Master Thesis, 30 HP:Joint High Resolution DOA Estimation

Your role

Background

Solutions based on phased array antenna technology with individual transmit- and receive channels for each antenna element, "digital arrays", have many advantages in for instance, Electronic Warfare, EW- and radar applications. In particular, a large number of different 'high-resolution' methods for simultaneous estimation of parameters such as, incident angles and Doppler frequencies of several targets can be applied. Examples of frequently applied Direction-Of-Arrival (DOA) estimation algorithms include MUSIC, WSF and TLS-ESPRIT, as well as variations thereof.

In practice, the gain- and phase response of the different transmit- and receive channels will vary. In addition, there might be uncertainties in actual sensor locations. Also, mutual coupling effects between different antenna elements may contribute to reduced performance. Different kind of errors may be uncorrelated or possibly correlated in time and/or space, (from one antenna channel to another).

In order to reduce this kind of imperfections, the antenna array is calibrated, using various test signals. These signals can be induced into the antenna elements using a transmitter with known characteristica operating in the near- or far field. Alternatively, internally generated or externally applied calibration signals can be inserted into the array receiver paths via switches located at the antenna elements. In general, the latter approach is less complicated although, effects of the antenna elements are excluded, which is a limitation.

The measured deviations from the nominal array response may be compensated for by "multiplying" each 'analog' antenna element input/output signal with a corresponding complex correction term. (Implemented using adjustable amplifiers/attenuators and phase shifters/time delay devices). As a possible alternative, this adjustment may be performed 'digitally' by means of correction of the array response used in the DOA estimation method. The latter approach might be easier to implement with respect to demands for high accuracy. Various forms of robust algorithms for arrays with calibration errors for instance, the Generalized Weighted Subspace Fitting, GWSF, have been proposed. It would be interesting to evaluate and compare these latter methods with algorithms which use training data, Machine Learning, ML based on neural networks, for obtaining accurate calibration corrections. ML algorithms can be expected to be particularly useful in presence of calibration errors exhibiting a time-varying, repetitive pattern. It might also be possible to combine more "traditional" model-based methods with ML algorithms.

Performance measures shall be evaluated by means of simulations using MATLAB and possibly also using measurements obtained from an experimental antenna array system, WIDAR.

Description of the master thesis

This project is suitable as a M.Sc. thesis work in Electrical Engineering or Engineering Physics.

* Using Matlab, evaluate the DOA-estimation performance achieved using the above-mentioned high-resolution parameter estimation methods together with "digital" array response compensation for various types of array imperfections. Methods applied for calculation of array response correction shall be based on array response perturbation models as well as based on training data, (Machine Learning).
* The resulting performance shall be compared to that obtained instead using "analog" calibration, as well as with that obtained for the case with no calibration.
* (If possible, the analysis shall also be made using broadband signals).
* Possible performance shall also be investigated in a similar way using measurements obtained from experiments with the WIDAR antenna array.
* Analyze and compare these latter results with those previously obtained using simulated receive signals.
* Document results in a report.

Your profile

Desired skills and areas of interest:

* Successfully completed courses in Antenna Theory, Digital- / Adaptive Signal Processing, Algebra or equivalent.
* Good skills in programming, preferably Matlab.
* Knowledge of basic RF receiver- and transmitter architectures.

As a condition for thesis employment, a candidate needs to participate in a background investigation and receive applicable Government security clearance in accordance with the Swedish Protective Security Act.

What you will be a part of

You will be part of an extremely competent team and have a very innovative work environment. We are responsible for the future functionality, technologies and performance of our products in order to meet present and future user needs. Read more about us here .

Last application day

30-11-2023



ADVERTISING PERIOD

01-10-2023 -- 30-11-2023



Contact information

Neda Tooloutalaie

Manager System Concepts and Studies

Ph.: +46 10 215 0055

Cell.: +46 734 460055

E-mail: neda.tooloutalaie@saabgroup.com



Anders Nordsjö

Master Thesis Supervisor

Ph.: +46 10 215 6192

Cell.: +46 734 376192

E-mail: anders.nordsjo@saabgroup.com



Saab is a company with a strong people-orientation. We offer a friendly work environment where we support and help each other to be at our best. Continuous learning, career & talent development and employee well-being are examples of areas where we always put the strongest effort to offer great opportunities.