Large conformal network of radars for 360° perception: system design and algorithms

Radar is becoming a ubiquitous technology for consumer, automotive, industrial and medical applications. Constant progress in semiconductor technologies enable to design low power, advanced digital processing and millimeter-wave circuits in highly miniaturized and low-cost ICs. Countless applications are now becoming possible. One of them, autonomous navigation, is witnessing a revolution leveraging advanced sensors such as radars. One of the key challenges for autonomous vehicles is the all-round, short-range sensing e.g. covering the distance from 0 to 50 cm at 360 degree or even 4 pi steradians around a vehicle.

The research of this PhD aims at solving this problem. We envision a network of ultra-low-power and ultra-low-cost radars arranged in a lattice topology and covering a vehicle surface. By suitably connecting the radars together, synchronizing them and jointly processing their signals, a 3D rendering of the immediate vicinity of the vehicle is extracted in real-time and further processed with detection, clustering, tracking algorithms. Higher level functionalities such as simultaneous localization and mapping, obstacle avoidance, etc. become then possible. The key is to use many simple low-resolution radars to achieve a high-resolution perception of the environment. The concept can find many applications in automotive, transportation, robotics, drone, etc. applications.

The goal of this PhD is to conceive the whole system of this 3D perception. The best radar frequency, bandwidth, waveform and processing must be selected with low-power and low-cost property in mind. The imaging and array processing algorithms must be developed, taking into account the partitioning between local, distributed and centralized processing. Implementation challenges such as synchronization between radars, communications, impact of hardware non-idealities must be investigated in detail. The proposed solutions will be validated first by simulation then with real radar hardware.

This concept will be researched by two PhD students working in close collaboration. The topic described here is for the system design and algorithm part. Another PhD student will work on the hardware realization (chip and antenna design, implementation in a lattice structure, interfacing between chips, ...).

The successful PhD candidate will be part of a large IMEC team working on the research, implementation and prototyping of future radar systems composed by experts in digital, analog and mm-wave design, radar and wireless communication systems, signal processing and machine learning algorithms. This is a unique opportunity to develop innovative, multi-disciplinary technology and shape future radar networks. You will publish your research in top-level journals and conferences.

Required background: Signal processing for wireless technology (communication or radar), knowledge in multi-antennas signal processing, proficiency with Matlab or Python is a must. Some knowledge of radar or SAR concepts is a plus.