Event

PhD Defence: mmWave Cognitive Radar: Adaptive Waveform Design and Implementation

  • Conférencier  Ehsan Raei (SPARC group)

  • Lieu

    Online & JFK building E004/E005

    LU

We’re happy to welcome you to the PhD Defence of Ehsan Raei (SPARC group) on 27 June 2022 at 14:00.

The event will take place digitally on WebEx. Click here to join.

Members of the defense committee:

  • Prof Dr Yves Le Traon, University of Luxembourg, Chairman
  • Dr Ezekiel Soremekun, University of Luxembourg, Deputy Chairman
  • Prof. Dr Michail Papadakis, University of Luxembourg, Supervisor
  • Prof. Dr Paolo Tonella, Università della Svizzera Italiana (USI), Member
  • Prof. Dr Zhang Lingming, University of Illinois at Urbana-Champaign, Member
  • Dr Thierry Titcheu Chekam, SES Satellites, Expert with advisory capacity

Abstract:

This research focuses on waveform design and implementation for mmWave cognitive radar systems. Cognitive radar refers to systems that can interact intelligently with their surroundings by adapting both the transmitter and the receiver. Indeed, the radio frequency spectrum will continue to become more crowded, and in this context, the new generation of radar systems will require to sense the environment and avoid making interference from other systems, like communications. To this end, cognitive radars require to have intelligent transmission strategies by utilizing waveform diversity and performing resources management. In general, the most essential resources available for radar systems are time (range), frequency, polarization, and spatial, and managing these results in waveform diversity for radar systems. Diverse waveforms and Multiple-Input Multiple-Output (MIMO) radars are concepts that are inextricably linked.

When compared to the traditional phased array systems, MIMO radar offers a variety of transmission strategies derived from different optimization objectives and constraints that improve angular estimation and detection performance.  In colocated MIMO radars, waveform design can be divided into two categories; uncorrelated and correlated waveform sets. In the first group, waveform optimization is being performed in order to provide a set of nearly orthogonal sequences to exploit the advantages of the largest possible virtual aperture. In this case, the sequences in the waveform set need to be orthogonal to one another in order to be separated on the received side. In addition, small autocorrelation sidelobes avoid masking weak targets by the range sidelobes of strong targets and mitigate the deleterious effects of distributed clutter. In this context, we propose Weighted BSUM sEquence SeT (WeBEST) approach to minimize the-norm ( and) of auto- and cross-correlation based on Block Successive Upper Bound Minimization (BSUM) method. This work offers a flexible framework to design waveforms with different properties. For instance, by choosing, and a waveform with a good Integrated Sidelobe Level (ISL), Peak Sidelobe Level (PSL), and sparse auto- and cross-correlation can be obtained, respectively. Through the numerical results, we compare the performance of our method with the state of the art. We show that the proposed method is able to meet the ISL lower bound when and decrease the PSL gap with the Welch lower bound when. Furthermore, by selecting the appropriate weights we can focus the sidelobe minimization in a specific range. In the second category, a correlated set of waveforms is transmitted to form a directional probing beam pattern on the transmit side. Because only the waveform correlation matrix needs to be optimized in this case, phase shifters can be removed on both sides of transmitting and receiving arrays, lowering hardware costs, which is important in mass production. As a result, the probing signal can be used as a tool to improve radar performance by increasing the SINR. This, however, necessitates knowledge from the environment, which can be obtained through a cognitive paradigm. In this context, we offer multiple beampattern shaping metrics for MIMO radar systems, including spatial-ISLR and beampattern matching. To tackle the resulting problems, we introduce several optimization strategies based on Coordinate Descent (CD), Semi-definite Relaxation (SDR), penalty approach, and BSUM. We exhibit the performance of the proposed methods and compare them to their state-of-the-art equivalents in the numerical results.

In the third investigation, we show that a good beampattern response results in a correlated waveform, whereas orthogonality necessitates as many uncorrelated waveforms as possible. As a result, beam pattern shaping and orthogonality are mutually contradictory. In this regard, we propose UNImodular set of seQUEnce design (UNIQUE) methods to make a trade-off between these two metrics, based on the CD approach. In this method, we consider the weighted sum of spatial-Integrated Sidelobe Level Ratio (ISLR) and range-ISLR as design metrics. Adjusting the weight between these two ISLR, plays an important role to make the trade-off. In the numerical results we show that, by choosing an appropriate weight, the waveform is able to discriminate the target and mitigate the interference simultaneously. However, this cannot be achieved when minimizing just the spatial- or range-ISLR individually.

We look at resource management for three different types of resources in the fourth study: time, frequency, and spatial domains. We propose Waveform design for beampattern shapIng and SpEctral masking (WISE) as a framework for tailoring the beampattern in MIMO radar systems while maintaining the proposed waveform’s unimodularity, desirable spectral occupancy, and orthogonality. The problem formulation leads to non-convex quadratic fractional programming. We propose an effective iterative to solve the problem, where each iteration is composed of a Semi-Definite Programming (SDP) followed by eigenvalue decomposition. Some numerical simulations are provided to illustrate the superior performance of our proposed over the state-of-the-art. 

The next research study focuses on maximizing Signal to Interference and Noise Ratio (SINR) through joint waveform and filter design. Two algorithms based on CD and the Alternating Direction Method of Multipliers (ADMM) are proposed. The numerical results reveal the improvements made by the proposed algorithms over the state of the art.

The final study looks at the problem of joint spectral shaping and waveform orthogonality in MIMO radar systems. The Parseval theorem is used in the first part to combine the two objectives related to orthogonality and spectral behaviour of the waveforms into one objective. In the second part, we use CD framework to optimize a weighted sum of the two aforementioned objectives. The waveforms, along with the receive processing, are designed to enhance the radar detection performance while avoiding certain frequency bands occupied by communications interference. For a representative scenario of cognitive radars, the designed system without loss of generality is then implemented using a custom-built Software Defined Radio (SDR) based prototype developed on Universal Software Radio Peripheral (USRP) from national instruments. These USRPs are quite flexible in terms of transmitting waveforms but operate at sub-6 GHz frequencies with a maximum instantaneous bandwidth of 160 MHz. However, the implemented framework and the design methodologies can be applied and utilized in mmWave frequencies.

This thesis concludes by summarizing the main research findings and some remarks on future directions and open problems.