Dr. Xuewu (Daniel) Dai
Research Associate in Communication Group
Department of Engineering Science
University of Oxford
I thought of that while riding my bike -- Albert Einstein, On the Theory of Relativity
I graduated (BEng) in Electronic Engineering at Southwest University, Chongqing, China, in 1999 and received MSc degree in Computer Science from Southwest University in 2003. I completed my PhD research on Observer-based Parameter Estimation and Condition Monitoring for gas turbine engines at School of Electrical and Electronic Engineering, University of Manchester in 2008.
As a volunteer of education aid group under the China Youth Volunteers Poverty Alleviation Project, I taught pupils at Datong, Qinghai in Aug 1999 - Jun 2000. I joined the School of Electronic and Information Engineering, Southwest University as a Lecturer Assistant in 2002, then was awarded the lectureship. From 2009 to 2011, I was a Research Associate at the Communications and Information Systems Group, Department of Electronic and Electrical Engineering in University College London, working for the TSB/EPSRC project "Wireless Data Acquisition in Gas Turbine Engine Testing". I have been a regular reviewer of several leading journals and served as a program member of a few international conferences.
More recently I have been actively involved in research activities in wireless sensor actuator networks for various applications (condition monitoring of aircraft engines and offshore energy systems, optic sensor systems in water industrial and energy consumption monitoring in building automation). In addition to academic research, I have close links with industrial partners and developed research collaborations with various companies, including Rolls-Royce, SELEX Galileo (Communications) and WRc. I was a R&D Engineer in Chuanyi (Chongqing) HiTech Co., Ltd. in 2001, developing CAN field-bus communication system for SCADA.
My research covers a broad range of both academic and industrial interests, centering around the robust adaptive signal processing, state estimation and its applications to wireless communication, time synchronization, wireless networked control systems and condition monitoring of aircraft engines and wind generators, etc.
Optical sensor for organic load monitoring in wastewater treatment and environmental sensing
-- a non-contactable sensor for online monitoring the organic load and suspension solid in the water with the purpose of saving energy in the wastewater treatment control systems.
Adaptive fast-fading OFDM channel estimation and equalization in 3GPP LTE
-- motivated by the application of 3GPP LTE advanced in high-mobility environment (high speed train), we apply the recently developed H∞ robust control and estimation techniques (e.g. dynamic observers) to the channel estimation.
Precision time synchronization for industrial wireless sensor networks
- The time synchronization is treated as the estimation of the global time from non-uniformed sampled clock observation. The precision is paid at the prices of communication bandwidth and computation costs.
Wireless sensor actuator networks and networked control systems
- Wireless data acqusition in gas turbine engine testing, realistic modelling of RF channels in harsh environment, software simulator of industrial wireless sensor actuator networks, WSN hardware testbed development for validating software simulator.
- MAC scheduling for time-sensitive applications.
- Media access delay estimation and delay compensation for wireless sensor actuator networks.
Condition monitoring and fault detection
- I studied dynamic modeling of complex system, model-based fault detection and diagnosis for aircraft gas turbine engines and DFIG wind generators.
Automatic code generation for hardware in loop simulation and embedded system development
- From theory to practice. I used to develop real-time software for CAN field-bus communication systems and TyniOS/Imote2-based WSN testbed, now I am working on Matlab/Simulink/Stateflow code generation for hardware in loop simulation and rapid prototyping on various platforms, Beagleboard, mBed, TI C6000, Arduino and PIC platforms. This lets us concentrate on algorithm development in a graphical fashion (Simulink/Stateflow) and validate the algorithms using the actual data and hardware similar to the final product, which significantly improves the algorithm development and reduces the time from ideas to products.