ESE406/505-MEAM513: Lecture 1 Introduction to Feedback and Control Ali Jadbabaie January 11, 2005 Goals: Give an overview of the course; describe course structure, administration Define feedback/control systems and learn how to recognize main features Describe what control systems do and the primary principles of control Reading (available on course web page): Astrom and Murray, Analysis and Design of Feedback Systems, Ch 1 “For the Spy in the Sky, New Eyes”, NY Times, June 2002. Course Administration Announcements : •First class is on Tuesday January 13th 2004 in Towne 313 from 12:00-1:30pm. Course Description: This course is an introduction to analysis and design of feedback control systems, including classical control theory in the time and frequency domain. Modeling of physical, biological and information systems using linear and nonlinear differential equations. Stability and performance of interconnected systems, including use of block diagrams, Bode plots, Nyquist criterion, and Design of feedback controllers. Suggested pre-requisites: Basic course on ordinary differential equations and linear algebra. For Systems Engineering Students: knowledge of ESE 210 (SYS 200) material. For EE students: Knowledge of signals and systems (ESE 325) Instructor: •Ali Jadbabaie , jadbabai@seas.upenn.edu , Office hours : Wednesdays 2:00-4:00pm, 365 GRW Moore bldg. Lectures: T- TR 12:00-1:30pm, Towne 313. Textbook: •Feedback Control of Dynamic Systems, by Franklin, Powell and Emami Naieni, 4th Edition, Prentice Hall, 2002. Other References: •Modern Control Engineering, 4th Edition, by K. Ogata, Prentice Hall, 2001 •Modern Control Systems, 9th Edition, by Dorf and Bishop, Prentice Hall, 2001. •Automatic Control Systems, by B. Kuo, Prentice Hall, 1995. Course Notes and Links Reading material for the class will be posted on blackboard Required reading sources •R. M. Murray (ed), Control in an Information Rich World: Report of the Panel on Future Directions in Control, Dynamics, and Systems, SIAM, 2002. Available online at http://www.cds.caltech.edu/~murray/cdspanel/ •K. J. Åström and Richard M. Murray, Analysis and Design of Feedback Systems, Preprint, 2004. Online access on blackboard •J. Doyle, B. Francis, and A. Tannenbaum, Feedback Control Theory, McMillan, 1992. Online access on blackboard •Grading : Homeworks : 20% Midterm I: 35% Midterm II : 45% •Teaching assistants: Nima Moshtagh , Ali Ahmadzadeh January 11, 2005 1 Controls Course Sequence Fall ESE500 – Linear Systems Theory Detailed description of state space concepts. Rigorous analysis and synthesis of time invariant and time varying systems. Spring ESE406/505-MEAM513 – Introduction to the principles and tools of control and feedback Summarize key concepts, w/ examples of fundamental principles at work Introduce MATLAB-based tools for modeling, simulation, and analysis Introduction to control design Provide knowledge to work with control engineers in a team setting ESE 617/MEAM 613- Nonlinear Systems • Tools and algorithms for analysis and design of nonlinear control systems January 11, 2005 2 What is Feedback? Miriam Webster: the return to the input of a part of the output of a machine, system, or process (as for producing changes in an electronic circuit that improve performance or in an automatic control device that provide self-corrective action) [1920] Feedback = mutual interconnection of two (or more) systems System 1 affects system 2 System 2 affects system 1 Cause and effect is tricky; systems are mutually dependent Feedback is ubiquitous in natural and engineered systems January 11, 2005 System 1 System 2 Terminology System 1 System 1 System 2 System 2 Closed Loop Open Loop 3 What do these two have in common? Boeing 777 Tornado • Highly nonlinear, complicated dynamics! • Both are capable of transporting goods and people over long distances BUT • One is controlled, and the other is not. • Control is “the hidden technology that you meet every day” • It heavily relies on the notion of “feedback” January 11, 2005 4 Example #1: Flyball Governor “Flyball” Governor (1788) Regulate speed of steam engine Reduce effects of variations in load (disturbance rejection) Major advance of industrial revolution Balls fly out as speed increases, Valve closes, slowing engine Steam engine Boulton-Watt steam engine January 11, 2005 Flyball governor http://www.heeg.de/~roland/SteamEngine.html 5 Other Examples of Feedback Biological Systems Physiological regulation (homeostasis) Bio-molecular regulatory networks Environmental Systems Microbial ecosystems Global carbon cycle Financial Systems Markets and exchanges Supply and service chains January 11, 2005 ESE 6 Control = Sensing + Computation + Actuation In Feedback “Loop” Actuate Sense Gas Pedal Vehicle Speed Compute Control “Law” Goals Stability: system maintains desired operating point (hold steady speed) Performance: system responds rapidly to changes (accelerate to 65 mph) Robustness: system tolerates perturbations in dynamics (mass, drag, etc) January 11, 2005 7 A modern Feedback Control System January 11, 2005 8 Two Main Principles of Control Robustness to Uncertainty through Feedback Feedback allows high performance in the presence of uncertainty Example: repeatable performance of amplifiers with 5X component variation Key idea: accurate sensing to compare actual to desired, correction through computation and actuation Design of Dynamics through Feedback Feedback allows the dynamics of a system to be modified Example: stability augmentation for highly agile, unstable aircraft Key idea: interconnection gives closed loop that modifies natural behavior X-29 experimental aircraft January 11, 2005 9 Example #2: Cruise Control disturbance reference mv bv uengine uhill uengine k (vdes v ) velocity vdes vss k 1 vdes uhill bk bk 1 as k 0 as k time January 11, 2005 + - Control + System Stability/performance Steady state velocity approaches desired velocity as k Smooth response; no overshoot or oscillations Disturbance rejection Effect of disturbances (hills) approaches zero as k Robustness Results don’t depend on the specific values of b, m, or k for k sufficiently large 10 Example #3: Insect Flight SENSING neural superposition eyes hind wing gyroscopes (halteres) specialized “power” muscles two wings (di-ptera) ACTUATION COMPUTATION ~500,000 neurons January 11, 2005 More information: M. D. Dickinson, Solving the mystery of insect flight, Scientific American, June 2001. 11 EXAMPLE # 4: Coordinated Control of Manned and Unmanned Systems January 11, 2005 12 Other Examples Temperature control Air bags EGR control Active suspension Electronic fuel injection Electronic ignition Electric power steering (PAS) January 11, 2005 Anti-lock brakes Electronic transmission Cruise control 13 Steering Brakes Anti-skid Cruise control Traction control Shifting Electronic ignition Wipers Mirrors GPS Temperature control Electronic fuel injection Seatbelts Bumpers Fenders Suspension (control) Airbags January 11, 2005 Radio Headlights Seats 14 Gene networks? essential: nonessential: unknown: total: January 11, 2005 230 2373 1804 4407 http://www.shigen.nig.ac.jp/ecoli/pec 15 essential: nonessential: 230 2373 Are these “redundant?” No! January 11, 2005 16 Cartoon of E. Coli metabolism January 11, 2005 Regulatory feedback 17 Regulatory feedback January 11, 2005 18 Actuation Decision January 11, 2005 Sensing Signaling 19 Organized complexity Simple behavior Robust and adaptive Evolvable Enormous “hidden” complexity January 11, 2005 20 Segway: The human Transporter January 11, 2005 21 January 11, 2005 22 Modern Engineering Applications of Control Flight Control Systems Modern commercial and military aircraft are “fly by wire” Autoland systems, unmanned aerial vehicles (UAVs) are already in place Robotics High accuracy positioning for flexible manufacturing Remote environments: space, sea, non-invasive surgery, etc. Chemical Process Control Regulation of flow rates, temperature, concentrations, etc. Long time scales, but only crude models of process Communications and Networks Amplifiers and repeaters Congestion control of the Internet Power management for wireless communications Automotive Engine control, transmission control, cruise control, climate control, etc Luxury sedans: 12 control devices in 1976, 42 in 1988, 67 in 1991 AND MANY MORE... January 11, 2005 23 The Internet: Largest feedback system built by man Applications Web FTP Mail News Video Audio ping napster Transport protocols TCP SCTP UDP ICMP IP Ethernet 802.11 Power lines ATM Optical Satellite Bluetooth Link technologies January 11, 2005 24 The Internet hourglass Applications Web FTP Mail News Video Audio ping napster TCP IP Ethernet 802.11 Power lines ATM Optical Satellite Bluetooth Link technologies January 11, 2005 25 The Internet hourglass Applications Web FTP Mail News Video Audio IP under everything ping napster TCP IP Ethernet 802.11 IP on Power lines ATM Optical everything Satellite Bluetooth Link technologies January 11, 2005 26 Network protocols. Files HTTP Files TCP IP packets packets packets packets packets packets Links Sources January 11, 2005 27 Protocol stack Applications TCP IP Hardware January 11, 2005 Modules Files packets packets packets packets packets TCP packets packets packets packets packets IPpackets packets packets packets packets packets Layerpackets 