Radar signal classification and fusion with heterogeneous senors using machine learning techniques

Introduction:

Radar technology has seen significant advancements, yet processing and interpreting radar signals effectively remains a challenge. The integration of radar with other sensor types, such as cameras and lidar, is promising to enhance signal interpretation and categorization. By combining data from multiple sensors, we aim to create comprehensive solutions for representative and highly current use cases.

Description:

This research aims to develop advanced techniques and machine learning models for processing and classifying radar signals. We will investigate the potential of integrating data from diverse sensors, such as cameras, to create robust final solutions and automate data annotation. The platform developed will be validated through practical use cases, like vehicle detection and classification OR gesture recognition in real-life scenarios (e.g. gesture recognition for car infotainment system control).

Expected Outcomes:

– Advanced Techniques: Introduction of novel machine learning techniques for radar signal processing and classification.
– Data Fusion Innovations: Development of effective data fusion methods that integrate radar with other sensor modalities for improved signal interpretation.
– Automated Annotation: Creation of automated systems for data labeling, which will significantly reduce the time and effort required for manual annotation.
– Practical Applications: Demonstration of the platform’s capabilities in real-world scenarios, offering practical solutions for vehicle detection OR gesture recognition in real-life applications.