Sebastian Hegler

Evaluation of OPC UA secure communication in web browser applications

OPC UA XML Web services mapping offers a Web service interface to access process data. Web services use XML technology for data exchange. Present-day Web browsers include XML functionality already as a standard feature, they are therefore very promising candidates for the implementation of monitoring and operating functions for industrial processes. However, the acceptance of Web services in industrial automation depends on adequate security realizations. For this purpose, the Web services security stack provides several specifications to meet the requirements for secure message exchange. The OPC UA XML Web services mapping refers to these specifications. The application of Web browsers for monitoring and operating of technical processes using OPC UA Web services demand the computation of cryptographic algorithms within the scripting engine of the Web browser. However, available scripting languages are not designed to compute complex mathematical, i.e., cryptographic, algorithms. Therefore, at the Institute of Automation of the Technische Universitat Dresden the suitability of a native Web browser for monitoring and operating of industrial processes with OPC UA based secure communication was analyzed. The paper shows representative measured computing times of cryptographic algorithms in JavaScript. The security specification XML signature - which is mandatory for OPC UA Web services mapping - requires about 700 ms to create a signature. Finally, the paper discusses methods to improve the performance.

The Semantic Web in action: semantically enabled Device Descriptions

In this paper, an approach to overcome the limits of XML encoded device descriptions by employing semantic Web technologies is presented. Starting with an illustration of the state of the art of working with XML device descriptions, reasons for a need to overcome these limits are given. The semantic Web, as idea and implementation, is then introduced shortly, followed by a presentation of the architectural concept. To conclude, this paper will highlight upsides and downsides of the system, and give a short overview over planned work.

Numerical computation of radar echoes measured by MARSIS during phobos flybys

The Mars Advanced Radar for Subsurface and Ionosphere Sounding, “MARSIS”, on board MarsExpress is the first and so far only space borne radar that observed the Martian moon Phobos. Radar echoes were measured for different flyby trajectories. The primary aim of the low frequency sounding of the crust of Phobos is to prove the feasibility of deep sounding. In this paper we present a numerical method that allows precise computation of radar echoes backscattered from the surface of large objects. The software is based on a combination of a Physical Optics calculation of surface scattering of the radar target, and a Method of Moments approach to calculate the radiation pattern of the whole space borne radar system, whereby the calculation of the frequency dependent radiation pattern takes into account all relevant gain variations and coupling effects aboard the space craft. This paper explains the simulation techniques and presents a comparison of simulation results for different orbits, and an interpretation of the backscattered signals.

Hybrid CPU-GPU computation of adjoint derivatives in time domain

For the solution of large-scale inverse scattering problems - in either acoustic or electromagnetic domain - gradient based optimization approaches are a method of choice, especially when the derivatives regarding the parameter of interest can be obtained from adjoint fields [1], [2]. Gradients regarding a parameter can be effectively computed using an adjoint approach where the direct and adjoint fields are integrated in opposite temporal direction. This yielding high memory consumption, the memory reduced computation of the gradients using checkpointing and recomputation of states from the checkpoint is a method of choice. We propose the use of graphics processing units (GPU) to accelerate the computation by solving the direct problem on the GPU and the adjoint problem on the CPU of the computer. The implementation of pipelining based on CUDA streams and pinned memory masks the memory transfer between host and GPU and allows for the computation of the adjoint derivatives at only a little more than twice the time of the solution of the direct problem.

PSTD-based approach to a large-scale inverse scattering problem

The three-dimensional finite-difference time-domain method (FDTD) simulation of electromagnetic wave propagation in large structures, like the cometary nucleus of 67P/Churyumov-Gerasimenko, as well as the inverse problem of reconstructing the permittivity distribution inside the comet nucleus is computationally expensive. With approximative methods not all propagation phenomena can be sufficiently modeled. For the efficient solution of this kind of inverse problem we propose the use of pseudo-spectral time-domain method (PSTD). This method overcomes the computational burdens and memory demands of the FDTD and allows accurate modeling of wave propagation for the CoNSERT case. In combination with optimal checkpointing and a state-of-the-art optimization tool, we demonstrate the solution of inverse problems at a scale of 50 wavelengths.

Enhanced GPR target classification by Compressed Sensing and radar polarimetry

Polarimetric technology has been one of the most important advances in microwave remote sensing during recent decades. The Entropy-α decomposition, which is a type of polarimetric analysis technique, has been common for terrain and land-use classification in polarimetric synthetic aperture radar. For certain scenarios, this kind of processing is also of interest for the interpretation of Ground Penetrating Radar (GPR) measurements. However, due to the limitation of the imaging resolution, which affects the performance of the polarimetric analysis, the classification of subsurface targets is not as reliable as in the original SAR scenario. In this paper we will apply blind support space (BSS) based compressed sensing (CS) to improve the performance of the polarimetric classification of GPR data. This is beneficial, because the support space is usually unknown and location-dependent. We demonstrate the feasibility of this approach based on full polarimetric data measured during a campaign in a controlled environment. The results from these measurements show that a BSS and CS based polarimetric GPR provides significant advantages in targets classification compared to standard GPR methods.

Simulative Ultraschall-Untersuchung von Pitch-Catch-Messanordnungen für große zylindrische Stahl-Prüflinge und gradientenbasierte Bildgebung

Zusammenfassung Große zylindrische Stahlprüflinge werden mittels der Methode der finiten Differenzen im Zeitbereich (engl. finite differences in time domain, FDTD) simulativ untersucht. Dabei werden Pitch-Catch-Messanordnungen verwendet. Es werden zwei Bildgebungsansätze vorgestellt: ersterer basiert auf dem Imaging Principle nach Claerbout, letzterer basiert auf gradientenbasierter Optimierung eines Zielfunktionals.

Reconstructing surface permittivity distributions from RADAR measurements using a sparsity inducing regularized least-squares approach

The estimation of surface properties of small bodies in our solar system like asteroids, comets and small moons from RADAR observations is a particularly challenging task. Due to limitations in orbitography and instrument dynamics the associated inverse problems are usually ill-posed. Common methods in the solution of these kind of problems are regularized least-squares approaches where the proper selection of the parameter weighting target function and regularization penalty is of great importance and not trivial. In this paper we propose a methodology based on a compressive sensing technique and demonstrate its feasibility on a surface permittivity reconstruction based on the well known method of physical optics (PO). We show that the proposed algorithm yields proper and stable reconstruction results in a sparse representation of the surface permittivity distribution for sub-optimal choices of the regularization parameter.

Method of iterative Physical Optics for the estimation of the scattering behavior of large layered dielectric bodies

The correct estimation of the scattering behavior of large dielectric objects is a particularly challenging task and well suited for the application of the Method of Physical Optics (PO). In this paper we propose the application of an iterative extension of the physical optics approach to the estimation of the scattering behavior of large layered dielectric bodies. With a suitable truncation and an efficient implementation layered structures can be modeled and simulated sufficiently accurate by the proposed method.

Acceleration of a physical-optics simulator using CUDA

This paper reports on the current state of a work-in-progress porting of a physical-optics simulation tool onto NVIDIAs CUDA platform. Current accelerator APIs are shortly presented. Our choice for the CUDA platform is explained, as well as the data flow of the simulation tool. The current state of the implementation of the port is presented, as are first run time measurements. The results are promising; however, the implementation is quite näıve at the moment. Further work will address the mentioned issues and increase the performance of the simulation tool.