MIMO Radar Antenna Array Calibration for DoA Estimation with Super-resolution Algorithms

Radars are seen as major components for applications requiring precise detection in 3D space with a small sensor (gesture recognition, autonomous driving cars, ...). This requires Multiple-input-multiple-output (MIMO) radars to estimate angular position.

The MIMO radar utilizes transmitter and receiver antenna arrays with multiple elements to form a virtual array to realize direction of arrival (DoA) estimation. The DoA estimation can be realized by digital beamforming (DBF). But to achieve higher angular resolution, required by autonomous driving cars for example, more advanced algorithms are needed.

Super-resolution algorithms like Capon, MUSIC or ESPRIT are popular for improving the angular resolution for DoA estimation. All the DoA estimation algorithms are degraded due to antenna array errors (amplitude and/or phase error, mutual coupling and position error). The impact of antenna array error for classical beamforming has been widely discussed in the literature.  However, the impact for super-resolution algorithm has not been sufficiently analyzed.

In this MS project, the student needs in a first phase to study the performance of super-resolution algorithms under various antenna arrays errors. In a second phase, he/she will propose calibration strategies and algorithms. The algorithm development will be done on Matlab or python. The validation will be done based on radar measurements done with a MIMO millimeter-wave radar.

The calibration is expected to work without the help of reference targets. Two environments will be considered: indoor environments, where the challenge is multipath propagation, and outdoor environments, where the challenge is the motion of the radar platform.

The work will include:

• Simulation of MIMO antenna error and impact on super-resolution DoA
• Implementation and analysis of phase calibration using state-of-the-art algorithms
• Investigation of innovative approaches of array calibration
• Validation using real MIMO radar recordings

The successful candidate must show a strong understanding of signal processing and linear algebra. Proficiency with Matlab and/or python is a must. Some knowledge of radar concepts is a plus

Type:

           - Master Thesis internship (6 months)

- Preceded by optional summer internship (max 3 months)

Responsible scientist(s): Marc Bauduin (Marc.Bauduin@imec.be)