Research directions:

This research theme deals with the issues related to the architecture design and implementation of embedded system's software and hardware. In addition, the aspects of CAD design, simulation, formal verification and test of such systems are of interest. The scope of the research ranges from very tiny computing systems up to networked high performance processors.

VYZAS main objectives

Objective 1: Embedded systems are becoming an unavoidable part of our life. This raises a big concern about their security, reliability and trustworthiness especially when privacy or human health is involved. Our objective is to research on techniques and approaches that can be widely applicable for increasing the embedded systems dependability and security.

Objective 2: Embedded systems are expected to "disappear" in the human environment. This trend is envisioned to constrain (the already limited) power budget by an additional order of magnitude. We aim at minimalistic, extremely low-power, networked architectures, implementations and design techniques of small computing nodes.

Objective 3: Reconfigurable computing "invades" the embedded systems domain. Wider application of reconfigurable hardware is difficult with the current technology constraints, e.g. reconfiguration speed and power. We investigate how to cope with the current technology limitations and actively look at emerging technologies that may be applied to build tiny reconfigurable computers.

Objective 4: The VLSI technology reliability is diminishing with the future size reductions. The contribution of soft errors in system failures will grow in a very near future. This imposes tough challenges not only for the VLSI testing process but also for the design trajectory. We investigate on testing (detection and repair) algorithms for off and online system monitoring.

Objective 5: Embedded systems often operate under real time constraints (of either hard or soft type). High degree of autonomy is expected since the software may or may not be accessible by the user. In addition to the severe size constraints, the Embedded software may have to deal with critical decisions. We mostly focus on back end compilation, software size minimization to fit hardware limitations via instruction set re-engineering, and fault tolerance using hardware assistance.

Objective 6: Recent embedded systems are characterized with their multidisciplinary nature and growing versatility. New design paradigms such as Systems on a Chip (SoC), Multiprocessor SoCs (MPSoC), distributed wireless sensor networks, systems on a package and Networks on a Chip (NoC) are inherent to that trend and need special attention. We address various applications and design assets of such heterogeneous systems not only from the processor point of view but also including sensing, data acquisition and extremely low power wireless communication systems. To address the heterogeneous nature of the above we have created alliances and partnerships within the TU Delft  Dimes and several electrical engineering research laboratories including: Electronic Instrumentation, Circuits and Systems and Electronics.

VYZAS projects

• SiMS (Smart implantable Medical Systems)
• MaLT (Memory and Logic Testing)
• RECouNT (Reconfigurable Embedded Computing and unemployed Novel Technologies)
• DepART (Dependable Architectures, Reliability and Trustworthiness)
• MinERAL (Minimalistic Embedded Reconfigurable Architectures for Low-power)
• MISP (Minmal Instruction Set Processor)
• PINE (Packet INspection Engine)