Projects

Henk Corporaal

Below are the currently active projects. Thereafter you'll find an overview of my finished projects.
Projects are executed either at TUD (Delft University of Technology, period 1986-2001), or at TU/e (Eindhoven University of Technology, since 2001).

Current projects

• PARSE: Parallel Architecture Research Eindhoven (started ~2009)
  
• This is the main site of our TU/e Computer Architecture research lab
• On this website you can also find:
• GPGPU symposium slides: Symposium organized by TU/e and Oce-Technologies, September 2010
  

• Convolve: Seamless design of Smart Edge Processors
  
• Horizon 2021 EU project
• Duration 2022-2025
• with 18 partners within the EU countries
  
  
• pAvis: Patient and Environment Aware Adaptive Intelligent Sensor Systems
• Duration 2021-2024
• Partners: Philips, TUE, GrAIMatterLabs, Infineon, Hamburg University, and many others
• Goal: Develop intelligent sensor systems for professional healthcare applications.
  With the aid of embedded AI algorithms and innovative electronics, pAvIs will realize a paradigm shift from “one size fits all” to real-time adaptability such that diagnosis and treatment are optimized for each individual patient towards best achievable performance.
  

• AI in the Sky: EuroStars project
• Duration: 2021 - 2024
• Partners: ViNotion (Eindhoven, project leader), DroneMatrix (Be), AbeonaConsult (Be), TUE-ES
  
• develop an autonomous drone-in-a-box that performs accurate image recognition for traffic measurements, provides real-time insight in unsafe situations, facilitates traffic and crowd safety management and supports emergency services.
  
  
• SmartMobility
  
• Deep learning used for autonomous driving
• Duration: 2020-2023
  
• Joined project with VCA group (Gijs Dubbelman) and NXP
• Bilateral funded by NXP and TUE
  
  
• EDL: Efficient Deep Learning program (2018-2023)
  
• Dutch NWO perspectief program headed by TU/e.
  
• EDL tackles many inefficiencies in deep learning, from faster learning using less data, exploiting temporal redundancy, to efficient implementation of deep learning by adapting Software and Hardware. More than 30 companies cooperate with universities and research institutes.
• The EDL program contains 7 projects, see next table. 
• We are cooperating in most projects, and leading project P7.
  
  

• P#	Name + Project Leader	Short description	Use cases
  P1	DL as a Service (DLaaS) PL: Mark Hoogendoorn (VU)	Tooling to improve the usability and efficiency of deep learning for third-parties; Several demonstrators will show how to exploit DL as a Service	-   Financial data analysis for banks -   Predictive maintenance for electro motors, printers and bank systems -   Airport surveillance -   Specializing SW based on sensory data
  P2	DL for 3D reconstruction, matching and recognition (RMR) PL: Theo Gevers (UvA)	Learning methods for reconstructing 3D computer graphics models directly from sensory data	-   3D Interior reconstruction -   3D cadaster creation and tooling
  P3	DL for Video Analysis and Surveillance (VAS) PL: Gijs Dubbelman (TU/e)	DL for automated video & streaming data analysis	-   Airport surveillance -   Fully automated driving
  P4	DL for High-Tech Systems & Materials (HTSM) PL: Rob van Nieuwpoort (UvA)	DL for large scale fabrication processes and large-scale scientific instruments (steel inspection, surveillance, electron microscopes, telescopes, detectors)	-   Cryo-electron microscopy -   Steel surface inspection -   Radio astronomy image denoising
  P5	DL for Human and Animal Health (HAAH) PL: Henkjan Huisman (Radboud)	DL for automatic monitoring and diagnostics for medical diagnostics	-   Health monitoring of humans -   Health monitoring of groups of cows
  P6	DL for Mobile Robotics (MRob) PL: Dariu Gavrila (TUD)	Learning methods to allow mobile robots to perceive and act intelligent on their environment	-   Autonomous driving -   Automated transit networks
  P7	DL Efficient Deep Learning Platforms (eDLP) PL: Sander Stuijk (TU/e)	HW-SW design for efficient processing of DNNs on low energy embedded systems	-   Small mobile platform for Deep Neural Network applications -   Radar image object recognition HW

   
  

• ZERO NWO-TTW perspectief program (2017-2023)
  Energy Autonomous Systems for IoT
• The ZERO program contains 5 projects:
  

• P1.  Parking sensors, such that cars can find a free parking slot autonomously.

