This video describes how to make an ideal PID controller more robust when controlling real systems that don’t behave like ideal linear models. Noise is generated by sensors and is present in every system. The derivative in an ideal PID controller amplifies high-frequency noise. Even if that noise is relatively low amplitude, the derivative will sense it and possibly amplify it enough to impact the controller. To protect against high-frequency noise impacting the system, you can modify the derivative path with a low pass filter to reduce the noise before it causes any problems.