#YJJ Architecture: Multi-level Feedback Control
Much of the discussion of the Internet of Things revolves around the myriad of intelligent sensors that can be used to collect data from almost anywhere (even pulse rate from your ear . But I really think the power of IoT will really be harnessed when data from all those sensors is used to make intelligent decisions and provide feedback control that improves the operation of whatever systems we are using.
I like to think of this in terms of Supervisory Control, a concept I first learned about and applied in implementing manufacturing and process control systems in the 1980’s.  Extrapolating that concept forward to the Internet of Things, I like to use the following diagram as a framework for discussion.
Â
Three levels of control are illustrated in this diagram:
1. Device Control
Each individual intelligent device may collect data from multiple sensors and exert control over the device. Â
For example, in a Yellow Jeep or other modern vehicle, the Engine Control Unit (ECU) depends on both sensors and actuators to control engine function. The following diagram illustrates how various input sensor are used by the Electronic Control Unit (ECU) to control various actuators (e.g. igniter, injector) and generate operator alerts (e.g. Check Engine Light).
Â
2. Gateway or Subsystem Supervisory Control
Gateways or subsystem controllers may be used to aggregate data from multiple devices and provide subsystem-level supervisory control over those devices.Â
For example, in the Yellow Jeep example, a video subsystem may aggregate video feeds from multiple cameras while providing synchronized control over video zoom, pan and tilt functions.
Â
3. Overall Supervisory Control
As data is gathered from various subsystems into the Cloud, that data can certainly be ingested and stored for historical analysis and visualization, but real time analysis of that data could be used to coordinate may subordinate systems. Â
For example, what if an army of Yellow Jeeps were exploring different parts of the world at the same time?  Each of these Jeeps may have one or more subsystems that collect data and perform their own local supervisory control functions.  However, data collected from all those Jeeps could be analyzed in the cloud in real time and fed back to the individual Jeeps to control how they are operated.  I’m not sure just what that supervisory control might be, but it is worthy of exploration.
Â
Please let me know what you think.
Roll on Yellow Jeep Journey!