  P2.  Monitoring of traffic patterns using advanced audio beam processing.

  P3.  Autonomous roadside monitoring with video.

  P4.  Ultra-low power transponders for vulnerable traffic users.

  P5.  Dependable autonomous computing platforms supporting mobile traffic users.

• We are cooperating in all above projects, and performing the Program Management.
• We are leading Projects P2 and P3: 
• Project P2: AAS - Autonomous Acoustic Systems
• Autonomous acoustic systems (AAS) can be found in various shapes and sizes, ranging from city wide acoustic monitoring systems to hearing aids worn by an individual. To deliver an improved quality and user experience, future generations of these systems should use adaptive signal processing algorithms, while staying within a stringent energy budget for autonomous operation. The AAS project uses two of these systems as driver cases to develop a novel programming paradigm and accompanying ultra-low power implementation platform for a wide range of autonomous acoustic systems.
• Industrial partners: Resound, Sorama and Altran
  
• Project P3: ARM - Autonomous Roadside Monitoring
• The ARM project will develop a sufficiently energy-efficient embedded processing platform that is suitable for advanced vision algorithms and radar algorithms that can interpret scenes. The scenes need to be interpreted locally to enable the use of low-cost wireless communication channels that have a low-bandwidth (and save energy). Therefore the scientific challenge of the ARM project is the development of programmable hardware and programming methods that allow to provide accurate real-time performance estimates before a detailed implementation has been made. Furthermore interpretation of radar images and combining radar information with vision information for road side monitoring is a new research topic.
  
• With Univ Twente
• Industrial partners: ViNotion, Topic and NXP
  
  

Finished Projects

• BrainWave project (2016 - 2021); see also brain-wave.nl
  
• Wearable device for continuous monitoring and analysis of EEG signals (multi-channel)
  
• For Epilepsy and Parkinson patients in non-hospital environments
  
• Extreme low power processing (CPUs + Accelerators) for EEG signal learning and analysis
• STW funded project
  
  
• MNEMOSENE: Computation-in-Memory architecture based on Resistive devices (2018-2022)
  • The MNEMOSENE project aims at demonstrating a new CIM, computation-in-memory architecture based on resistive devices, together with its required programming flow and interface
  • A H2020 ICT project

• NeMeCo (2016-2021)
  
  • Near Memory Computing, bringing computing much closer to main memory, substantially reducing the data access energy
  • Cooperation between IBM research in Zurich (CH), ASTRON radio astronomy lab in Dwingelo (NL) and TU/e. In 2020 also ETH and Univ Dresden joined.
    
  • Funded by H2020 EU, Marie Sklodowska-Curie EID (European Industrial Doctorate) program.
    
    
• PACSAL (2016-2020)
  
• Joined project with TUE-CS and OCE Copiers and Printers, Venlo
• Among others we heavily perform research on finding near optimal loop schedules using the Halide language.
  
  
• BrainSense: Autonomous EEG headset for Brain Controlled Applications (2018-2020)
• An NWO demonstrator project.
  
• BrainSense develops a wireless headset integrating all the processing, which, for the first time, makes robust brain (EEG based) monitoring and control possible. This new brain-computer interface will be demonstrated for 2 applications, brain-controlled typing and automated learning of a 2^nd language. Its flexibility opens many markets in the area of EEG monitoring and brain controlled applications, like health care market, home automation, assisted learning, gaming and professional virtual reality.
  
