Embedded Systems for Automation & Embedded Software

The automation applications of interest include sensor/actuator control of domestic and industrial infrastructures like, security systems, smart buildings, vehicle automations, production line control and medical applications. The controlled sensors/actuators may be simple analog components like temperature, pressure, humidity, infrared sensors, microphones, leds, switches etc, or may have a more complicated structure with an intelligent digital interface like cameras, LCD displays, DTMF generators/checkers etc. Sensors having an analog interface are connected to the processing unit through Analog-to-Digital Converters while analog actuators may be driven by the processor through Digital-to-Analog Converters.

One or more processing units are used in such systems with appropriate embedded peripherals as well as reconfigurable hardware for the implementation of special functions with strict time constraints. The processing units can communicate with each other through dedicated interface, shared buses, shared memory etc. The application software can be executed directly on a microcontroller in simple systems or under the control a real time operating system if it consists of multiple threads/processes that have to be scheduled and assigned to various processors. The application software can be described in assembly or in a higher level language like C and is developed using simulator or emulation environments based on development/evaluation boards. In real time operation the software is either stored in a non-volatile memory like Flash or EEPROM or is downloaded from an external source through an interface like USB, Ethernet, Serial Port and is temporarily stored in RAM.

Biomedical Systems and Applications

The main objective of this direction is the design and development of ICT platforms facilitating the in-depth study, monitoring and research on demanding medical applications, exploiting dominant technologies from the fields of sensors, wireless sensor networks, embedded systems, IoT and data processing.

In the context of the specific research activities, we study novel communication architectures, develop enhanced, robust and flexible back-end infrastructures as well as integrating a wide range of heterogeneous sensor aiming towards holistic, detailed and objective monitoring biomedical processes and phenomena. Additional, emphasis is paid in the development of novel approaches of data processing of the acquired raw data in order to develop multiscale, modeling of biomedical processes through which new methods can be envisioned towards treatment as well as forecast of specific conditions.

Architectures for Visual Perception

Special Architectures for Visual Perception. The need for machines that can see and recognize what they see is becoming very important as robotics come out in the real market. Traditional hardware architectures for machine vision are too complex and too expensive as they involve arrays of dedicated DSP cores. We are in need of new architectural templates that can provide the necessary processing power with reasonable need for resources. And we also need special OS and software structures to work with these new architectures.

Embedded Multicore/Multithreaded Architectures

The focus of this team is to deal with the design challenges that arise when multiple processors or/and highly multithreaded processors are present in a single chip. More specifically, the research performed aims at exploring the current trends towards heterogeneous architectures in relation to the still growing memory wall and power wall problem.

Embedded Hardware Design

Embedded System Hardware design team is specialized in hardware implementation for embedded telecommunication systems. For the design of such systems, system level design languages like SysML, ImpulseC and SystemC are usually used, while for the implementation languages like VHDL and VERILOG are employed.

The hardware implementation is mainly based on the use of FPGAs.