Feedback Architecture

students, welcome to Feedback Architecture in Control and Mechatronics 🤖🔁 This lesson explains how a feedback system is built, how signals move through it, and why this structure matters in real machines. By the end, you should be able to identify the main parts of a feedback loop, explain the roles of the reference, error, and output signals, and connect this idea to everyday systems like cruise control, thermostats, and robot arms.

What feedback architecture means

Feedback architecture describes the arrangement of parts in a control system and how information flows from one part to another. In a feedback system, the output is measured and compared with a desired value. That comparison helps the system decide what action to take next 🔧

A basic feedback structure includes these main parts:

Reference input $r(t)$, the desired value
Comparator or summing junction, where the desired value is compared with the measured output
Error signal $e(t)$, the difference between desired and actual values
Controller, which uses the error to decide a control action
Plant or process, the system being controlled
Output $y(t)$, the actual result of the process
Sensor or measurement device, which sends output information back to the comparator

The heart of feedback architecture is the comparison step. If the output is not what we want, the system creates an error signal and responds. This is why feedback systems are useful in engineering: they can correct themselves when conditions change.

A simple mathematical relationship for the error is:

$$e(t) = r(t) - y(t)$$

If the output is below the reference, the error is positive. If the output is above the reference, the error is negative. In many real systems, the measured output is first passed through a sensor and may be written as $b(t)$ instead of $y(t)$, so the error becomes $e(t) = r(t) - b(t)$.

The signal path in a closed-loop system

A closed-loop system is a system where the output is fed back into the input side of the system. This is the classic feedback architecture used in control engineering. The signal path usually works like this:

A desired input or reference is set, such as a target speed of $60\ \text{km/h}$.
A sensor measures the actual output, such as the car’s current speed.
The measured output is sent back to a comparator.
The comparator calculates the error signal.
The controller uses that error to adjust the plant.
The output changes, and the cycle repeats.

This loop continues many times each second. Because of this repetition, the system can respond to changing conditions, like hills, wind, or load changes in a vehicle 🚗

Feedback architecture is not just about drawing arrows. It is about understanding how each signal has a purpose. The reference tells the system what is wanted. The output tells the system what is actually happening. The error tells the system how far off it is. The controller uses that error to reduce the difference.

A practical example is a room thermostat. If the reference temperature is $22^\circ\text{C}$ and the room is at $20^\circ\text{C}$, then the error is $e(t) = 22 - 20 = 2^\circ\text{C}$. The thermostat uses that error to turn the heater on. When the room gets closer to $22^\circ\text{C}$, the error becomes smaller and the heater output is reduced.

Open-loop and closed-loop architectures

Feedback architecture is easiest to understand when compared with open-loop systems. In an open-loop system, there is no feedback path from output back to input. The control action is chosen without checking the actual result.

For example, a toaster with a fixed timer is an open-loop system. If the bread is thicker, thinner, colder, or already partly toasted, the toaster does not measure the toast color before deciding when to stop. It simply follows the timer setting.

A closed-loop system, on the other hand, checks the output and adjusts. An oven with a temperature sensor is closer to closed-loop control because it measures the temperature and turns heating on or off to stay near the set value.

The key difference is this:

Open-loop: no feedback, simpler, less accurate when conditions change
Closed-loop: uses feedback, more accurate, can correct disturbances

In Control and Mechatronics, feedback architecture is especially important because many systems must react to real-world changes. A robot arm lifting a package, for example, may need to adjust motor torque if the package is heavier than expected. That correction is possible because the output is measured and compared with the desired motion.

Reference, output, and error signals

The three most important signals in feedback architecture are the reference, the output, and the error.

Reference signal

The reference signal is the target or command value. It can be a speed, position, temperature, voltage, pressure, or any measurable quantity. In equations, it is often written as $r(t)$.

Examples:

A drone altitude target of $100\ \text{m}$
A motor speed target of $1500\ \text{rpm}$
A tank water level target of $80\%$

Output signal

The output signal is the actual measured result, often written as $y(t)$. It comes from the plant or process. The output may change because of the controller’s action or because of outside disturbances.

