2013 ACC ADRC Workshop - Program PDF Print E-mail
Written by Administrator   
Wednesday, 13 February 2013 06:06










Different from regular academic lectures and conference presentations, this workshop is designed to show, instead of tell, the principles and practice of ADRC. It is question driven with a single goal: to maximize participants’ learning experience by observing the long and fascinating history of the idea, by witnessing how the idea is engaged today in solving problems in various domains of engineering applications, and by directly interacting with over ten speakers, covering topics from basic principles, methods, implementations, to current and future research and developments. In each of the three sessions, the speakers will lead, but not dominate, the proceedings by giving short, concise presentations on the latest developments and addressing a key question, followed by a forum where questions from the participants, either submitted on-line or in person, are asked and answered. Practical problems and examples are used throughout the discussions to help participants connect novel principles to their own experience. Matlab simulations are shared with the participants to show simplicity and effectiveness of ADRC. Benchmark problems from various domains of applications are solved on the spot to show breakthroughs.


  • Session I: Principles, Methods and Impacts ( Saturday, June 15th, 2013, 8:00 am to 12:00 am )
  • Session II: Technologies and Applications ( Saturday, June 15th, 2013, 1:00pm to 6:00pm )
  • Session III: Theoretical Foundation ( Sunday, June 16th, 2013, 8:00 am to 12:00pm )
  • Session IV: Hands on Session for Students ( Sunday, June 16th, 2013, 1:00 pm to 3:00pm )

Session I: Principles, Methods and Impacts

  • Saturday, June 15th, 2013, 8:00 am to 12:00 am
  • Speakers: Zhiqiang Gao, Yi Huang, Enrico Canuto, Lei Guo and Hebert Sira Ramirez

Description:There are three parts to this session: 1) a critical reflection on the foundation of automatic control and the exposition of the history and principles of a different paradigm, ADRC; 2) the methods of estimating internal and external disturbances, with or without a model of the physical process; 3) the impacts of ADRC in asking a different set of questions in research and in creating a disruptive technology in industrial control.

In his landmark paper of 1989, at the beginning of ADRC development, Prof. J. Han pointed out an irreconcilable difference between the process uncertainties a controller must deal with and the model based control theory that has been dominant for several decades at that time. To honor his spirit of inquiry, this workshop begins with a critical examination of the foundation of automatic control. Taking his inquiry further, we’ll show that at the foot of this great body of knowledge called control theory there is much confusion, that the concept of feedback and feedback control need to be distinguished, as suggested in vein by Dr. Bode in 1960, and that to equate the science of control with that of feedback, an unexamined belief, is narrow-minded. Through an exposition of history of ideas in automatic control, we let a picture of this discipline unfold that is rich and colorful. Recovering of these great ideas, we believe, will animate another round of furious growth in the science and engineering of automatic control. In particular, we trace the idea of ADRC to the conception by Jean Victor Poncelet in 1829, through its germination in the then Soviet Union from 1930s to 60s and rebirth in China in 1990s, and finally, its industrialization in the U.S. financed in part by a Silicon Valley venture capital firm, U.S. Venture Partners. As the workshop unfolds, we will demonstrate a new toolbox of principles, methodologies and analysis.

We will summarily introduce various methods of disturbance estimation in different forms and under different names from the engineering practice and academic literature, spanning over four decades, including unknown input observer (UIO), disturbance observer (DOB), extended state observer (ESO), and many others. We’ll explore the effectiveness and limitations of these techniques. We will also show that by actively rejecting disturbances, the traditional boundary is broken down in terms of dynamics systems: linear vs. nonlinear, time-varying vs. time invariant, etc. and this will have a huge impact in how we see the problems of control, in research and in practice.

