Edwin

Learning new technologies: • Ansible • Deep Learning • TensorFlow/Keras • Vue.js • Zephyr

NXP’s chips possesses advanced security features. In this project I focused on improving the ease of using them from a software development point of view. My responsibilities in this project are: • writing code to assess the ease of use of security features • developing a prototype tool for project setup • developing a prototype VSCode extension for project setup and analysis. The technologies that I used in this project are: C/C++, Git, CMake, Rust, Python, VSCode, Typescript, Javascript, NXP tools, Windows, Linux.

The Allseas Jacket Lift System is complex high-end lifting system for off-shore lifting of jackets. This system is in demand of advanced control software to enable an operator to perform lifting operations. In this project I contributed to the overall control system (OCS). My responsibilities in this project are: • adding and completing features of the OCS • developing unit tests and testing on the HIL setup • refactoring the communication layer to support multicast The technologies that I used in this project are: C++, Conan, Git, CMake, Google Test, gMock, Azure DevOps, QT, Windows, Linux, QNX.

High Dynamic Range (HDR) video involves the processing of video data streams. Algorithms and their efficiency play an important role in such video pipelines. In this project my focus is on the (re)design and implementation of such algorithms to benefit from the computational capabilities of modern GPUs, both for PC and Mobile. My responsibilities in this project are: • refactoring existing code • algorithm development and optimization • implementation • verification and validation The technologies that I used in this project are: C++, Jupyter, PythPython, Java, OpenGL, fragment shaders, compute shaders, Git, CMake, Android.

The Physics Based Camera is an important sensor in the PreScan simulation system. This sensor is different from the outset in that the simulated images are high dynamic range and result from multi-spectral light interactions. This is required to accurately simulate the current state of the art automotive cameras. My responsibilities in this project are: • improving and extending the simulator (model specification, design, and implementation) • verification and validation of the product • full re-implementation of the optics and sensor simulation • writing documentation • setting up Camera Validation methodology • contributing to both simple and advanced Lidar sensor models The technologies that I used in this project are: C++, Matlab, Jupyter, Python, OSG, OpenGL, fragment shaders, compute shaders, C# .NET, Mercurial (Hg), CMake, Doxygen, Jira, and Jenkins.

Responsible for the specification, design, and implementation of an alarming system of a multi-board embedded system that is used for loading and blending liquids in tanks. This system is responsible for always giving an actual and consistent view of the error status of the product, even when the communication system fails. The technologies that I used in this project are: C, C++, C# .NET, Make, Subversion, Doxygen, and IAR Workbench.

Responsible for high quality sensor modeling and simulation in the PreScan simulation system. The focus of this project was to accelerate the post-processing of an abstract Radar sensor by using OpenCL. The technologies that I used in this project are: C++, OpenCL, OpenGL, OSG.

To load tank cars with fuel, dedicated loading and blending systems are required. These systems consist of a diversity of hardware board, interconnected by a CAN bus. Because it is required that each board must be able to initiate a transaction and because the standard CAN-bus level of abstraction is not appropriate for the type of products build, a dedicated protocol stack is built on top of the CAN protocol. My main responsibility in this project is the development of PC tooling to support application development and debugging, and further analyzing, implementing, and integrating the new protocol stack for the embedded platform, and designing a test framework for the CAN communication. The technologies that I use in this project are: C, C++, C\# .NET, Make, Subversion, Doxygen, and IAR Workbench.

Cars get increasingly more intelligent. Intelligent Vehicle Systems or Advanced Driver Assistance Systems are examples of that. In such systems, cars are equipped with sensors to observe the environment. Based on the sensor data cars can respond with, for example, automatic breaking or notifying the driver to take some action. PreScan is a software package that allows these systems to be rapidly evaluated with a high degree of flexible and at low-cost. My main responsibility in this project is the development and maintenance of the sensors in the PreScan simulation system. The sensor suite ranges from abstract bounding box based sensor to sensors that model RADAR, LIDAR, and Ultrasonic systems. The technologies that I use in this project are: MatLab, Simulink, Real-Time Workshop, C, C++, Java, JNI, Java-profiling Kit, and Subversion.

Radio finger printing is a successful method for room-level accurate localization. However, this method suffers from the calibration effort that is required to construct a database of measurement-location pairs. This research focusses on an affordable multi-modal localization system that can be used for finger printing calibration. The focus is to combine vision based techniques with inertial sensors and to use the process dynamics of a mobile camera to allow for sparse marker deployment and Kalman filter-like approaches. My role is to perform research on this system. This includes overall system architecture and algorithms. I also guide students and cooperate with them to set up an experimental system. Implementation will be done in Matlab and/or C(++)/OpenCV. My responsibilities in this project are: • leading project • performing research, architecting, and design experimental system • guiding graduation students • built prototype implementation • writing scientific publications

The research takes place with the DevLab members but also in larger consortia settings. ALwEN (Ambient Living with Embedded Networks) is such a project. The goal of ALwEN is to further maturize the underlying MyriaNed wireless sensor network technology. It does so by selecting an application driver from the elderly care domain. In ALwEN I fulfil two roles: work package leader and researcher. The work package I lead focusses on methods and tools for wireless sensor network creation and the creation of a reference architecture. My activities are: • organizing and chairing meetings; • guiding the research of Ph.D. students active in the project; • developing the reference architecture; • formulating graduation assignments.

