An Overview of Control Loops

Aug 18, 2022

A brief introduction to systems and control loops.

status of this document: MOSTLY DONE

TODO: Link to glossary, spell check…

It’s All About Loops

Before diving into O-RAN architecture, it’ll be helpful to take a step back and talk about control loops because RAN in general, and O-RAN in particular is built upon control loops.

We’ll dive into what kinds of control loops O-RAN has later, in greater detail. Yet, before we even get there, it’s only fitting to talk about what a control system is and how control loops work.

But before we even dive into all these, what is a system in the first place. And what is a control system?

This Is Just an Overview

Control Theory and Digital Signal Processing are extensive disciplines that take years to learn and master. This article you are reading will give you just enough information to find your way through the contents of this website. If you want to drill down deeper, please check the References and Further Reading section at the end of this article.

What Is a System?

A system can be defined as a collection of components (or blocks, or objects) grouped together to perform an action (or execute a process, or attain an objective)

Another definition of a system is a group of elements connected (typically sequentially) to perform a specified task or function.

Here’s a block diagram of a simple system:

System: Block diagram.

Keep in mind that, systems don’t have to be physical entities. They can also be abstract concepts often used in economics, mathematics, or algorithmic constructs in computer sciences.

What Is a Control System?

A control system is a system where the output is controlled by adjusting the input.

Another definition of a control system is a combination of subsystems and processes working together to obtain the desired output given a specified input.

A control system defines a clear relationship between the input and output of what it’s controlling.

A control system that takes an input and some external disturbance to generate output.

The above diagram shows the control system as a function of input and external factors modeled here as a disturbance.

Isn’t “Disturbance” an Input Too?

The difference between disturbance and input is that you typically have control over the input, whereas the disturbance is an external factor (like the speed and direction of the wind, ambient temperature, humidity, atmospheric pressure… etc.) that we don’t have any direct control over.

For an example of a control system, let’s consider an elevator: In an elevator, when you press a button (say, 3), the elevator takes you to a floor (3rd floor, in this case).

The elevator’s control system takes the input signal from the button you pressed, runs its logic, then operates the motors of the elevator to take you to the destination floor (which is the “output”).

Types of Control Systems

There are several ways to categorize control systems. Some frequent ones are:

  • Open-loop and closed-loop systems,
  • Linear and non-linear systems,
  • Continuous and discrete systems,
  • Digital and analog systems,
  • And Time-dependent and time-independent systems.

In this article, we’ll look into feedback-based control systems; namely open-loop and closed-loop systems.

Open-Loop and Closed-Loop Systems

Open-loop and closed-loop systems are collectively called feedback-based systems.

But what is a “loop” anyway? A loop—or, in our context, a control loop—can be thought of as a process manager designed to keep the system it’s controlling in the desired state.

Open-Loop Systems

The control function does not depend on the system’s output in an open-loop system. So, for example, when you set a timer for the clothes drier to operate for an hour, the drier will turn off at the end of the hour even if the clothes in it are still damp.

Example of an open-loop system.

Moreover, suppose there is a disturbance in the system, for example. Let’s say, the disturbance is some insulation damage leading to a more-than-usual amount of heat leaking out of the system. Then, the drier will need longer than usual to dry the clothes. In the end of the hour, the clothes will be damper than expected, and thus they’ll need extra time to be dried.

In our open-loop system, this extra information is not considered. Therefore, the system does not adjust the timer to give the system more time to operate. The system does not take the clothes’ current dampness state into account.

Open-loop systems cannot tolerate disturbances.

Open-loop systems are typically used when the output is hard to measure or when the output is not immediately available.

Closed-Loop Systems

On the other hand, in a closed-loop system, the control action also takes the system’s output as a parameter. An example is an air conditioner which stops heating the room when it reaches a specific temperature.

A closed-loop system is a better fit for automation scenarios. The reason for that is because it can tolerate disturbances in the system better and constantly strives to eventually push the system to the desired state.

An AC unit is an example of a “closed-loop” system.

The astute reader might have realized that an open-loop system can be converted to a closed-loop system by introducing a feedback loop that feeds back the system’s output into itself—hence the “closed loop.”

Principles of closed-loop systems are becoming more prominent in modern system design. For example, orchestration systems such as Kubernetes utilize closed control loops to orchestrate and reconcile large scale distributed systems. Again, closed loops are used in the Internet of Things (IoT), where the devices ingest data from the various sensor to readjust and realign their current state and make decisions based on the system’s current state. Similarly, AI/ML algorithms can be thought of as self-learning closed-loop controls.

The Benefits and Liabilities of a Closed-Loop System

The benefits of closed-loop systems are:

  • They can compensate for deviations and external factors better.
  • They produce a more reliable and stable output.
  • They are generally more resource-efficient.

The liabilities of closed-loop system are:

  • They are more complex than their open-loop counterparts.
  • They often require tuning and integration with external systems.
  • They are susceptible to oscillations that occur in second (or higher) order systems.
  • Sensor failure can result in indeterminate outcomes and undesired system performance.

Control Loops in O-RAN

Open RAN (O-RAN) has three main control loops:

  • Realtime control loop,
  • Near-realtime control loop,
  • And, non-realtime control loop.

We’ll discuss them in detail in the following sections. But, for now, suffice it to say that all of these control loops in O-RAN are closed-loop systems.

Conclusion

This article briefly overviewed various kinds of control systems and control loops and how some of these control loops apply to O-RAN.

We’ll talk more about O-RAN in the upcoming sections, and control loops will be critical to understanding O-RAN. That’s why we introduced them as early as possible.

References and Further Reading

← previous
O-RAN Architecture
Introduction
next →
O-RAN Use Cases
Introduction