Error signal

The error signal is the difference between the reference and the measured output. It can be written as:

$$e(t) = r(t) - y(t)$$

The error is the “gap” the system tries to reduce. If the output is exactly equal to the reference, then $e(t) = 0$. That means the system is on target 🎯

A real-world example is speed control in an electric scooter. Suppose the rider sets a target speed of $25\ \text{km/h}$. If the scooter is only moving at $20\ \text{km/h}$, then the error is $25 - 20 = 5\ \text{km/h}$. The controller increases motor power until the speed gets closer to the reference.

Why feedback architecture matters in mechatronics

Mechatronic systems combine mechanics, electronics, sensing, and computing. Feedback architecture is what helps these parts work together smoothly.

Without feedback, a system may work only under ideal conditions. With feedback, it can deal with changes like:

extra load on a motor
friction changes in gears
temperature shifts in electronics
wind or vibration in moving systems
wear and tear over time

For example, an automatic door uses sensors to detect whether someone is nearby. The desired behavior is for the door to open when needed and close safely afterward. Feedback helps the system confirm what is happening and adjust its motion.

Another example is a camera stabilization system. If the camera shakes, sensors detect the movement and motors make small corrections. The output becomes steadier because the error between desired and actual camera position is reduced.

Feedback architecture also helps explain why some systems are more stable than others. If the feedback is designed well, the output approaches the reference smoothly. If the design is poor, the system may overshoot, oscillate, or react too slowly. Those ideas belong to later control topics, but they begin with the architecture itself.

A simple block diagram view

A feedback system is often shown as a block diagram. Even without drawing it, you can imagine the signal flow:

$$r(t) \rightarrow \text{Comparator} \rightarrow e(t) \rightarrow \text{Controller} \rightarrow \text{Plant} \rightarrow y(t)$$

Then the measured output returns through the sensor back to the comparator.

This structure highlights an important idea: the controller does not directly compare the reference to the physical system. It compares the reference to a measured version of the output. That matters because sensors are never perfect. They may add noise, delay, or small measurement errors.

For example, a level sensor in a water tank may report a value that is slightly different from the true water level. The controller still uses that measured signal because it is the available information for decision-making.

Feedback architecture therefore includes both the physical process and the information path. In mechatronics, that information path is often just as important as the mechanical parts.

Connecting feedback architecture to Feedback Fundamentals

Feedback architecture is a central idea inside the wider topic of Feedback Fundamentals. The broader topic includes open-loop and closed-loop systems, plus the signals and structure that make feedback possible.

Here is how the ideas connect:

Open-loop systems show what happens when there is no output measurement used for correction.
Closed-loop systems show how feedback is added to improve control.
Feedback architecture explains the internal arrangement of the closed-loop system.
Reference, output, and error signals are the key signals that move through the architecture.

So, if Feedback Fundamentals is the big picture, feedback architecture is the map that shows how the system is organized. It is the part that helps you understand where each signal starts, where it goes, and why it matters.

A useful way to remember this is:

The reference says what you want.
The output says what you got.
The error says how far off you are.
The controller decides what to do next.
The feedback path keeps the system informed.

Conclusion

students, feedback architecture is the structure that makes closed-loop control possible. It describes how the reference, error, controller, plant, sensor, and output work together in a loop 🔁 By comparing the desired value with the measured output, a feedback system can adjust itself and respond to disturbances in the real world. This is why feedback is so important in Control and Mechatronics: it helps machines behave more accurately, safely, and reliably.

Study Notes

Feedback architecture is the arrangement of parts and signals in a control system.
In a closed-loop system, the output is measured and fed back to the input side.
The reference signal $r(t)$ is the desired value.
The output signal $y(t)$ is the actual measured result.
The error signal is often written as $e(t) = r(t) - y(t)$.
If the output matches the reference, then $e(t) = 0$.
Open-loop systems do not use feedback from output to input.
Closed-loop systems use feedback to reduce error and handle disturbances.
A comparator compares the reference with the measured output.
A controller uses the error to decide how to change the plant.
Sensors are important because they provide the measured output for feedback.
Feedback architecture is a core part of Feedback Fundamentals in Control and Mechatronics.
Real examples include thermostats, cruise control, robot arms, and camera stabilizers.