Authors:

Siddharth Murali

Date Created:

Not specified

Course Title:

Professor:

Not specified

About Paper:

Many modern autonomous safety systems, such as the Autonomous Braking System (AEBs) rely on cameras, LIDAR and radar data to detect the presence of Vulnerable Road Users (VRUs). One major problem that these systems face, however, is that they can only detect the presence or absence of an object, and not its identity. This limitation makes it impossible to engineer a custom response for each object, which is often a necessity when dealing with different road users. In order to overcome this limitation, this research proposes identifying VRUs by their unique micro-doppler signature patterns. Micro-doppler signature patterns are unique patterns generated by the relative internal motion between different parts of the object. Identifying patterns in the micro- doppler patterns of different objects allows for object identification and classification. However, this hypothesis remains largely unverified experimentally. This project uses a DRVEGRD 152 radar to record and analyze the micro-doppler radar data to verify the hypothesis and to develop a method for object detection, identification and classification. The radar is attached to a vehicle along with a calibrated camera for data collection. Data for various VRU scenarios is then recorded and analyzed using SmartMicro's DriveRecorder software. The project validates the hypothesis that Vulnerable Road Users can be uniquely classified through micro-doppler signatures by analyzing collected data in controlled environments. Viability of the hypothesis is then tested against data collected in non-controlled environments. Proving the validity of the hypothesis allows for future research and development into the use of micro-doppler signatures in object identification and classification. Keywords: Micro-Doppler Effect; Vulnerable Road Users; Object Identification

Source:

Purdue University / 2024

Topics:

No topics listed

Co-authors:

Siddharth Murali