Integration and Implementation Platforms
Integration & Implementation Platforms
Integration and implementation platforms include several setups and equipment that support, facilitate and implement all the research activities of the center.
These include embedded systems, mobile communications and networking technologies, IoT platforms and IT infrastructure.
Embedded Systems Technologies
KIOS provides technologies and solutions focused on the design, development, implementation and verification of low-power, high-performance and highly reliable systems on chip, suitable for embedded and mobile environments. Moreover, these novel technologies include computer aided design, testing and reliability of modern VLSI circuits and embedded systems, with expertise in CAD algorithms for automatic testing, diagnosis, verification, and resilient design, applicable to different hardware platforms (VLSI, SoCs, NoCs, on-chip multiprocessors).
Activities include hardware acceleration of domain-specific computing for real-time processing, with a broad range of applications such as computational intelligence and machine learning, computer vision, robotics and multimedia applications, bioinformatics and wearable biomedics and intelligent data processing.
Key equipment includes:
- Various visual sensors (cameras, LiDAR, time-of-flight)
- NVIDIA Jetson TK1 and TX2 embedded GPU platforms for real-time edge computing
- Intel Neural Compute Stick 2 and Google Coral Edge TPU platforms able to perform real-time object detection using Deep Neural Networks on the Edge
- Embedded Computing Systems (Raspberry Pi, Gumstix Boards, Odroid Platforms)
- Smart Camera Nodes based on the Raspberry Pi Computers
- Hybrid Multi-FPGA/Multi-GPU Platform for large scale circuit and system emulation
- State-of-the-art low-end and high-end FPGA boards for acceleration of compute intensive applications (computer vision, image processing, machine learning)
- Multi-GPU supercomputers used for machine learning and computer vision applications
- Multi-SOC FPGA-Based BeeCube Platform for large-scale hardware emulation
- Bumblebee Stereoscopic Camera for stereo vision applications
- High-end servers/workstations and logic analysers
- State-of the-art CAD tools (Synopsys, Cadence, and Mentor Graphic) for development and simulation purposes
Mobile Communications & Networking Technologies
Mobile Communications & Networking Technologies examine the modelling, simulation, emulation and design of architectures, protocols, algorithms and technologies for next-generation communication systems (e.g. machine-to-machine, 5G systems, relaying, cognitive radio, optical/telecommunication networks), with a focus on communication theory, wireless communications and networking.
Key equipment includes:
- NI USRP: Software Defined Radio Platform
- Energy Harvesting Development Kit for Wireless Sensors
- Commercial/Custom-made simulation software for optical/telecommunication networks
- IP routers/switches and network testing equipment
Smart Buildings IoT
The Smart Buildings IoT platform supports the research and development of solutions related to monitoring and control of building systems, focusing on indoor environmental monitoring.
All available sensors communicate using the Z-WAVE protocol. The data are collected through a gateway and analyzed by the Domognostics platform and API. The sensors are installed within the KIOS Laboratory and are continuously monitoring the environmental conditions of the building.
The Smart Buildings IoT infrastructure is currently equipped with:
- Aeotec Multisensor (temperature, humidity, light, UV, motion, vibration)
- Aeotec Door/Window Sensors
- Samsung SmartThings
- CO2 Sensors
- Weather Sensors
IT Infrastructure
The main IT infrastructure of KIOS CoE includes servers which are used to support, facilitate an implement all the research activities of the center.
The several types of physical or virtual servers are used to support the flow, storage, processing and analysis of research data. Additional servers or groups of personal computers connected in local networks are used for simulations covering demands such as MATLAB models, 3D Modeling, CFD calculations and machine learning models including deep learning.
Some of the servers include:
- LINUX SIMULATION SERVERS for simulations on different kind of programming that requires LINUX OS
- KIOS Windows server for simulations
- KIOS GPU server having the latest technology GPU to be shared among researchers for deep learning
