CSE hosts Early Career Workshop 2019

Aiming to provide opportunities for early career researchers in automation and control to establish international research linkages and promote their research, a two-day workshop was held during Nov 28-29, 2019 at Zhejiang University, Hangzhou, China. College of Control and Science Engineering (CSE) of Zhejiang University, which is one of the leading “A+” institute in control and automation area in China, hosted this event.  

This is the first workshop since Automation and Control Global Research Network is established in August 2019, which is a new collaboration between five of the leading research institutions in the field of automation and control, including: The University of Newcastle (UoN), Australia; Zhejiang University (ZJU), China; KTH Royal Institute of Technology (KTH), Sweden; Nanyang Technolog ical University (NTU), Singapore; and The University of California, Santa Barbara (UCSB), USA. The objective of the global research network is to promote joint research activities among these universities through joint program, academic publication, conference and workshop and so on.

This workshop is one of the activities fulfilling the above objective, which contains several conference style presentations for the first day, followed by some networking events and laboratory visits in the next day. The workshop starts with an opening remark addressed by Prof. CHEN Jiming who is one of the general chairs of this workshop. Speaking on the occasion, Prof. Chen said “This will be an excellent opportunity for the early career researchers to establish international research collaboration as well as to interact with local students here.” He is also very grateful for the great support from Prof. Andrew Fleming from UoN and all the other steering members, such as Prof. Lihua Xie from NTU, Prof. Andrew Teel from UCSB and Prof. Hakan Hjalmarsson from KTH. Following the opening remark, there are four sessions of talks throughout the first day as detailed below.

In the first session, Dr. ZHAO Chengcheng gave a presentation on “Data Privacy-Preserving Consensus-Based Distributed Energy Management (DEM) in Smart Grids”, where the vulnerability of consensus-based DEM algorithm due to data disclosure is discussed and the strategies based on secret function (i.e., adding some vanishing noise), is used to enhance the security level of the strategy. 

A final year PhD student, Yipeng Pang from NTU, then spoke on “Gradient-Free Distributed Optimization”, where he talked about how to solve the distributed optimization problem when only the local measurement of the cost functions is available. He also mentioned that they have implemented their algorithms in HVAC systems for verification purpose. 

After this, Dr. Sindri Magnússon, from KTH, further presented a talk on “On the Convergence of Distributed Optimization under Limited Communication”, where he showed that a constant number of bits is sufficient for a general distributed algorithm to maintain the linear convergence rate. He also discussed about the specific tradeoff between the communication cost and the convergence performance for a class of distributed algorithms.

In the second session, a PhD student, Nathan Bartlett from UoN, delivered a talk on “Extended Target Tracking with Random Matrices”, where he proposed a random matrix model and a corresponding closed-form solution to estimate the kinematics of a target such as position, velocity, and acceleration, as well as the shape, size, and orientation of the target body. He also talked about the potential extension of this method to multi-target tracking problems arising in mail delivery robotic systems. 

Then, another PhD student, Kun Cao from NTU, presented a work on “Relative Docking and Formation Control via Range and Odometry Measurements”, where he proposed a scheme to not only localize but navigate a robot to a desired docking position relative to a fixed landmark, whose position is unknown, based on merely range and odometry measurements. Besides, he also discussed about the potential application to multi-robot systems for formation control at both spatial and temporal dimensions. 

Dr. Dario Paccagnan from UCSB, then gave a presentation on “Control of Socio-technical Systems: Traffic Routing, Congestion Pricing, and Beyond”, where he proposed an effective and robust way to solve a class of congestion games so as to make efficient use of the existing technological infrastructure, such as transportation system and power networks, based on proper toll design leveraging only local information. He also showed that the optimal toll designed based on local information is surprisingly comparable with that based on full information.

