news

       Researchers from Western Michigan University and Oak Ridge National Laboratory (ORNL) have proposed a new approach to improving the efficiency and safety of self-driving cars: embedding chips in road surface markings to better detect lane markings.
       “We’re working to make self-driving features more accurate and safer in more remote areas,” said Ali Ekti, a researcher at Oak Ridge National Laboratory who collaborated with lead author Sachin Sharma and colleagues. “We’re doing this by transforming the virtual infrastructure into something that can be used for multiple purposes.”
       While maintaining this basic functionality, the team aimed to introduce some intelligence into the markup to make it more useful for both autonomous vehicle systems and human drivers.
       The smart device, called a Chip-Elevated Road Marker (CERPM), fits entirely into the existing road marker market that was eliminated to make room for it. It’s built around a Heltec Automation WiFi LoRa 32 development board and programmed with the Arduino IDE to transmit the marker’s precise GPS coordinates and a unique ID tag. This information is then received by another WiFi LoRa 32 development board (in a 2016 Kia Niro Hybrid autonomous-driving research vehicle) and processed by a Python program using the Robot Operating System (ROS).
       The end result: Even if the camera or lidar is blocked, the car can clearly detect the shape of the road and the roadway, while reducing the computational load on the autonomous vehicle’s computer system. The detection range can reach 410 meters (approximately 1,345 feet), more than four times the original target distance. In field tests, prototype markers worked well in all weather conditions, but long-term power supply remained an issue.
       “Looking back on a decade of massive investment in research and development, we now realize that software and cameras alone will not be enough,” said co-author Zachary Asher. “A more patient approach, collaboration with state transportation agencies, and the use of infrastructure equipment may be the path to truly zero-failure, zero-emission vehicles.”
       The team’s research is published under proprietary conditions in the journal IEEE Sensors, with open-access copies available from the U.S. Department of Energy’s Office of Scientific and Technical Information.