Engineering safety
Chau recalls his time as a PhD student at Imperial College as a period of discovery, challenges, and intellectual development. The mindset required at PhD level study was entirely different to that required as an undergraduate or master’s student, commanding time management and efficiency to benefit from the autonomy it offered. It was also a time during which he learnt how to think and problem solve by recognising and finding solutions to problems. It was also a challenging time. Until Chau learnt to think differently; challenging his sources to find loopholes. These loopholes would then unlock a new area of research.
However, Chau was able to utilise this freedom of independent study to explore and develop intellectually, talking to and learning from people with brilliant ideas. These interactions often led to unexpected collaborations. During one point in his research, Chau happened upon a postdoctoral researcher working on medical robotics. They found that their research had something in common and that there would be value in a collaborative project. Chau was able to help the robotics team improve the calculation speeds of the human-robot collaborative control in medial surgery, a major issue that the team had been struggling to apply to clinical setting due to complex models. This project - born out of a perchance conversation - turned out to be a chapter of Chau’s thesis: “Human-robot collaborative control for surgery”.
Another project Chau remembers particularly fondly was a cross-discipline collaboration with Cambridge University, who had a research group in Control Theory. They collaborated to look into air traffic control problem which is concerned with the routing, safety and scheduling of aircraft in airspace and on ground. Air traffic control systems are notoriously inefficient as they rely on too many human interventions. Busy hub airports in the world, such as Heathrow Airport, often suffer as victims of their own success. An excess of air traffic results in aircrafts being forced to circle.
Before Chau joined the research project, the Cambridge group had developed a control system to automate the air traffic control process to be more efficient, but the traditional software solution was far too slow. The algorithm was only able to make calculations every ten minutes, when it needed to do the same in ten seconds. Using Chau’s expertise in reconfigurable technology, substantive speed up has been achieved to allow the air traffic control problem to be solved in real time for dense vehicle scenarios. It also inspired a number of final year students that Chau had involved in the project to pursue PhDs in the department.
Applying Research in Real World Scenarios
Both Chau’s research on custom computing during his PhD and his ability to question and push boundaries, have lent themselves naturally to his role as a systems solution engineer at Altera. Whereas during his research years at Imperial, much of Chau’s applications were more theoretical in nature, because Altera is more customer-driven, the applications are more commercially applicable.
Chau’s work at Altera is specifically related to control and functional safety in automotive and industrial applications. He uses Field Programmable Gate Array (FPGA), a device produced by Altera that does reconfigurable computing. Control is crucial in factory, process automation and energy infrastructure where there are many machines interacting with one another and therefore reliable and precise control is key. It is also important in automobiles - with the requirements constantly changing, as technologies evolve. For example, Advanced Driver Assistance Systems (ADAS) include functions such as collision avoidance, adaptive cruise control and vehicle-to-vehicle network, which enhance driving experience and safety. And again – semi-autonomous and driverless vehicles need robust system to meet the escalating performance requirements while adapting quickly and cost effectively to evolving market and standards. Another factor Chau must consider is functional safety - that the device reduces the risk of failure and adheres to the latest safety regulations.
Chau has recently collaborated with a number of automotive companies manufacturing hybrid electric/battery electric vehicles. Electric vehicle environment is very challenging on batteries, thus one part of the work would be to provide a Battery Management System (BMS) that is able to precisely estimate battery state of charge, available power, health condition, and lifetime. Chau’s job is to provide a high performance and robust BMS systems using the power of FPGA and his knowledge in reconfigurable computing.
Having studied Computer Science and Engineering at the Chinese University of Hong Kong (2004-2010), Chau pursued a PhD (2014) at Imperial College London, under the supervision of professors Wayne Luk and Peter Cheung, as a Croucher scholar. His Thesis, Optimising Reconfigurable Systems for Real-time Applications - and the majority of his research - revolved around using reconfigurable technology to address evolving computation challenges.
To view Thomas’ personal Croucher profile, please click here.