Sonar-Based Sensors for Autonomous Vehicles Using Passive and Active Metamaterials

Project #3.15 began in 2019 and was completed in 2020.

To operate autonomously in dynamic and uncertain environments, autonomous vehicles (AVs) must construct and continuously update a world model. Hence, they require robust, reliable, low power sensors that provide a 3D map of the surroundings in real-time while operating stealthily for long periods of time and in a variety of adverse, changing conditions (rain, snow, fog, bright light, and darkness). LIDAR is the most prominent sensor system in use.

We proposed that ultrasound (US) sensor systems acting as the acoustic analogues of LIDAR maintain efficiency in any kind of weather including rain, snow and fog. Moreover, they are stealthy because sound is rapidly absorbed in air and does not travel too far. Nevertheless, the current generation of US sensors have significant limitations that have prevented their widespread use. Their biggest drawback is that they produce maps of the surroundings of very low resolution. In addition, US beams spread much faster than laser beams due to their much larger wavelength, which leads to a significant reduction in sensor range. Current US sensors have a range of less than 10 meters.

We envision that manipulating acoustic waves after their generation and before their sensing with nonlinear artificial materials having carefully engineered micro-structure will address the limitations of current acoustic imaging technology and will lead to performant US sensors for 3D mapping of the environment.

This research addresses outstanding challenges of sonar-based sensors and seek to answer several fundamental research questions:

• What are the physical principles that enable single-ultrasound-transducer solid state sensors (SUTSS) with minimal moving parts that can provide 3D maps of surroundings?
• How to evaluate and enhance the resolution, range and performance of SUTSS in various weather conditions? (a single transducer approach ensures low power and cost)?
• How to classify the various objects appearing in the 3D maps produced by the SUTSS?

The main goal of this project is to create sonar-based sensors for autonomous vehicles using passive and active metamaterials. We envision acoustic sensors that create high-resolution images of same quality as low-end LIDAR systems, employ a single active transducer element and thus are low power and low cost, are stealthy, and operate for long periods of time in any weather.

Patent:

• US12190857B2 published 2025-01-07. Metamaterial-based Acoustic Sensor Beamforming. Inventors: Bogdan Epureanu; Hyung Suk Kwon; Bogdan Ioan Popa.  

3.15 / 3.A61