As a researcher, I am mainly active in defining research directions and to translate them in graduation assignments together with guiding (graduation) students. I also built a prototype simulator that allows the rapid execution of embedded WSN code. Apart from the already mentioned tasks I fulfil in the DevLab a number of additional tasks: • member (shared chair) of the technical committy: here we discuss technical issues and translate them into concrete proposals for the DevLab board; • chair of the DevLab Technical Forum: this is the meeting where DevLab students meet experienced technical employees from the participating DevLab companies to get input for their project; • coordinator DevLab: coordinate the dayly activities at the DevLab.

The DevLab is a cooperation between twelve technological SMEs. The aim is to gain knowledge to create business on the long turn. The DevLab has its own accommodation where Ph.D. students and graduation students (M.Sc and B.Sc.) can perform their projects. The DevLab research activities focus on three themes, viz.: Dutch Clay, MyriaNed, and Atalanta. My primary role within the DevLab is to coach the graduation students. Next to that I am increasingly busy with gaining knowledge myself, and to spread this knowledge within the DevLab and our own company. The result of my contribution to the DevLab is the guiding/coaching of six graduated students and the establishment of a Wireless Sensor Network special interest group to transfer DevLab knowledge to Alten PTS.

In this home automation project eight products are developed to be taken in mass production. For four of the eight products Vitelec has product responsibility, for the other four products Vitelec produces an important component that is composed of hardware and software. A home automation system is set up by combining several of these products. They will organize themselves into a wireless Zigbee network to accommodate wireless data transport between the components. The goal of the home automation system is to increase the living comfort. In the role of project leader my responsibilities were: • formulation project planning and proposal of project budget (fixed price); • align with and report to external client; • align with and report to two external product owners (together they have the responsibility for four products); • align with external mechanical designers; • align and direct purchase and production support; • align with electronic service providers (ESP); • leading (technical) of development team consisting of three software (design, implementation, test) and three hardware developers (schema, layout, test).

PTS wants an intelligent game for her employees. In this game, small animals (PeTS) compete autonomous in a virtual world. Each employee "programs" his PeT from a flash-enabled web browser. The PeTS themselves compete on a customer's server. My responsibilities in this project were: • development of game engine; • development of game logic. An object-oriented Python implementation has been made on a Linux PC platform. The engine connects to a MySQL data base. Eclipse and CVS are used.

To safeguard her advising role for the future Alten PTS is continuously busy with keeping her knowledge up-to-date. This project contributes to Java knowledge development. The customer wants an evaluation of the PERC Ultra VM (virtual machine). This VM provides soft real-time program execution. In this project, my responsibilities were: • setting up the evaluation environment; • formulating the evaluation criteria; • composing the benchmark set; • reporting the evaluation results. As evaluation platform, a Linux PC has been selected. The PERC VM is benchmarked against the standard Sun JVM (J2SE 1.5) and against gcj native compiled Java. The evaluation criteria are response times and execution times of the programs in the benchmark set. I selected the SciMark2 benchmark set and the Java Grande Forum Benchmark Set. From this last set we selected a subset of programs that best matches the application domain of Alten PTS.

In this period, I worked on several projects, varying from business applications to embedded systems. In particular, the latter is worth mentioning. Our project is developing a Terrain Exploration System. This system basically consists of three different components; a home station, a base station, and a number of drones. The home station is located on the Earth and is the graphical user interface to the base station which is located on the planet Mars. The base station controls the drones by requesting them to explore regions of the planet. The drones then will autonomously explore the assigned region and send exploration information back to the base station when they return. The base station constructs a map from the information it received from the drones and sends this map back to the home station where is is visualized. I took up various responsibilities in this project, viz., technical project leader, architect, and developer. As technical project leader, I provide the technical steering of the project team. This role combines nicely with the architect role in which I am responsible for the overall system architecture. In the time that is left I contribute to the development of the embedded software that runs on the drone. The technologies used in the project include C, C++, Eclipse, TCP/IP, Infra-red, Linux, Windows, QNX, CVS, UML, and RUP.

In future consumer electronics chips, found in TV sets or mobile phones for example, interconnecting the on-chip components is getting increasingly difficult. This is caused by the increasing complexity of these chips together with the tightening real-time constraints of the on-chip communication. Within the AEthereal project we provide a scalable solution to this interconnection problem. The solution is a communication network on a chip; a network of routers (switches) and network interfaces that provide scalable communication services to the connected IP blocks on the same chip. Our focus has been the design and exploitation of these networks. We developed a method to cost efficiently provide guaranties for both communication throughput and latency. Based on this method a software tool-chain has been set-up that automatically generates network instances that meet the communication requirements specified by the system designer. Within the project I mainly was responsible for the development of the method that provides the performance guarantees, the implementation and verification (in VHDL and ncsim) of a number of network components composing the network, and contributing to the development of the software tool-chain (in C++/SystemC and TCL). This includes issues like scheduling, performance analysis, and deadlock analysis. I also wrote scientific publications and patent proposals.

My research goal was to develop a parallelizing compiler for the class of so called ``piece-wise regular programs''. These programs are characterized by a high degree of regularity in their computational structure. Many digital signal processing algorithms belong to this class. The compiler's job is to convert a program (written in Matlab) into a dataflow description, that is, a highly parallel version of the program. The internals of the compiler are based on polyhedral theory and enumerative combinatorics. I used Java, C, and the JNI to implement the compiler. I further got introduced in technologies like design patterns and UML. As part of the work and in cooperation with the NFRA (Netherlands Foundation of Radio Astronomy (Astron in Dutch)) I contributed to the design of a beam forming algorithm.

During my graduation project, I researched the applicability of fractal image coding for the compression of video data. The aim was to compactly describe the video sequence but to keep the displayed video visual attractive. The result of my project is a method and prototype application written in C.