Self-driving electric vehicles continue to face significant challenges on their journey towards reliability. However, researchers from the Department of Energy’s Oak Ridge National Laboratory (ORNL) and Western Michigan University (WMU) are collaborating to drive solutions from outside the vehicle itself, by integrating sensors and processing into road infrastructure.
In partnership with other organizations, ORNL engineers are placing low-powered sensors within reflective raised pavement markers, commonly used to assist drivers in identifying lanes. These microchips transmit road shape information to passing cars, even in situations where vehicle cameras or remote laser sensing (LiDAR) may be unreliable due to factors like fog, snow, or glare.
The aim is to make autonomous driving features more accurate and safe in remote areas by repurposing existing infrastructure. Beyond providing precise environmental information, this technology offloads some of the processing load from the vehicle’s software onto the infrastructure. Consequently, it saves electric vehicle battery power, extending the driving range and promoting wider adoption of electric vehicles. In fact, according to a technical paper, the chip-enabled pavement markers can reduce navigational power consumption by up to 90% compared to leading camera and LiDAR-based autonomous driving technologies.
The potential applications of this technology extend beyond future self-driving vehicles and can enhance existing autonomous driving features such as lane assist. WMU leads a larger project involving research and industry partners to develop related sensor and autonomous driving technologies like radar retro-reflectors, high-definition mapping, computational offloading, and weather sensing. Field testing is being conducted by WMU researchers, who are also measuring the reduction in vehicle energy use enabled by these technologies using a vehicle driving on a closed course.
ORNL researchers experimented to determine the optimal combination of transceiver, battery, and antenna for the sensor package within standard road markers, including those designed to withstand snowplows. They developed a communications protocol that utilizes radio frequency spectrum hopping up to 50 times per second, offering resistance against interference, low cost, and low power consumption. The equipment is designed to ensure that the sensor’s battery lasts for the same replacement cycle as the pavement markers, typically a year.
The ORNL team created algorithms that leverage GPS coordinates of lane markers to reconstruct an image of the drivable area. One algorithm is embedded in the microchip inside the pavement marker, while a decoding algorithm is integrated into the vehicle’s software. The sensor platform underwent rigorous field testing in various weather conditions, even in a remote national park in Montana with no wireless access. Impressively, it exceeded the initial goal of transmitting within a 100-meter range, offering a transmission distance of more than five times that.
The sensor-equipped markers have the potential to signal temporary lane shifts or closures in construction zones, providing real-time information when high-definition maps are outdated. In the future, these marker sensors could convey additional data about temperature, humidity, and traffic volume. The project team plans to collaborate with students to develop a smaller microchip for the markers, offering a cost-effective alternative to off-the-shelf products.
To facilitate widespread implementation, road demonstrations are being planned for stakeholders, including transportation departments in Tennessee and Michigan, the Michigan Office of Future Mobility, and the City of Chattanooga. Involvement from government agencies responsible for infrastructure decisions is vital to drive the adoption of these technologies.
While self-driving vehicles have traditionally been viewed as a software problem, Asher from WMU highlights that a more patient approach, combining infrastructure-based hardware with government transportation agencies, may be the key to achieving zero-accident vehicles that use energy sustainably. The integration of sensors within road infrastructure presents a promising path towards safer and more efficient autonomous driving.
By Impact Lab
