Signal processing for environment-aware radar (SPEAR)

Featured
to Research
Automotive radar Distributed signal processing Radar-LiDAR fusion Occupancy grid mapping

SPEAR studies signal processing for connected automotive radar systems. Its goal is to use vehicular connectivity and complementary sensing to improve perception in mixed traffic, where vehicles do not all have the same sensors or view of the road.

Project data

  • Researchers: Geethu Joseph, Nitin Myers, Peiyuan Zhai, Rupam Chakraborty
  • Starting date: February 2023
  • Closing date: February 2027
  • Funding: 2000 kE; related to group 1000 kE
  • Sponsor: NXP, TKI
  • Partners: NXP
  • Contact: Geethu Joseph

Research themes

SPEAR brings together four PhD sub-projects: digital mmWave radar waveform design, interference-aware distributed radar sensing, multimodal sensing for automotive radar, and surface-aided radar sensing.

My work belongs to the multimodal sensing for automotive radar sub-project. It investigates how radar can be complemented by situational information from sensors and vehicle systems such as cameras, LiDAR, GPS, control, and navigation units.

The project develops advanced signal processing methods for next-generation automotive radar. Rather than treating each sensor as isolated, it considers how radar data can be processed across a communication backbone so connected vehicles can cooperate.

The project is part of TU Delft’s Signal Processing Systems research line on distributed autonomous sensing systems.

Project context: TU Delft Signal Processing Systems.

SPEAR project architecture

Research question

How can connected vehicles combine radar, communications, and complementary sensors to perceive difficult traffic scenes more reliably?

Contribution

SPEAR develops signal-processing methods that make connected automotive sensing more cooperative, informative, and robust.

My PhD sub-project

Multimodal sensing for automotive radar

These connected research strands form my contribution within SPEAR.

Spatial sparsity-aware fusion of radar and LiDAR for automotive occupancy-grid mapping.

Radar-LiDAR fusion · Occupancy grid mapping · Automotive perception

Using camera-derived prior information to improve automotive occupancy-grid mapping.

Occupancy grid mapping · Camera sensing · Sparse Bayesian learning

Developing occupancy-grid mapping methods that better represent changing automotive scenes.

Dynamic occupancy-grid mapping · Autonomous perception · Automotive radar