Joint DSRC and Radar Advanced Localization

Rapid MIMO-OFDM Prototyping

Joint DSRC and Radar Advanced Localization

Collision detection and prevention software has become a necessity for safety in vehicular environments. The majority of current collision detection systems use radars, lasers, or cameras. The goal of this project is to demonstrate that systems that rely solely on radar are not as effective as systems with both radar and dedicated short-range communications (DSRC). A team of undergraduate students has created a system that combines radar and DSRC data into a single stream of possible collisions that can be consumed by car manufacturers for functions such as automatic braking or alerting the driver. The functionality and effectiveness of this joint system in vehicular environments is demonstrated.

System Implementation

The joint system is composed of a Delphi Electronically Scanning Radar (ESR) and the CohdaWireless MK5-OBU DSRC module. In addition, two-way communication between radar and DSRC is enabled with a CAN bus. Data is retrieved and processed by a Python script, which has been made available on github. Once the received data has been logged and parsed, the combined detection from both modules are visually represented. The script performs image processing to overlay results from both modules and records the data in video format.

Rapid MIMO-OFDM Prototyping

Recent Results

Performance of the joint DSRC and radar system was tested in various real-world vehicular scenarios, such as blind intersection, lane change, or long-distance scenarios. Data was collected by attaching the joint system to the CAN bus of two different vehicles, and driving around certain areas of the UT campus multiple times. Results have demonstrated that the joint system performs no worse than each individual module alone. In several cases, the percentage of correctly identified vehicles experienced an increase of up to 25% when using the joint system in place of the best performing module (DSRC or radar). Future work will address collision avoidance algorithms based on these results.

This research is partially supported by the U.S. Department of Transportation through the Data-Supported Transportation Operations and Planning (D-STOP) Tier 1 University Transportation Center and by the Texas Department of Transportation under Project 0-6877 entitled Communications and Radar-Supported Transportation Operations and Planning (CAR-STOP). Special thanks to AutonomouStuff for giving us a hardware discount.

Comments are closed.