  • 8:00-8:15 Introduction and Opening Remarks
  • 8:15-8:30 On the Centrality of Disturbance Rejection, Zhiqiang Gao (Please click here to download the slides)
  • 8:40-9:00 Active Disturbance Rejection Control: From Concept to Practice, Zhiqiang Gao (Please click here to download the slides and MATLAB simulation models)
  • 9:10 to 9:30 On Methodology of Active Disturbance Rejection Control, Yi Huang and Wenchao Xue (Please click here to download the slides)
  • 9:30-10:30 Coffee Break: Coffee break activities: interactive and computer simulation demonstrations
  • 10:30-10:50 Active Disturbance Rejection and the Control of Differentially Flat Systems, Herbertt Sira Ramirez (Please click here to download the slides)
  • 11:00-11:20 Disturbance Estimation and Rejection: the Embedded Model Control Approach, Enrico Canuto (Please click here to download the slides and MATLAB simulation files)
  • 11:30-11:50 Refined Anti-Disturbance Control for System with Multiple Disturbances, Lei Guo (Please click here to download the slides)
  • 12:00-1:00 Lunch Break: Lunch is provided. Lunch time discussion: What’s new in Active Disturbance Rejection Control?
Questions from Participants

1. What is ADRC? Where does it apply and what is the scope of its applications? What are its limitations?

2. How is ADRC related and compared to other disturbance rejection methods from the last four decades?

3. Why do you say the problem of industrial control is that of disturbance rejection?

4. Why do we have so much trouble answering the simple question of "What is Control?"

5. Why are we so single-minded about feedback control? How do we explain the vast use of feed-forward?

6. Why is it that after so many decades of modern control theory, PID still dominates in industry?

7. What if I already have a good plant model? Can we combine it with ADRC for the better?

8. As more information about the plant comes in, can it be used to make ADRC perform better?

9. Can the concept of stability be redefined in the ADRC framework, to reflect a new conception of control?

10. The goal seems to be similar between ADRC and SMC, can the two be compared or combined?

11. Where do we start to learn ADRC? Is there a dedicated website?

Suggested Readings for Session I

Session II: Technologies and Applications

  • Saturday, June 15th, 2013, 1:00pm to 6:00pm
  • Speakers: Zhiqiang Gao, Hebert Sira Ramirez, Enrico Canuto, Lei Guo, Hui Xie, Wen Tan, Shen Zhao, Rafal Madonski, Qing Zheng

Description: ADRC as originally proposed by J. Han has three components: tracking differentiator, nonlinear feedback control, and nonlinear extended state observer. The combination of the three proves to be a powerful tool for disturbance rejection control. As an industrial control technology, however, it has been streamlined, simplified and parameterized so that it can be easily deployed across various hardware-software platforms and easily tuned by factory personnel, leading to explosive growth in its applications that transcend engineering disciplines. In this session we show how this is accomplished and how an advanced principle is married into a domain of engineering applications that often leads to an order of magnitude improvement. Most importantly, it will be shown that the key in application of ADRC is the reformulation of the control problem as that of disturbance rejection, as shown in applications in several key technology areas such as internal combustion, power generation, space applications, aeronautics, process control, and high energy physics.

  • 1:00-1:20 ADRC as a Viable Industrial Control Technology: a Summary of Recent Developments, Zhiqiang Gao (Please click here to download the slides)
  • 1:30-1:50 Active Disturbance Rejection in Internal Combustion Engines: Problems and Solutions, Hui Xie
  • 2:00-2:20 Linear Active Disturbance Rejection Controllers (LADRC) for Boiler-Turbine Units, Wen Tan (Please click here to download the slides)
  • 2:30-2:50 Hardware Validations of the Concept of ADRC, Hebert Sira Ramirez ( Please click here to download the slides)
  • 3:00-3:30 Coffee Break
  • 3:30-3:50 Disturbance Rejection in Space Applications: Problems and Solutions, Enrico Canuto (Please click here to download the slides)
  • 4:00-4:20 Active Disturbance Rejection in a Lower-limb Rehabilitation Manipulator, Rafal Madonski (Please click here to download the slides)
  • 4:30-4:50 Refined Anti-Disturbance Control in Aeronautical Engineering: Problems and Solutions, Lei Guo
  • 5:00-5:20 On The Unity of Decoupling Control and Active Disturbance Rejection, Qing Zheng
  • 5:30-5:50 Application of Active Disturbance Rejection Control in Modern Linear Particle Accelerators, Shen Zhao (Please click here to download the slides)
  • 6:00pm Adjourn for the day

