Professor Glen Tian

Professor, School of Computer Science

PhD (University of Sydney), PhD (Zhejiang University)

Research theme: Computer Science

Research areas: 

• Big data computing, cloud computing, distributed computing, data centre management and optimisation, bioinformatics computation
• Computer networks and communications, smart grid communications, mobile networks, sensor networks, Internet of Things, networked systems
• Dynamic systems, fractal and multi-fractal analysis,  bioinformatic
• Artificial intelligence, machine learning, heuristics and meta-heuristics
• Control systems, networked control, smart grid control

Potential PhD/Masters/Honours Projects

1. Distributed and/or Parallel Processing of Big Data. Big data are data with large volume, fast and dynamic generation, and diversity of data formats. Their management, storage, retrival and processing are challenging due to these features. Hadoop and map/reduce are being widely used for big data management and processing, but they are not suitable for many real-world applications such as some bioinformatics problems we are investigating at the moment. Also, the efficient utilisation of the resources of each of the distributed machines is still a challenging task for big data processing. These call for emerging development of new front-end programming models and back-end technology support for big data processing. We have established a big data lab for experimental HPC and networks at QUT, and the lab provides state-of-the-art facilities to carry out the research in this area.
2. Communications and Big Data Processing in Smart Grid. Smart grid is a new concept to design and operate power generation, transmission, distribution and other related systems in an integrated environment with emerging services. There are a number of challenges in smart grid systems. One of the challeges is smart grid communications for real-time applications. The existing standards for smart grid communications are based on IP networks, which were not designed for real-time control systems and thus do not provide real-time performance guarantee. New network protocols need to be developed to complement existing smart grid network technologes to enable real-time communications for real-time QoS performance. Another challenge in smart grid systems is the management, storage, retrival, and processing of big data generated through a huge number of sensors in high frequencies for system monitoring, fault diagnosis, control and operation.This requires development of new technologies in this area. We have collaborations with QUT’s power engineering discipline, and have also established a big data lab for experimental HPC and networks at QUT, facillitating excellent research in this area.
3. Data Centre Management and Optimisation. Data centres have been increasingly built to provide a wide range of data, network, and other cloud services. The increasing demand in various services requires more and more data center resources and consequently more and more energy is consumed in data centres. It is estimated that about 30% of the running cost of a data centre is the energy consumption. Therefore, it is significant to manage and optimise data centre resources and energy consumption in data centres. New techniques are to be developed to tackle the problems of virtual machine management, task placement, task migration, and QoS maintainance for effiient management of data centre resources and energy comsumption. An understanding of operations research and data centre management is essential to carry out the research.
4. Distributed/Cooperative Computing for Reducing Computational Complexity in Bioinformatics. Bioinformatics computing demands huge computing and memory resources, and becomes one of the main problems in computational bioinformatics. Investigating traditional sequential computing, parallel computing, cluster/distributed/cooperative computing and/or other computing approaches, the projects to be undertaken on this topic aim to develop architecture/frameworks, algorithms and scheduling strategies, dynamic memory management policies, implementation methods and/or web tools for complicated bioinformatics problems. They would make the best use of the computing resources including memory and CPUs for improved computation efficiency and/or more complicated bioinformatics computing. In particular, the projects will consider the scenarios of building Composition Vector (CV) Trees for phylogenetic analysis using complete genenomes without sequence alignment. Preliminary studies have been carried out in our group and working c++ code is already available for stand-alone sequential computation of CV tree building.
5. Cooperative Computing of Ad-Hoc Wireless Networks and GPS for Seamless Vehicle Navigation Services. GPS (Global Position System) is the most widely used navigation technology for automobile systems. The GPS applications can be used in traffic navigation, emergency assistance, collision avoidance and vehicle tracking in automobiles. Seamless navigation means continuously navigating users over the applicable areas of those sensors and maps. Therefore, seamless navigation in GPS will provide reliable navigation, reduce the fuel consumption of the vehicles, minimize dangerous driving, and reduce the possibility of collision. However, current GPS devices can provide location information, route guidance and location-sensitive services only when there is a direct line of sight to four or more satellites. When a vehicle passes through a tunnel or urban areas, the tunnel or high skyscrapers block the signals from the satellites. As a result, the GPS device in the vehicle could not communicate with the satellites effectively, and continuous navigation of the vehicle becomes impractical. This motivates the research on seamless global navigation technologies. The projects to be carried out on this topic aim to develop leading edge technologies for seamless navigation services through corporative computing of ad-hoc wireless networks with GPS. They will focus on network architecture, routing protocols, cross-layer optimization, computing frameworks, and computing algorithms.
