1. Pioneers in Engineering, UC Berkeley Pioneers in Engineering Week 8: Sensors and Feedback

2. Pioneers in Engineering, UC BerkeleyControl Announcements 2

3. Pioneers in Engineering, UC BerkeleyControl Lesson Agenda  Definition of systems  Overview of controls  PID control 3

4. Pioneers in Engineering, UC BerkeleyControl Lesson Agenda  Definition of systems  Overview of controls  PID Control 4

5. Pioneers in Engineering, UC BerkeleyControl Systems 5 INPUT OUTPUT  A system is a function. SYSTEM

6. Pioneers in Engineering, UC BerkeleyControl Examples of Systems - 1 6  Coffee Machine INPUT? OUTPUT?

7. Pioneers in Engineering, UC BerkeleyControl 7  F1 Car Example of Systems INPUT? OUTPUT?

8. Pioneers in Engineering, UC BerkeleyControl Lesson Agenda  Definition of systems  Overview of controls  PID Control 8

9. Pioneers in Engineering, UC BerkeleyControl What is controls theory?  The study of how to choose the input to a system to give the desired output.  The entity which generates the input is called a controller  Vex 9

10. Pioneers in Engineering, UC BerkeleyControl Why is controls hard? 10  Usually we don’t know what the system is exactly  Model Mismatch–Model not the same as real system  Disturbances–gravity, wind, etc. Laboratory Model Reality

11. Pioneers in Engineering, UC BerkeleyControl Solution: Feedback Control 11 SYSTEM SENSOR CONTROLLER r y u  Measure output and send to controller  Controller usually acts on the error e= r-y’  Common algorithm is PID control y’

12. Pioneers in Engineering, UC BerkeleyControl Videos  Inverted Pendulum Inverted Pendulum  Line Following Car Line Following Car  Quadrotors Quadrotors 12

13. Pioneers in Engineering, UC BerkeleyControl Lesson Agenda  Definition of systems  Overview of controls  PID Control 13

14. Pioneers in Engineering, UC BerkeleyControl The PID Controller  Probably the most commonly used controller architecture.  Sufficient for basic linear dynamical systems (basic robot locomotion, arm placement, etc.) 14

15. Pioneers in Engineering, UC BerkeleyControl The PID architecture 15 SYSTEM SENSOR  The PID controller has three basic parts: The proportional (P), integral (I), and derivative (D) blocks. I D P r u e=r-y’ y’

16. Pioneers in Engineering, UC BerkeleyControl The PID architecture 16 I D P u  Corrects the current error. Largest input of the three  Improves stability and shapes response  Hammers out residual error. Rejects disturbances

17. Pioneers in Engineering, UC BerkeleyControl PID Tuning – Manual Method 17  Set K i and K d to 0.  Increase K p until system starts to oscillate, then reduce by about half  Increase K i until system steady-state (residual) error is acceptable  Increase K d to shape response

18. Pioneers in Engineering, UC BerkeleyControl PID Demos 18  Matlab demo  Robot arm demo

19. Pioneers in Engineering, UC BerkeleyControl PID Activity 19  Download code from wiki page  Change the gains using the tuning technique taught until the arm behaves well  Add a way for the user to control the arm position (suggestion in code)