A new UK project will research how ‘look-ahead’ and ‘non-line-of-sight’ (NLOS) perception can be delivered to any connected and autonomous vehicle (CAV) from infrastructure deployed sensors using low-latency over-the-air messaging.
The CORAM (Connected On-Road Autonomous Mobility) project will allow self-driving vehicles to perceive obstacles with pin-point accuracy around blind corners, pushing CAV sensing well beyond what they can currently recognize using onboard sensors.
The project will undertake a feasibility study as part of a CCAV3 funding competition sponsored by Innovate UK and the country’s Center for Connected And Autonomous Vehicles (CCAV). CORAM is being led by Propelmee, an AV technology company that is developing an operating system (OS) to enable full autonomy on any vehicle, globally, to enable Autonomous Mobility-as-a-Service (A-MaaS) applications.
The other partners in the CORAM consortium are the Transport Systems Catapult (TSC) and Cranfield University. The project’s work will pave the way for infrastructure-based connectivity to enable safer on-road autonomy, culminating in a technology demonstration at Cranfield’s new £19m (US$26m) state-of-the-art Multi-User Environment for Autonomous Vehicle Innovation (MUEAVI) research facility, using TSC’s autonomous pods.
Propelmee will develop the core capability of converting the infrastructure perception sensor inputs into meaningful scene understanding, as look-ahead and NLOS perception, beyond the line of sight and range of an AV’s onboard sensors.
The partners will develop and research methods for achieving these technology goals, and for efficient transmission of NLOS perception Over-the-Air (OTA) for use by CAVs, while assessing the latency and bandwidth constraints of chosen communication protocols.
The research outcomes of CORAM will support the creation of new message structures and communication protocols for CAVs, new use case concepts for connected on-road autonomy, and will pave the way for shared vehicle understanding that goes beyond current vehicle-to-vehicle (V2V) and vehicle-to-infrastructure/location (V2I) systems for coordinated autonomous driving.
“Connectivity has a key role to play in autonomy, and infrastructure sensing has the potential to greatly aid and enable full autonomy, especially when on-vehicle sensing struggles to accurately perceive the world or 3D maps are broken,” explained Zain Khawaja, founder and CEO of Propelmee.
“Infrastructure can assist autonomy in urban city centers where there is a large density of pedestrians, buses, taxis, cyclists, and complex and unpredictable driving scenarios often arise. The CORAM consortium will take the first steps in developing a proof of concept of such a system, and will help establish the groundwork for the UK to become a world-leader in connected infrastructure for fully autonomous mobility.”
Zain Khawaja will be speaking about autonomous mobility at The Future of Transportation World Conference 2018, which is taking place in Cologne, Germany, on June 19-20.