Constantin Timm

Dynamic web service orchestration applied to the device profile for web services in hierarchical networks

Based on the idea of Service Oriented Architectures (SOA), Web Services paved the way for open and flexible interaction between heterogeneous systems with a loose coupling between service endpoints. The Device Profile for Web Services (DPWS) implements a subset of WS-* specifications in order to make the advantages of the Web Service architecture available to a growing embedded systems market. In this paper we are proposing a service orchestration mechanism applied to services on top of a DPWS-based middleware. The approach is complementary to the rather complex and resource intensive Web Service Business Process Execution Language (WS-BPEL) and focuses on service orchestration on resource constrained devices deployed in hierarchical network topologies. We validate our service orchestration concept through its resource consumption and illustrate its seamless integration into the service development cycle based on the underlying DPWS-compliant middleware.

Towards Sensor-Actuator Coupling in an Automated Order Picking System by Detecting Sealed Seams on Pouch Packed Goods

In this paper, a novel concept of coupling the actuators of an automated order picking system for pouch packed goods with an embedded CCD camera sensor by means of image processing and machine learning is presented. The picking system mechanically combines the conveyance and singularization of a still-connected chain of pouch packed goods in a single machinery. The proposed algorithms perform a per-frame processing of the captured images in real-time to detect the sealed seams of the ongoing pouches. The detections are used to deduce cutting decisions in order to control the system’s actuators, namely the drive pulley for conveyance and the cutting device for the separation. Within this context, two controlling strategies are presented as well which specify the interaction of the sensor and the actuators. The detection is carried out by two different marker detection strategies: enhanced Template Matching as a heuristic and Support Vector Machines as a supervised classification based concept. Depending on the employed marker, detection rates of almost 100% with a calculation time of less than 40 ms are possible. From a logistic point of view, sealed seam widths of 20 mm prove feasible.

Design space exploration towards a realtime and energy-aware GPGPU-based analysis of biosensor data

In this paper, novel objectives for the design space exploration of GPGPU applications are presented. The design space exploration takes the combination of energy efficiency and realtime requirements into account. This is completely different to the commonest high performance computing objective, which is to accelerate an application as much as possible.As a proof-of-concept, a GPGPU based image processing and virus detection pipeline for a newly developed biosensor, called PAMONO, is presented. The importance of realtime capable and portable biosensors increases according to rising number of worldwide spreading virus infections. The local availability of biosensors at e.g. airports to detect viruses in-situ demand to take costs and energy for the development of GPGPU-based biosensors into account. The consideration of the energy is especially important with respect to green computing.The results of the conducted design space exploration show that during the design process of a GPGPU-based application the platform must also be evaluated to get the most energy-aware solution. In particular, it was shown that increasing numbers of parallel running cores need not decrease the energy consumption.

Trade-Off Analysis Considering Tomogram Quality and Performance of a Parallel Computing Hardware Realization of Katsevich's Reconstruction Algorithm

In this paper, the trade-off between the target variables tomogram quality and reconstruction time of an implementation of Katsevichs exact reconstruction algorithm for helical cone-beam computed tomography (CT) is analyzed. This is accomplished by means of an OpenCL-based, parallel and portable realization of Katsevichs algorithm. The detailed examination is carried out on a phantom object which is composed of cubes with different attenuation coefficients. For this phantom object, X-ray cone-beam projections are simulated with different helix trajectory parameters. From the simulated projection data sets 3D tomograms are reconstructed. The impacts of the helix parameters on the tomogram quality and the reconstruction time are measured. Moreover, the speedup in dependence of the number of computing units on the OpenCL device is examined. To sum up the results: the tomogram quality increases monotonically with ascending number of projections while keeping the helix pitch fixed. A linear relationship between the overall runtime and the number of projections exists. The trade-off analysis between the tomogram quality and the reconstruction time proves that an optimum is reached if the criteria for the lateral and vertical sampling of the 3D space are fulfilled.

