New strategic Associate Professor: Zhonghai Lu
in electronic system design specially dependable autonomous systems.
2000 Zhonghai Lu came to Kista to study at the new pilot master’s program System-on-Chip Design. He stayed on for his PhD 2007, and is now one of five new strategic associate professors. His research area is electronic systems with a special focus on the new field Network-on-Chip, NoC.
Zhonghai Lu has been appointed Associate Professor in Electronic Systems with special focus on dependable autonomous systems. He is very modest about his personal achievements. He prefers to highlight his team and the results that are achieved within the Dependable Autonomous Systems group that he is directing. But when he is asked about his research he gets very talkative.
Embedded Systems
His research covers two main areas: Embedded System (System-on-chip - SoC, Network-on-chip, NoC and multicore architectures) and Wireless Sensor Network (WSN).
“I have my main focus on network-on-chip, NoC,” Zhonghai Lu explains. “This research area started about ten years ago, and KTH in Kista was one of the pioneering groups on NoC. It is an exciting area”
Zhonghai Lu came to Kista 2000 as one of the students in the first (international) master’s program, System-on-Chip Design, at KTH in Kista. The program started off as a pilot, but has established itself as a cornerstone in the electronic systems field. The NoC area caught his interest during his studies.
“I was accepted as a PhD student at KTH in Kistaright after completing my master’s studies.” Zhonghai Lu explains. “My thesis was about Design and Analysis of On-Chip Communication for NoCs.”
In the embedded systems area his research interest covers the communication architecture, performance, analyzing and application mapping in SoCs, NoCs and multicore systems. His aim is to design dependable autonomous electronic systems.
“Until now the NoC research has mainly focused on the network architecture and network interface” Zhonghai Lu says. “The next move is going to be more system-centric and focus on integrating processors, memories, IP blocks etc. to optimize and customize for applications.”
Memory architecture for multicore NoCs is one of the focus areas for future research. Memory access is usually a bottle-neck in electronic systems and also energy-consuming. One solution could be Network-Memory codesign.
“We are currently investigating a comprehensive solution for efficient and flexible support for distributed shared memory (DSM) in multicore architectures,” Zhonghai Lu tells.
Other focus areas for the future will be performance analysis of Multicore 2D and 3D NoC platforms using Network Calculus and flow regulation for optimizationof on-chip communication.
Intel project
Zhonghai Lu is the principal investigator for an Intel project in future communication fabric.
“We are addressing the key fundamental challenges for future chip designs. It is important to stay on top, closely collaboration with leading industrial giants”, Zhonghai Lu comments.
Wireless sensors networks
He started his research into Wireless Sensor Network (WSN) in 2007. WSNs are off-chip networks that need specific and optimized solutions. They use shared radio-links with limited range and introduce interferences. They must be reliable and tolerate the failures of some of the nodes and links.
“High reliability must be achieved without comprising performance at reasonable cost”, Zhonghai Lu stresses. “We need to provide techniques to self-healing and self-organizing mechanisms to deal with various faults. We will conduct research both for on-chip and off-chip networks.”
Team-building
Zhonghai Lu is supervising a few PhD students together with other professors at the department. He is directing the Dependable Autonomous Systems group.
“We are a small, strong and well-connected group”, Zhonghai Lu stresses. “Research needs teamwork, but building a team is a challenge. I am not only looking for the strongest on paper, I want people to contribute to the team.”
Zhonghai Lu is concerned for the future – will it be possible to recruit the best and the right students to the programs. Even though KTH has a good reputation, it will be difficult to get top students. He is working hard to build his research group.
“We need to have our own identity – our own research areas. But we cannot be too individual - it is very important that we contribute the strategic research area electronic systems at KTH ICT", Zhonghai concludes.