In the third session, Dr. WANG Mengzhi from ZJU, delivered a talk on “Stabilization of Perturbed Continuous-time Systems Using Event-triggered Model Predictive Control”, where he introduced two event-triggered control schemes for model predictive control of (EMPC) of continuous-time nonlinear systems with bounded disturbances. He showed that event-triggering conditions can be properly designed in order to balance the computation and communication cost and the inter-event time should be bounded below to avoid the zeno behavior. He also provided the conditions of recursive feasibility for the proposed triggering strategies in terms of the prediction horizon, the triggering level and the disturbance bound. 

Dr. Keith Paarporn from UCSB, then spoke on the topic of “Risk-Security Tradeoffs in Networked Systems”, where he discussed the mitigation strategies for a distributed networked system against adversarial attacks. He also characterizes the level of vulnerability of the system to the attacks and the operator's fundamental trade-off between security against worst-case broad attacks and vulnerability from focused attacks. Besides, he highlighted the design challenges a system operator may face in maintaining resilience of networked distributed systems.

After this, Prof. Uday Tupakul from UoN, delivered a brief lecture on “SDN-based Security for Critical Information Infrastructures”, where he first gave an overview of cyber security research at Advanced Cyber Security Engineering Research Centre (ACSRC), The University of Newcastle, Australia, and presented some opportunities at ACSRC for further collaboration. He then briefly discussed his current research work which makes use of Software Defined Networking for securing Critical Information Infrastructures, aiming to enable secure and dependable computing information systems and services.

In the final session, Dr. Michael Ruppert from UoN, presented a talk on “Design and Control of Microelectromechanical Systems”, where he discussed new advances in integrating actuation and sensing at chip level onto the micro-cantilever, a key component in a variety of instruments for manipulation and interrogation at nanoscale (e.g., atomic force microscopes), so as to improve the dynamics and control performance of the micro-cantilever especially when interacting with nonlinear forces arising at the micro- and nano-scale. He also demonstrated the ability to control the quality factor in order to manipulate the cantilever tracking bandwidth and tip-sample dissipation. 

Dr. Mohamed Rasheed Hilmy Abdalmoaty from KTH, then spoke on the topic of “System Identification of Stochastic Nonlinear Models Using Estimating Functions”, where he introduced a prediction error method based on suboptimal predictors for system identification of stochastic nonlinear models. He said the proposed method is relatively simple compared to estimators based on sequential Monte Carlo smoothing algorithms as it does not compute the likelihood function. He also described a more general identification framework to systematically construct optimal estimators based on the estimating functions approach. 

Finally, Dr. Riccardo Sven Risuleo from KTH, delivered a talk on “Bayesian approaches in learning dynamical systems”, where he introduced a universal modeling tool based on Gaussian-process model for complex structured systems that can be modeled as static nonlinear blocks and linear dynamical blocks. He also showcased the approach on a blind problem of occupancy estimation from CO2 data with tractable approximate inference leveraging Markov Chain Monte Carlo approach and the expectation-maximization method.

After a fruitful presentation day, the next day mainly covers some networking events and laboratory visits. Prof. CHENG Peng, the other chair of this workshop, first delivered a lecture introducing Zhejiang University, College of Control and Science Engineering as well as the Networked Sensing and Control Group led by Prof. CHEN Jiming. Then, it is followed by a bunch of laboratory visits prescheduled to allow attendees have a concrete view of the research here in a form of demonstration and report, which include unmanned systems lab, industrial control security lab as well as the network sensing lab.

In summary, the workshop was basically designed to bring the early career researchers from automation and control area on a single platform and share their research experiences and also career plans, promoting research collaboration. The workshop attracts around 80 researchers (including faculties and students) for participation from the control and automation area. The workshop was organized jointly by Zhejiang University, Faculty of Information Technology, College of Control Science and Engineering, State Key Laboratory of Industrial Control Technology, and Networked Sensing and Control Group. It was, again, an excellent opportunity for all attendees to establish international research linkages, promote their research as well as seek potential collaboration and, of course, enjoy one of the most beautiful city in China.

Finally, the organizers would like to thank all the attendees for making this workshop successful and wish all of them a promising career future!