Mobile Detektion viraler Pathogene durch echtzeitfähige GPGPU-Fuzzy-Segmentierung

Die vorliegende Arbeit stellt einen neuartigen Fuzzy-Logik-basierten Segmentierungsalgorithmus zur Detektion von biologischen Viren in stark Artefakt-behafteten Bildsequenzen vor, der konform ist zu den differenzierten Ressourcenbeschrankungen mobiler Endgerate. Als Sensor kommt der neuartige PAMONO-Biosensor zum indirekten Nachweis von Viren mittels optischer Mikroskopie zum Einsatz. Die Segmentierungen weisen eine hohe positive Ubereinstimmung bei idealisierten synthetischen Segmentierungen auf, die durch den Fuzzy-Ansatz insbesondere bei kleinen Viren/schlechtem Signal-Rausch-Verhaltnis nochmals verbessert wird. Ferner wird gezeigt, dass eine GPU-gestutzte Datenanalyse die Detektion viraler Strukturen in Echtzeit auf mobilen Endgeraten ermoglicht, und im Vergleich zur CPU den Energieverbrauch im Durchschnitt um Faktor 3.7 senkt.

Feedback-Based global instruction scheduling for GPGPU applications

In the face of the memory wall even in high bandwidth systems such as GPUs, an efficient handling of memory accesses and memory-related instructions is mandatory. Up to now, memory performance considerations were only made for GPGPU applications at source code level. This is not enough when optimizing an application towards high performance: The code has to be optimized at assembly level as well. Due to the spreading of GPGPU-capable hardware in smaller and smaller devices, the energy consumption of a program is --- besides the performance --- an important optimization goal. In this paper, a novel compiler optimization technique, called FALIS ( F eedback-based and memory- A ware g L obal I nstruction S cheduling), is presented based on global instruction scheduling and multi-objective genetic algorithms. The approach uses a profiling-based feedback in order to take the measured performance and energy consumption values inside a compiler into account. Profiling on the real hardware platform is important in order to consider the characteristics of the underlying hardware. FALIS increases runtime performance of a GPGPU application by up to 13.02% and decreases energy consumption by up to 10.23%.

Parameteroptimierte und GPGPU-basierte Detektion viraler Strukturen innerhalb Plasmonen-unterstützter Mikroskopiedaten

Die lokale Verfugbarkeit von effizienten und leistungsf ahigen Biosensoren, z.B. an Flughafen, gewinnt durch die zunehmende Verbreitung viraler Infektionen zunehmend an Bedeutung. Die zentralen Herausforderungen fur entsprechende in situ Virusdetektionssysteme sind eine schnelle und sichere Erkennung der Viren respektive die Adaptivitat an unterschiedliche Auspragungen von Erregern. Optische Verfahren, wie die neuartige Plasmonen-unterstutzte Mikroskopie von Nanoobjekten erlauben es, diesen Anforderungen zu entsprechen. Aufgrund starker multipler Artefaktbelastung des Signals und hohen Datenmengen (zeitlichen und ortlichen), werden nachhaltige Anforderungen an die Bildrestauration und -analyse gestellt. Hier setzt die vorliegende Arbeit an, welche eine GPGPU-basierte Bildrestaurations- und Bildanalysepipeline vorstellt. Uber eine Kombination aus lokaler und globaler Parameteroptimierung mittels Genetischer Algorithmen kann eine hohere Effektivitat der einzelnen Stufen der Verarbeitung erzielt werden, aber auch im ubergreifenden Verbund – dies zeigt sich nachhaltig in der Erkennungsrate des Biosensors fur Viren.

GPGPU-basierte Echtzeitdetektion von Nanoobjekten mittels Plasmonen-unterstützter Mikroskopie

Die Verfugbarkeit echtzeitfahiger und mobiler Biosensoren gewinnt durch die zunehmende Verbreitung viraler Infektionen zunehmend an Bedeutung. Im Gegensatz zu Virusdetektionsmethoden wie beispielsweise ELISA erlaubt die neuartige Plasmonen-unterstutzte Mikroskopie von Nanoobjekten, Proben innerhalb von wenigen Minuten auf Viren analysieren zu konnen. Die Herausforderung fur ein, auf dieser Analysemethode beruhendes In-situ-Virusdetektionssystem, besteht in der Echtzeitverarbeitung von extrem hohen Datenmengen. Hier setzt die vorliegende Arbeit an, welche eine hoch parallele GPU-basierte Verarbeitungspipeline zur echtzeitfahigen Virusdetektion vorstellt. Durch die konsequente Ausnutzung der GPGPU-Fahigkeiten von Grafikkarten kann auf teure Spezialhardware verzichtet werden, um eine echtzeitkonforme Beschleunigung notwendiger Bildverarbeitungs- und Bildanalysealgorithmen bereitzustellen, die auch den Anforderungen an ein eingebettetes Virusdetektionssystem gerecht wird.