Dr. Xi Wang
Humboldt Research Fellow at the Chair for Control Systems Engineering
Xi Wang received the B.S. degree in Automation from Liren College, Yanshan University, Qinhuangdao, China, in 2008, and the M.S. and Ph.D degree in Mechanical Electronic Engineering from Xidian University, Xi’an, China, in 2011 and 2016, respectively. He joined Xidian University in 2016. From 2013 to 2015, he was a visiting Ph.D. student at the Systems Control Group, Department of Electrical and Computer Engineering, University of Toronto, Canada. Dr. Wang has been awarded a 24-month Humboldt Research Fellowship to work on his research in FAU, supervised by Prof. Thomas Moor, as a Postdoctoral Researcher during February 2018 and January 2020.
The city of Erlangen is a quiet city, there are many gardens and forests that are very good for me to relax and think. They are both very important for my research.
What is your field of research?
I am interested in supervisory control theory (SCT) and its industrial applications. SCT is a methodology for the modelling and supervisor (controller) synthesis of discrete-event systems (DES) such as manufacturing systems and traffic systems. Informally, a DES is a discrete-state and event-driven system. For instance, a traffic light has three discrete states: red, yellow, and green and the state changes after a permission signal is sent (event-driven). With a DES to be controlled and requirement specifications, SCT automatically synthesizes a supervisor to restrict the behavior of the system in the least restrictive way such that the requirement specifications are always fulfilled.
What were your reasons for choosing FAU as your host institution for a research stay abroad?
One reason is that Professor Thomas Moor at the Chair for Control Systems Engineering is a well-known expert in the supervisory control of discrete-event systems community. He has many significant contributions in the theories of supervisory control of DES, hierarchical supervisory control – which is a layered architecture in which control authority is structured according to scope – , and their industrial applications, which is the reason I choose Chair for Control Systems Engineering at FAU for a research stay abroad. Another reason is that Germany is an industrial country and FAU has a long history, I am so interested to know FAU and understand Germany, that’s why I choose FAU as my host institution for a research stay abroad.
How well known is FAU internationally in your field of research?
Thanks to the DES research group in FAU, led by Professor Thomas Moor, that can provide both software and hardware environments for supervisory control of DES and simulations, FAU is well known very much internationally.
How do you find the interaction between researchers at FAU?
The interaction between researchers at FAU is quite nice.
Could you give us a short summary of the project your research group is working on?
The applicability of SCT to real-world industrial systems is limited due to the two main technical issues: the lack of compact model representation and high computational complexity of controller synthesis. In order to tackle these two issues, state-tree structures (STS) are developed, which can model hierarchical and concurrent organization and synthesize controllers efficiently in the case of large-scale DES. Since 2015, collaborating with W. M. Wonham, who is one of the founders of SCT, I have been developing a novel DES framework “hierarchical timed STS (HTSTS)”, in which the equipped time of events is hierarchical, inspired by time scale “hour, minute, and second” in our daily life. HTSTS provides possibilities of computing the supervisors, i.e. the controllers for large-scale real-world systems. The main line of research in FAU is to continue the initial work on HTSTS towards a self-contained framework.
Furthermore, we plan to work on the supervisor synthesis of HTSTS under partial observation. Arising from that some sensors are located in dangerous environments or their signal collections and passing are difficult or expensive. Supervisor localization of HTSTS is another aspect we investigate. A system is usually composed of several subsystems. Supervisor localization is a distributed control method by which the supervisor for the whole system is decomposed into several local supervisors being assigned to subsystems. In this way, less computer memory is demanded. We also plan the implementation of supervisory control based on STS to real-time systems. It will intertwist the wo research fields’ advantages: The STS framework will be utilized to find the optimal behaviour of a real-time system with the state explosion problem managed significantly, in which an optimal schedule can be found, which satisfies the criteria defined by users.
Currently, our research group has almost finished the development of the hierarchical timed state-tree structures (HTSTS) to bring hierarchical clock structure in to hierarchical discrete-event systems. Meanwhile, we are using DES and STS to model and schedule/reconfigure real-time systems.
My wife Ms. Jun Guo is a great supporter. Without her support, my life and my research will not be success.
And what is your role in the research group?
I provide most of the ideas and frameworks for the research we are working on. And I develop the HTSTS framework and implementing the supervisory control of STS framework to real-time scheduling and reconfiguration by myself.
What are the most important results of your research at FAU to date?
Up to date, our research group has almost finished the development of the hierarchical timed state-tree structures (HTSTS). Currently, our research group is working on the development and supervisory control of novel hierarchical discrete-event system frameworks. By the end of 2018, the development of a software tool for a new approach to control of the STS framework is finished, which provides a new perspective to the supervisory control of STS. Moreover, the development of a new real-time scheduling framework is finished, which can statically and dynamically schedule real-time systems based on the state-tree structure (STS) framework.
How does your research benefit or affect society?
My research on HTSTS will provide a new approach to model hierarchical timed DES. This approach allows users who intend to use HTSTS framework to solve their problem to think locally instead of globally. Therefore, the modelling process is simplified. The modelling principle can manage the state explosion problem in the supervisory control of DES. Generally, a large-scale real-world system could be decomposed into several hierarchical levels and the time scales on different levels are different from each other. By following these general ideas, we are developing a computational toolbox for supervisor synthesis in STS/HTSTS, which will be implemented in industrial cases. Based on these synthesis tools, users can draw their input files by a user-friendly interface according to the logical decomposing depth of the large-scale system.
Our research on the real-time scheduling based on supervisory control theory of STS will provide a two-fold contribution. On the one hand, the developed scheduling framework can find the optimal behavior (all the safe execution sequences) of an RTS by the supervisory control of STS. On the other hand, a few sequences are selected, which rank at the top according to some specified optimality criterion. With the state explosion problem managed and user-friendly interface provided, the newly developed real-time scheduling framework will provide the society a new approach to model and schedule real-time systems. Users can make decisions based on a readable list that is similar to a set of traffic signal.
What were your first (and subsequent) impressions of the Erlangen-Nuremberg region?
The city of Erlangen is a quiet city, there are many gardens and forests that are very good for me to relax and think. They are both very important for my research.
Can you already tell us a highlight, a moment from your stay, which you find particularly memorable?
The Welcome Centre is always good. The kind staff provide great helps every time when I face some troubles. All the activities organized by welcome centre are memorable and lovely.
What are your favourite places at FAU and in Erlangen or Nuremberg?
The Burgberggarten in Erlangen is my favourite place. In Nuremberg, the Germanische Nationalmuseum is my favourite place.
Is there anything else you would like to mention?
My wife Ms. Jun Guo is a great supporter. Without her support, my life and my research will not be success.
Thank you for the interview, Dr. Wang.