• BrainSense is a followup of BrainWave and of oCPS P3 (see below)
  
  
• oCPS: Platform-aware Model-driven Optimization of Cyber-Physical Systems (2016-2020)
  
• Innovative Training Networks, (Marie-Curie EU project)
• Sub-project: Low power imaging / Approximate computing
  
• with Intel & Univ Vienna
  
  
• REXUS project, rocket with our PR3 payloads,
  
• Student project from TU/e (mainly from our ES group) and the Radboud University
• Team, lead by our PhD student Mark Wijtvliet.
• Successful launch on March 11, 2019
• PR3: Payload for  Radiation measurement and Radio-interferometry in Rockets
• Experiment 1: precise rocket localization (10 cm precision)
• Experiment 2: radiation measurement using low cost off-the-shelve cameras (and comparison with scintillation counter)
  
  
• ASSUME project (2015-2019)
  
• Design and Programming of Safety Critical Multicore Systems
• ITEA3 European project
  
  
• rCPS2:  High-performance control for nano-precision systems (2013-2018)
  
• This project is part of the Dutch rCPS (robust Cyber Physical Systems) program.
  
• Industrial Partner: ASML
  
  
• rCPS3: Energy-Efficient Computer-Brain Interaction (2013-2019)
  
• This project is part of the Dutch rCPS (robust Cyber Physical Systems) program.
  In this project Ultra-Low Power Sensor Processing Architectures are developed suitable for Health (EEG) Monitoring
• Project partners: IMEC-NL, RECORE, Twente University, Leiden University

• 
  NEST : Netherlands Streaming project
  In this project the major Dutch expert groups on data flow processing and architectures are researching future streaming systems and corresponding design flows. Key issues are SoC, NoC, predictable and composable design.
  Period: 2009 - 2017
  

• RECA: STARS project on Reconfigurable SIMD accelerator for an MPSoC platform, used for intelligent learning applications.
  Project with RECORE Enschede
  2015
  

• JIIC-RP: STARS project on Just in time compilation for reconfigurable processing
  Project with THALES Hengelo
  2015
  

• ASAM: 2010 - 2015
  
• Automatic Architecture Synthesis and Application Mapping
• Joined project with STM (Italy), UNICA (Italy), DTU (Denmark), Silicon Hive (Eindhoven), TUBS (Germany), ACE (Amsterdam) and Compaan (Netherlands).
  
  
• MNEMEE: Memory management technology for adaptive and efficient design of embedded systems
• Period: 2008-2011
• Partners: IMEC Leuven (coordinator), Democritus Univ (Greece), TU/e, ICD (Informatik Centrum Dortmund), INTRACOM (Greece), and THALES Communicationos (France)
  

• Sensafety: Sensor Networks for Public Safety.
  A COMMIT project 2011-2016
  

• EVA: Embedded Vision Architectures (2009-2014)
  
• EVA is the successor of the SmartCam project.
  
• Supported by "Pieken in de Delta"; joined research with Prof Pieter Jonker (TUE/TUD)
  
• Companies: Assembleon (Veldhoven, penvoerder), OTB (Eindhoven), CHESS (Best), Philips Applied Technologies (Eindhoven),
  
  
• COBRA (2009-2011)
  
• Computing fabric for high performance applications
  
• European Catrene Project, lead by ST-Microelectronics, Grenoble.
  
• Dutch partners: ACE, COMPAAN, ST-Ericsson, TUDelft, TU/e
  
  
• MORPHEUS (2011-2014)
• Motion analysis with contactless camera sensing for professional polysomnography and baby sleep-watching at home
• National R&D project funded by "Point-One"
• Involved: Prodrive, Luvion, Arcuccio, Philips, TU/e and MMC
  
  
• ViewCorrect (2006-2010)
  
• Predictable Co-Design for distributed embedded mechatronic control systems
  
• STW project funding
  
• with Twente University
  
  
• Atalanta (2005-2006)
  
• Autonomous flying systems
• Joined project with DEVlab, Eindhoven, and Prof. Henk Nijmeijer (TU/e)
  
  
• PreMaDoNA project (2004-2009)
• Predictable Matching of Demands on Networked Architectures 
• About predictable design of Networks-on-Chip (NoC) based systems
• Supported by STW-PROGRESS
• Partners: LogicaCMG (Peter de WIth), Philips Research (Marc Duranton, Bart Barenbrug) and Philips Semiconductors (Nijmegen, Rene van den Berg)
  
  
• SmartCam Project (2002 - 2007)
• Multi-Processor Architecture and Application Design Environment for Smart Cameras
• Joined project with TUD (Prof. Jonker).
  
