Higher Resolution Millimeter Wave Imaging Using Radar Fusion
Event details
| Date | 23.08.2023 |
| Hour | 15:00 › 17:00 |
| Speaker | Samah Hussein Youssef |
| Location | |
| Category | Conferences - Seminars |
EDIC candidacy exam
Exam president: Prof. Sabine Süsstrunk
Thesis advisor: Prof. Haitham Al Hassanieh
Co-examiner: Prof. Mathieu Salzmann
Abstract
The usage of millimeter-wave radars for perception is becoming increasingly prominent in multiple fields such as automotive, industry and security. This is due to their robust perception capability in different environments, as well as carrying higher depth information than other common perception sensors like cameras. However, these devices face several challenges that hinder their usage as the main sensor in such applications, such as low angular resolution, multipath, and specularity. These challenges make it difficult to reach a meaningful scene understanding. In this work, we explore the usage of multiple mmWave radars for imaging. We show that combining the reflections from two radars provides a higher resolution scene mapping and more accurate depth information than just one radar or stereo cameras.
Background papers
1. See Through Smoke: Robust Indoor Mapping with Low-cost mmWave Radar (millimap) https://dl.acm.org/doi/pdf/10.1145/3386901.3388945
2. Hidden Gems: 4D Radar Scene Flow Learning Using Cross-Modal Supervision https://arxiv.org/pdf/2303.00462.pdf
3. Pointillism: Accurate 3D Bounding Box Estimation with Multi-Radars https://arxiv.org/pdf/2203.04440.pdf
Exam president: Prof. Sabine Süsstrunk
Thesis advisor: Prof. Haitham Al Hassanieh
Co-examiner: Prof. Mathieu Salzmann
Abstract
The usage of millimeter-wave radars for perception is becoming increasingly prominent in multiple fields such as automotive, industry and security. This is due to their robust perception capability in different environments, as well as carrying higher depth information than other common perception sensors like cameras. However, these devices face several challenges that hinder their usage as the main sensor in such applications, such as low angular resolution, multipath, and specularity. These challenges make it difficult to reach a meaningful scene understanding. In this work, we explore the usage of multiple mmWave radars for imaging. We show that combining the reflections from two radars provides a higher resolution scene mapping and more accurate depth information than just one radar or stereo cameras.
Background papers
1. See Through Smoke: Robust Indoor Mapping with Low-cost mmWave Radar (millimap) https://dl.acm.org/doi/pdf/10.1145/3386901.3388945
2. Hidden Gems: 4D Radar Scene Flow Learning Using Cross-Modal Supervision https://arxiv.org/pdf/2303.00462.pdf
3. Pointillism: Accurate 3D Bounding Box Estimation with Multi-Radars https://arxiv.org/pdf/2203.04440.pdf
Practical information
- General public
- Free
Contact
- edic@epfl.ch