Questions from Participants

1. What is technology? How is it different from “an application”?

2. Can each speaker explain the thought process that leads to the successful application?

3. What is the signature of ADRC that is essentially distinct from the existing technology?

4. What is the downside of ADRC?

5. How is ADRC implemented in PC, PLC, DSP, FPGA, etc.?

6. When will the TI DSP chip with ADRC algorithm hardwired in it be available on the market?

7. Can Nonlinear ADRC be parameterized as in the linear case?

8. How do we use ADRC for systems where the number of inputs and outputs are not the same?

9. Can the parameter b0 in ADRC be estimated in real-time, as opposed to be given?

10. How is ADRC applied to solve problems in high energy physics? Is it scalable to other kind of scientific experimentations?

11. How is ADRC affected by uncertainties in sampling, noise, time-delay, etc.? How can it be improved?


Suggested Readings for Session II

Session III: Theoretical Foundation

  • Sunday, June 16th, 2013, 8:00 am to 12:00 pm
  • Speakers: Bao-zhuGuo, WenchaoXue, Qing Zheng, ZhiqiangGao, Enrico Canuto.

Description: This session is designed to answer critical questions about the theoretical foundation of ADRC and stability. It will be shown that rigorous mathematical proof has been recently established in all three parts of Han’s theory: tracking differentiator, nonlinear ESO and nonlinear feedback. For practitioners familiar with the language of frequency response, the bandwidth of the observer and of the control loop is analytically related to the tracking error. Frequency response of a typical ADRC loop is shown to have unsurpassed quality in robustness of performance and stability. The fascinating question is where do we go from here to develop another kind of control theory: the theory of disturbance rejection? Initial principles and methods are examined. The structural outlook of a reconstructed control theory is presented in the context of active disturbance rejection.

  • 8:00-8:20 On the Mathematical Proofs of Convergence of Active Disturbance Rejection Control, Bao-Zhu Guo (Please click here to download the slides)
  • 8:30-8:50 Frequency Domain Analysis of ADRC, Zhiqiang Gao (Please click here to download the slides)
  • 9:00-9:20 Theoretical Analysis of ADRC, Wenchao Xue, Yi Huang (Please click here to download the slides)
  • 9:30-10:00 Coffee Break
  • 10:00-10:20 Time Domain Analysis of Linear ADRC: Relationship Between Bandwidth and Tracking Error, Qing Zheng (Please click here to download the slides)
  • 10:30-10:50 Embedded Model Control calls for disturbance modeling and rejection, Enrico Canuto ( Please click here to download the slides)
  • 11:00- noon Panel Discussion – A theory of disturbance rejection: past, present and future.
  • 12:00-1:00 Lunch Break: Lunch is provided

Questions from Participants

1. Exactly what is proven in ADRC?

2. What are the pros and cons between the time domain and frequency domain analysis of ADRC?

3. What can be proven in nonlinear ADRC as opposed to linear ADRC?

4. Beyond ADRC, what are the questions and methods towards a theory of disturbance rejection?

5. What are the main issues/challenges facing ADRC?

6. How do we systematically explore the advantage of using nonlinear gains in ADRC?

7. How to reduce the conservatism in the stability analysis?

8. Can ADRC be applied to a non-minimum phase system? How?

9. Can ADRC be used in nonlinear filtering? How?

Suggested Readings for Session III

Session IVHands on Session for Students

  • Sunday, June 16th, 2013, 1:00 pm to 3:00pm

Description: In this newly added session, students, assisted by some speakers and their assistants, will tackle the control problems of their own, using the simulation modules provided. At the end of session, students will present their research problems and preliminary findings, followed by discussions.


Last Updated on Monday, 02 December 2013 14:17
Joomla 1.5 Templates by JoomlaShine.com