Disturbance Rejection in Feedback Systems

Imagine students is riding a bike on a windy day 🌬️. A gust of wind pushes the bike to the side, but students steers a little to stay on the path. That idea is the heart of disturbance rejection in control systems: when something unwanted affects a system, feedback helps push the output back toward the desired value. In Control and Mechatronics, this is one of the biggest reasons feedback is used at all.

Introduction: Why Disturbance Rejection Matters

A control system takes a desired input, called the reference or setpoint, and tries to make the output match it. In real life, outside forces often interfere. These unwanted influences are called disturbances. They can come from changes in load, temperature, friction, wind, voltage fluctuation, sensor noise, or sudden external pushes.

The key learning goals for students in this lesson are to:

explain what disturbance rejection means and why it is important
identify how disturbances enter a system and affect the output
compare open-loop and closed-loop behavior when disturbances occur
connect disturbance rejection to reference, error, and output signals
use examples to understand why feedback improves performance ⚙️

The main idea is simple: a feedback system measures the output, compares it to the reference, and corrects errors caused by disturbances.

What Is a Disturbance?

A disturbance is any unwanted input that pushes a system away from its desired behavior. It is not the command you want the system to follow. Instead, it is something that interferes with the system after the command is given.

For example:

In a room heater, opening a window lets cold air in.
In a motor speed control system, adding a heavier load slows the motor.
In a drone, a wind gust changes its position or angle.
In a water tank, water being used from another pipe lowers the level.

Disturbances can act at different points in a system. Some affect the process or plant directly, while others affect the sensor or measurement. In many cases, the system cannot stop the disturbance from happening, but it can respond to reduce its effect.

A useful way to think about it is this: the reference says what should happen, and the disturbance tries to spoil it. Feedback helps the system fight back 💡.

Open-Loop Systems and Disturbances

In an open-loop system, the controller does not use output feedback to correct the result. The system sends a command and assumes the output will be correct.

For example, a toaster that heats bread for a fixed time is open-loop. If the bread is thicker than usual or the room is colder, the toast may come out too light or too dark. The toaster does not measure the bread color and adjust automatically.

Because there is no output measurement used for correction, open-loop systems are usually poor at disturbance rejection. If a disturbance changes the output, the system has no built-in way to notice and respond.

A simple example in mathematics can help. Suppose a process output is

$$y = u + d$$

where $u$ is the control input and $d$ is a disturbance. In an open-loop system, $u$ is set without checking $y$. If $d$ changes, then $y$ changes too. The system cannot automatically compensate.

This is why open-loop systems are often used only when disturbances are small or when very precise output is not needed.

Closed-Loop Systems and Disturbance Rejection

A closed-loop system uses feedback. It measures the output, compares it with the reference, and uses the difference to make corrections.

The difference between the reference and output is called the error:

$$e = r - y$$

where $r$ is the reference and $y$ is the output.

If a disturbance changes the output, the error changes too. The controller sees this and adjusts the control input to push the output back toward the reference. This is the main reason closed-loop systems are better at disturbance rejection.

Real-world example: room temperature control

Suppose students sets a heater to keep a room at $22^a0^eC$. The reference is $r = 22^a0^eC$. If a door opens and cold air enters, the room temperature drops. That temperature drop is a disturbance.

A thermostat detects the lower output temperature, so the error increases. The controller turns the heater on more strongly. As the heater adds heat, the output moves back toward $22^a0^eC$.

Without feedback, the room could stay cold for a long time. With feedback, the system actively resists the disturbance.

Real-world example: motor speed control

In a conveyor system, a motor may need to keep a belt moving at a constant speed. If more boxes are placed on the belt, the load increases. That load is a disturbance. A closed-loop controller measures the motor speed and increases the drive signal if the speed drops.

This shows an important rule: the larger the effect of the disturbance, the more the controller must correct.

How Feedback Architecture Supports Rejection

Most feedback systems follow a basic architecture:

A reference is given.
The system output is measured.
A comparator finds the error using $e = r - y$.
The controller creates a control signal to reduce the error.
The plant or process responds.
The cycle repeats continuously.

This loop is what makes disturbance rejection possible. If a disturbance pushes the output away from the reference, the error signal becomes nonzero. The controller uses that error to act in the opposite direction.

In many systems, this happens very fast. For example, a drone flight controller may correct tiny tilts many times each second so that wind disturbances do not cause a crash. The more quickly and accurately the system reacts, the better the disturbance rejection.

The Role of Reference, Error, and Output Signals

To understand disturbance rejection, students should keep the main signals straight:

Reference $r$: the desired value or target
Output $y$: the actual measured result
Error $e$: the difference between desired and actual, given by $e = r - y$

When a disturbance changes $y$, the error changes. That is the signal the controller uses to respond.

Here is a simple sequence:

The reference is a target water level.
The output is the current level.
A disturbance drains some water.
The output falls below the reference.
The error becomes positive.
The controller opens a valve to add more water.
The output rises back toward the target.

This pattern is the same in many mechatronic systems, even if the physical details are different.

Why Feedback Usually Rejects Disturbances Better Than Open Loop

Closed-loop systems do not make disturbances disappear instantly, but they usually reduce their effect a lot. This happens because feedback creates a self-correcting action.

A useful way to compare the two is:

Open-loop: disturbance changes output, and the system does not automatically fix it.
Closed-loop: disturbance changes output, error appears, and the system corrects itself.

Feedback also helps because the correction is based on the actual result, not just the expected result. That makes the system more robust when conditions change.

However, feedback is not perfect. If the controller is too slow, too weak, or badly tuned, disturbance rejection may still be poor. Also, too much feedback can cause overshoot or oscillation. So the goal is not just to use feedback, but to use it well.

Example With Simple Control Thinking

Suppose a tank should stay at a water level of $h_r$. The measured level is $h$, so the error is

$$e = h_r - h$$

Now imagine a drain opens and removes water. That is the disturbance. The level $h$ drops, so $e$ becomes larger. A controller opens the inlet valve more to bring the level back up.

If the controller is strong enough and the system is stable, the output returns close to $h_r$. If the controller is weak or too slow, the tank may stay below the desired level for too long.

This example shows the essential purpose of disturbance rejection: maintain the output close to the reference even when the world changes.

Conclusion

Disturbance rejection is a central reason feedback exists in Control and Mechatronics. Disturbances are unwanted influences such as load changes, wind, friction, and temperature shifts. In open-loop systems, these disturbances can move the output away from the target with little or no correction. In closed-loop systems, the output is measured, compared to the reference, and corrected through the error signal.

For students, the most important takeaway is that feedback makes systems more robust. It does not stop disturbances from occurring, but it helps the system respond so the output stays closer to the desired value. That is why disturbance rejection is a key part of Feedback Fundamentals 🔧.

Study Notes

A disturbance is an unwanted influence that changes a system’s output.
Common disturbances include load changes, wind, friction, temperature changes, and supply variation.
In an open-loop system, the controller does not use output feedback, so disturbance rejection is usually weak.
In a closed-loop system, the output is measured and compared with the reference.
The error is given by $e = r - y$.
If a disturbance changes the output $y$, the error changes, and the controller responds.
Feedback systems usually reject disturbances better because they correct based on actual output.
Disturbance rejection is important in heaters, motors, drones, tanks, and many other mechatronic systems.
Good disturbance rejection depends on the controller, the plant, the sensors, and the tuning of the system.
Feedback improves robustness, but it must be designed carefully to avoid instability or excessive oscillation.