Australian researchers have developed technology that can allow autonomous vehicles to track running pedestrians behind buildings and cyclists that may be obscured by other vehicles. The vision technology, which the researchers liken to “X-Ray”-style vision, is capable of penetrating through to pedestrians in blind spots and detecting those who may be otherwise obscured. The iMOVE Cooperative Research Centre-funded project, along with the University of Sydney’s Australian Centre for Field Robotics and connected vehicle solutions company Cohda Wireless, released its findings after three years of R&D. The technology’s applications, which Cohda is commercializing, involve a technology called collective (or cooperative) perception (CP) that supports intelligent transportation systems (ITS). Using roadside ITS information sharing units equipped with sensors such as cameras or lidars, vehicles can share what they “see” with others via vehicle-to-X (V2X) communication. This allows autonomous vehicles to tap into various viewpoints that they might perceive. By virtue of being hooked up to the one system, the range of perception increases significantly to allow vehicles to see what they otherwise would not be able to see. Tests showed that a vehicle, when connected to the system and using CP information, was able to track a pedestrian who was visually obstructed by a building. “This was achieved seconds before its local perception sensors or the driver could possibly see the same pedestrian around the corner, providing extra time for the driver or the navigation stack to react to this safety hazard,” said Eduardo Nebot, a professor from the Australian Centre for Field Robotics. Australian researchers have developed technology that lets autonomous vehicles perceive and track pedestrians hidden behind buildings and cyclists obscured by larger cars, trucks, and buses. Here, a CP-enabled vehicle indicates detection of a cyclist behind a bus. Courtesy of Cohda Wireless. Another experiment demonstrated the CP technology’s ability to safely interact with walking pedestrians. In this case, it generated its response based on the perception information provided by the roadside ITS station. Tests also demonstrated the expected behavior of a connected vehicle when interacting with a pedestrian rushing toward a designated crossing area. “Using the ITS system, the connected autonomous vehicle managed to take preemptive action,” Nesbot said. The vehicle braked and stopped braking before the pedestrian crossing area, based on the predicted movement of the pedestrian. “The pedestrian tracking, prediction, path planning, and decision-making were based on the perception information received from the ITS roadside stations. Collective perception enables the smart vehicles to break the physical and practical limitations of onboard perception sensors.” Other vehicles — those not connected to the system — could also benefit from the technology, members of the development team said. And use of CP could improve awareness in multiple traffic scenarios. “CP enables the smart vehicles to break the physical and practical limitations of onboard perception sensors and embrace improved perception quality and robustness,” Cohda Wireless CTO Paul Alexander said. “This could lower per vehicle cost to facilitate the massive deployment of connected and automated vehicles technology." Additionally, Alexander said, CP for manually driven connected vehicles could be an attractive option to enable perception capability without retrofitting the vehicle with perception sensors.