2 packets packets packets packets packets packets Bits 28 Animation of the protocols Files HTTP Files TCP packets packets packets packets packets TCP packets January 11, 2005 29 Animation of the protocols Files HTTP Files TCP packets packets packets packets packets TCP packets IP packets packets packets packets packets TCP packets packets packets packets packets packets packets January 11, 2005 30 Animation of the protocols Files HTTP Files TCP packets packets packets packets packets TCP packets packets packets packets packets IPpackets packets IP packets packets packets packets packets TCP packets packets packets packets packets IPpackets packets packets packets packets packets packets Layer 2 packets Links Sources January 11, 2005 packets packets packets packets packets Bits 31 TCP IP January 11, 2005 Vertical decomposition Protocol Stack Application IP Application Application Each layer can evolve TCP TCP independently provided: 1. Follow the rules 2. Everyone else does IP IP with IP “good enough” their layer Routing Provisioning 32 Application Application TCP TCP IP Application IP IP TCP IP IP Horizontal decomposition Each level is decentralized and asynchronous January 11, 2005 Routing Provisioning 33 Vertical decomposition • Entirely different from the telephone system, although the parts are essentially identical (VLSI, Application Application Application copper, and fiber) • The Internet is much more like biology and TCPrelies on feedbackTCP TCP regulation at every level. • Only recently has a coherent theory of the Internet started to emerge and pay off. IP IP IP IP IP January 11, 2005 Routing Horizontal decomposition Provisioning 34 Internet Interface Application Application TCP TCP Operating System Simplify IP IP Computer Board Device January 11, 2005 Link 35 Links Sources January 11, 2005 36 Routers packets January 11, 2005 Hosts 37 Files Routers Hidden from the user packets January 11, 2005 Hosts 38 Routers packets January 11, 2005 Hosts 39 Routers packets January 11, 2005 Hosts 40 Control Tools Modeling Input/output representations for subsystems + interconnection rules System identification theory and algorithms Theory and algorithms for reduced order modeling + model reduction Analysis Stability of feedback systems, including robustness “margins” Performance of input/output systems (disturbance rejection, robustness) Synthesis Constructive tools for design of feedback systems Constructive tools for signal processing and estimation (Kalman filters) January 11, 2005 MATLAB Toolboxes SIMULINK Control System Neural Network Data Acquisition Optimization Fuzzy Logic Robust Control Instrument Control Signal Processing LMI Control Statistics Model Predictive Control System Identification µ-Analysis and Synthesis 41 Magic of Feedback • Feedback is used to regulate the value of a quantity in a system to a desired level, by measuring the error, i.e., difference between the desired value and the sensed value. •Sometimes the decision is based on the instantaneous value of error, and sometimes is based on the history of the error, and/or predictions on the future value of the error. Some times we use all three. •The performance of a feedback system is measured based on the response to a “step” change in the reference, or in tracking a sinusoid. • Feedback regulation will work even when the “components” are uncertain. • The down side of using feedback is that It can cause instability It makes the design more complicated • The main components of a feedback loop are sensing, decision/computation, and actuation. • We will use theory of differential equations, linear algebra and complex variables to analyze feedback systems. January 11, 2005 42 Overview of the Course Wk 1 2-3 Tue/Thur Introduction to Feedback and Control System Modeling/Analysis, Review of ODEs, and Laplace Transform January 11, 2005 4-5 Stability and Performance 6-7 Tests for stability 8-9 Root locus analysis. Design for time domain specs. 1011 Frequency Domain Design: Bode plot. 1214 Loop Analysis of Feedback Systems. Nyquist criterion 15 Fundamental Limits on Performance 16 Uncertainty Analysis and Robustness 43 Summary: Introduction to Feedback and Control Actuate Sense Control = Sensing + Computation + Actuation Feedback Principles Robustness to Uncertainty Design of Dynamics Compute Many examples of feedback and control in natural & engineered systems: BIO ESE BIO ESE CS January 11, 2005 44 Summary Feedback control is Every where you just have to look for it January 11, 2005 45 Welcome to ESE406/505- MEAM513 Control Systems Instructor: Ali Jadbabaie jadbabai@seas.upenn.edu Course website: on Blackboard January 11, 2005 46