A control system automatically regulates a process to maintain a desired output. It consists of four elements: a set point (the desired value), a sensor (measuring the actual value), a controller (computing what action to take based on the error), and an actuator (executing the action). The simplest controller is on-off control (like a thermostat), but more sophisticated proportional control adjusts the actuator output in proportion to the error -- a large error produces a large correction, while a small error produces a small correction. Control systems are the foundation of automation, enabling everything from cruise control in cars to temperature regulation in buildings to precise manufacturing.
Build a proportional controller for a simple system: a fan speed controller that adjusts based on temperature. Start with on-off control (fan fully on or fully off) and observe the oscillation. Then implement proportional control (fan speed proportional to the temperature error) and observe the smoother response. Compare the two approaches on a graph of temperature vs. time. Discuss why a cruise control system that only applied full throttle or no throttle would result in jerky, uncomfortable driving.
You have already learned about sensors that measure the physical world and feedback loops that compare measurements to desired values. Control systems formalize these ideas into a complete engineering discipline: the science of making systems regulate themselves automatically.
The basic control loop has four parts. The set point is what you want -- 72 degrees in a room, 60 mph on a highway, 1,000 RPM on a motor. The sensor measures the actual value. The controller calculates the difference (error = set point - actual) and decides what to do. The actuator carries out the controller's decision -- opening a valve, adjusting a motor speed, or activating a heater. The loop runs continuously: measure, compare, adjust, measure, compare, adjust.
The simplest controller is on-off control (also called bang-bang control). If the temperature is below the set point, the heater is fully on. If above, it is fully off. This works but produces constant cycling: the heater overshoots the target because it is at full power right up until the temperature crosses the set point, then the room cools below the target before the heater kicks on again. A thermostat clicking on and off repeatedly is on-off control in action.
Proportional control is a significant improvement. Instead of full-on or full-off, the controller adjusts the actuator output in proportion to the error. If the room is 10 degrees below the set point, the heater runs at high power. If it is only 1 degree below, the heater runs at low power. As the temperature approaches the target, the heater gradually reduces its output, resulting in a smooth approach rather than a jarring overshoot. The proportional constant (called the gain) determines how aggressively the controller responds.
However, proportional control has a subtle limitation: it often leaves a small residual error. If the heater needs to run at 30% power to maintain the set point temperature, the proportional controller can only produce 30% output if there is some non-zero error. This means the steady-state temperature settles slightly below the set point. Advanced courses cover how integral and derivative terms (forming the famous PID controller) address this and other issues. For now, the key insight is that control systems transform passive systems into active, self-regulating ones -- one of the most powerful ideas in all of engineering.