• Involved companies: Philips Natlab, Philips CFT, TNO-FEL, HP-Brisot and IN-3D
  
  
• Soft Reliability Project (2006-2011)
  
• Control soft reliability problems in technical, end-user, and business process terms.
• Lead by Prof. Aarnout Brombacher (dep of Technolgy Management)
• Project embedded into DTI, Design Technology Institute at TU/e and NUS (National University of Singapore)

• Octopus: 2007-2011
  
• ESI (Embedded Systems Institute) project
• Partners: ESI Eindhoven, OCE Venlo, Radboud Univ Nijmegen, TUDelft, Twente University of Technology, and TU/e
  
• Printer Data Path Design: SDF modeling approach
  
  
• SPITS: Strategic Platform for Intelligent Traffic Systems (TU/e, 2009-2011)
  
• High Tech Top project
  
• With NXP, TomTom, Logica CMG, and others
  
  
• BODERC: Beyond the Ordinary: Deisgn of Embedded Real-time Control
• Period: 2003-2007 (TU/e)
  
• ESI (Embedded Systems Institute) project with OCE Venlo
  
  
• AIRPRO (TU/e, 2009)
  
• Advanced Image Processing and Parallel Architectures
  
• SenterNovem (Kenniswerkers) project with OCE Venlo
  
  
• 4MP: SenterNovem Kenniswerkers project (TU/e, 2009-2010)
  
• With ST Ericsson, Eindhoven
• Low power vector processing for 4F mobile phones
  
  
• DxRadio (TU/e, 2009-2010)
  
• Signal processing and digital implementation methodologies for wireless communication
• Joined SenterNovem (Kenniswerkers) project with Catena
  
  
• FAME (2002-2006, TU/e)
  
• Code generation for Embedded Systems
• Joined project with Leiden University (Dr. Peter Knijnenburg)
• NWO fundamental research funding
  
  
• MOVE (1990-2002) and MOVE-ASP (1995-2002) projects (TUD)
  
• TUD spearhead project
  
• MOVE project  home page TUDelft (1990 - 2002)
  TCE site (TTA-based Codesign Environment) of Tampere university
  Research on the semi-automatic design of embedded processors and systems, based on Transport Triggered Architectures
  
• Involved companies: Hewlett-Packard, Oce Venlo, Sagantec Europe B.V., STW, NWO, TNO-FEL, SPaSE Nijmegen, NEC USA, CME
  
• 7 processor chips were realized as a result of this project:
  
• Note, more MOVE/TTA processor were realized later through the TCE (TTA-based Co-Design Environment) project in Tampere
  
  
• 
  
  
• Multiprocessor embedded system project (1997-1999, TUD)
• STW funding (as part of MOVE-ASP project)
• Semi-automatically generate embedded systems based on a single chip multi-processor
  
  
• ADOSS: Automated Design of Simulation Systems (1994-2000, TUD)
  
• Beek funding of TUD
• Cooperations with TUD-FI-CP and TUD-ENS
• Design of simulation systems for Lattice-Bolztmann equations
  
  
• ICOS: Intelligent COde Scheduling (1995-1999, TUD)
• Funded by SION-NWO
• Cooperation with UvA, department of Computer Systems
  
  
• CHESS (1992-1995, TUD)
  
• EU ESPRIT project (# 6498) 
• Parallel VHDL simulation, using optimistic & speculative execution
• With Genua University, Prof Alexandro de Gloria and dr. Paolo Faraboschi (Italy)
  
  
• Communication Processor Design project (1989-1994, TUD)
  
• VLSI framework has been developed for the automatic generation of High Bandwidth, Low Latency Communication Processors (CPs), supporting massive parallel, message passing and systolic processing
• Funded by Sagantec Europe B.V. (Tool support)
  
  
• Lisp/Prolog project (1986-1990, TUD)
• Investigators Henk Corporaal (TUD-EE), Frank .J. Dijkstra (TUD-EE), and Ruud Sommerhalder
  (TUD-TWI)
• Co-operation between: Delft University of Technology (dept. of EE and Computer Science) and IBM.
• Funded by IBM (equipment support with a value of about 0.8 Mf)