligent decision making, more and more data is being collected, and a greater need to process that data quickly.
This is where Nvidia and Qualcomms technology is stepping in, and Edge Computing more generally.
William Dare 