6. Positioning Services and Clock Synchronization through Wireless Data Networks. Wireless is now a pervasive aspect of everyday life, and has become ubiquitous. This makes it possible to develop emerging and location-aware services that cannot be provided by existing network systems. Potential applications of location-aware services include emergency services, aged care and assisted living, to just mention a few. Although the Global Navigation Satellite Systems (GNSS) are an ideal positioning technology for open sky operation, they have limited availability in urban settings, especially in indoor applications, due to the blockage of signals. Preliminary studies have shown several possible solutions using wireless networks for positioning services. One is to use access points; however, noise characteristics will result in inaccuracy in location determination. An alternative and distributed solution is to establish relative clocks of the mobile nodes; however, accurate relative clocks and their synchronization in wireless networks are still challenging. Addressing both challenges, the projects to be carried out on this topic aim to develop effective technologies for accurate location determination in wireless networks. Specific objectives include: 1) Comprehensive requirements analysis for wireless location determination; 2) Development of fast and precise relative clock synchronization protocols, 3) Development of innovative methods, algorithms, and protocols for wireless location determination; and/or 4) Evaluation and demonstrations of the developed technologies.
7. Cross-Layer Design and Optimization for Wireless Network Applications. Wireless networked systems are being increasingly investigated for various applications due to their good scalability, fast deployment, and low implementation and maintenance costs. However, compared with wired systems, wireless ones introduce longer transmission delays, bigger jitter and higher rates of packet loss. These problems could result in unpredictable system behaviors which may not be acceptable in many services and applications. Our projects on this topic aim to improve overall performance of wireless networked applications through avoiding data packet dropouts, reducing network-introduced communication latency and jitter, and better scheduling wireless network resources. The methodology of the cross-layer design and optimization will be used in this project to develop new technologies and network protocols. Cross-layer design is an efficient and popular method for designing protocols especially for wireless networks; however, it will actively violate conventional layered architecture for better performance.
8. Integrated Modelling and Design for Wireless Networked Control Systems. There has been strong interest in wireless networked control, e.g., networked mobile vehicles, uninhabited aerial vehicles. Existing and emerging applications of wireless networked control are wide-spread including irrigation networks, water resource monitoring, disaster monitoring and emergency services, and aged care and assisted living. However, the full potential of wireless networked control is hindered by the lack of theory and technology support to address current challenges. The projects to be carried out on this topic will try to overcome these hurdles by creating new knowledge in the emerging area. They aim to develop innovative integrated design theory and methodologies for wireless networked control systems (WNCSs) for significant improvement in real-time performance. They are significant because if successful as we anticipate they will fundamentally change our philosophy to design networked control systems for emerging as well as existing applications, provide a total solution in a uniform framework to three significant challenges, and provide theoretical support to WNCS implementations. Expected outcomes include a uniform WNCS modelling framework under various scenarios, new methodologies for stability analysis and controller design, new network protocols, new techniques for feedback scheduling, and/or frameworks for integrated design.
9. Coordination and Networked Operation of Large Groups of Intelligent Agents. With the rapid development of the technologies in multi-robot and multi-agent systems, multi-robot and multi-agent systems have more and more applications. Many practical applications can benefit from the use of multi-agent systems, such as highway traffic control, battlefield environment, emergency services, and underwater and space exploration. In these challenging application domains, multi-agent systems can often deal with tasks which are difficult to accomplish by an individual robot/agent. While some topics in this area have been well addressed over the years, such as mission assignment, path planning, and formation generation and keeping, other significant issues are still challenging. Moreover, implementation of the theory in real world environment is also very limited. Our projects on this topic aim to optimize system structure, behaviour function and cooperation mechanism for large scale multi-robot/agent systems. Optimization strategies and algorithms will be established to enhance the overall performance of the large scale multi-robot/agent systems.
10. Machine learning for fault diagnosis from multiple sources of big data.

For further information, please refer to my homepage: users.tpg.com.au/yctian

Additional information