Control systems are used to make machines and processes behave in a desired way. They appear in cruise control, room thermostats, robot arms, aircraft autopilots, and factory automation. The main idea is to compare what you want with what is actually happening, then use that difference to make corrections.
This matters because real systems face disturbances, delays, and changing conditions.
A feedback loop connects sensors, controllers, and actuators so the system can continuously adjust itself. In a closed loop, the output is measured and fed back to the controller, which reduces error between the setpoint and the measured value. Engineers often model these systems with block diagrams, transfer functions, and time response graphs.
Good control design balances accuracy, speed, stability, and robustness rather than maximizing only one feature.
Understanding Control Systems: Feedback Loops
A real feedback system has limits at every stage. A sensor does not measure perfectly. A thermometer may be placed near a window, where it feels a draft that does not represent the whole room.
A car speed sensor can give noisy readings when wheels slip. The controller receives this imperfect information, decides on an action, then sends a command to an actuator. The actuator is the part that creates physical change, such as a heater, throttle valve, motor, or pump.
The physical system then responds over time. Heat spreads slowly through a room, while a motor can respond much faster. These delays strongly affect the design.
The error signal gives the controller a direction and a size for its correction. If a room is cooler than the selected temperature, the error tells the heating system to add heat. A simple proportional controller makes a stronger response when the error is larger.
This is useful, but it may leave a small remaining error. For example, a heater may need the room to stay slightly below the chosen temperature before it produces enough heat to balance heat loss through walls. Integral action keeps track of past error.
It gradually increases the command until that lasting difference is removed. Derivative action responds to how quickly the measurement is changing. It can reduce excessive motion when a value is approaching its target quickly.
Fast correction is not always good correction. If a controller reacts too strongly, it can push the output past the target. It then notices the mistake and pushes back too hard.
The result is oscillation. A poorly tuned thermostat can cause a room to alternate between being too warm and too cool. A poorly tuned robot joint can shake around its commanded position.
In severe cases, the motion grows instead of fading, which means the system is unstable. Engineers test how quickly a response begins, how far it goes beyond the target, and how long it takes to become calm. They must account for sensor noise, communication delays, friction, battery changes, and changing loads.
Students can spot feedback ideas in many everyday devices. A refrigerator switches cooling based on temperature. A phone adjusts screen brightness from a light sensor.
A drone uses motion sensors to stay level despite wind. In each case, identify the target value, the measured value, the controller decision, the actuator, and the disturbance. It helps to draw the signal path as a loop rather than memorizing separate parts.
Pay close attention to units and signs. A controller must act in the direction that reduces error.
If a sensor wire is reversed or the sign is chosen incorrectly, a correction becomes a disturbance. That simple mistake can turn a useful system into an unstable one.
Key Facts
- Error signal: e(t) = r(t) - y(t), where r(t) is the setpoint and y(t) is the output.
- Open-loop systems do not use output feedback, while closed-loop systems use feedback to correct error.
- For a proportional controller, control effort is u(t) = Kp e(t).
- For a PID controller, u(t) = Kp e(t) + Ki integral of e(t) dt + Kd de(t)/dt.
- A common closed-loop transfer function is T(s) = G(s) / (1 + G(s)H(s)).
- Key performance measures include rise time, settling time, overshoot, steady-state error, and stability.
Vocabulary
- Setpoint
- The setpoint is the desired target value that the control system tries to achieve.
- Feedback
- Feedback is information about the actual output that is sent back to the controller for comparison.
- Controller
- A controller is the part of the system that decides how to adjust the input based on the error.
- Actuator
- An actuator is a device that converts the controller signal into physical action, such as motion or force.
- Stability
- Stability means the system response stays bounded and eventually settles instead of growing without limit.
Common Mistakes to Avoid
- Confusing open-loop and closed-loop control, which is wrong because only closed-loop systems measure the output and use it to reduce error.
- Assuming more gain always improves performance, which is wrong because excessive gain can cause oscillation, overshoot, or instability.
- Ignoring sensor dynamics or delay, which is wrong because real feedback signals are not instant and can strongly affect stability and response speed.
- Treating zero steady-state error as the only design goal, which is wrong because a system can be accurate in the end but still be too slow or too unstable during the response.
Practice Questions
- 1 A temperature controller has setpoint r = 75 degrees C and measured output y = 68 degrees C. Find the error e = r - y. If Kp = 4, find the proportional control effort u = Kp e.
- 2 A unity-feedback system has forward transfer function G(s) = 8. Use T(s) = G(s) / (1 + G(s)H(s)) with H(s) = 1 to find the closed-loop gain.
- 3 A robot arm reaches its target quickly but oscillates around the final position for several seconds. Explain whether this suggests a stability or tuning problem, and describe one controller change that could reduce the oscillation.