Engineers introduce human -type driving technology for autonomous vehicles

Autonomous cars will soon be able to “think” as human drivers in complex traffic environments, thanks to a cognitive coding frame built by a multidisciplinary research team from the Hong Kong University of Sciences and Technology School (HKUST).

This innovation considerably improves the safety of autonomous vehicles (AVS), reducing the risk of overall traffic by 26.3% and reducing potential damage to a high risk Road users like pedestrians and cyclists by an impressive 51.7%. Even the AV themselves benefited, their risk levels lowered by 8.3%, paving the way to a new framework to advance the automation of vehicle safety.

Existing AVs have a common limitation: their decision -making systems can only carry out risk assessments in pairs, by not considering the interactions between several road users in a holistic manner. This contrasts with a competent driver who, for example, can skillfully navigate in an intersection by prioritizing pedestrian protection while slightly compromising the safety of vehicles nearby. Once pedestrians are confirmed to be sure, the driver can then focus on nearby vehicles. This risk management capacity presented by humans is known as “social sensitivity”.

To empower AVS with social sensitivity, a research team led by Professor Yang Hai, Professor President of the Department of Civil and Environmental Engineering at HKUST, was inspired by neuroscience, human cognitive processes and ethics to develop a human -plausible coding diagram. This system allows AVS to perceive, assess and behave in a way resembling a thoughtful human conductor.

The article, entitled “Socially sensitive autonomous vehicles using human plausible coding” published in the newspaper Proceedings of the National Academy of Sciences.

This new system incorporates three innovative features:

1. Individual risk assessment – assesses the risk facing each road user, including pedestrians, cyclists, motorcyclists and nearby vehicles. This implies evaluating their speed, their distances from each other and behavioral predictability. For example, a child walking near the road would be considered a high risk.
2. Socially weighted risk mapping – adds an ethical layer to decision -making by prioritizing the safety of vulnerable participants. In practice, this means that AV could give in to a pedestrian even when the technical rules allow it to proceed.
3. Encoding of behavioral beliefs – predicts the way in which the AV actions will affect the overall situation of traffic. For example, it examines whether a rapid change of track could cause the drivers braking nearby or increase congestion.

To determine the safety performance of this cognitive coding scheme, the research team estimated the new framework using 2,000 reference traffic scenarios, and the results showed that the framework reduced the risk of global traffic by 26.3%.

Remarkably, these safety improvements came with better operational efficiency. In the aforementioned simulations, AVS equipped with this system finished driving tasks of 13.9% more quickly on average, demonstrating that ethical driving and performance can go hand in hand.

“By emulating human capacity for the treatment of holistic risks and moral reasoningWe allow AVS to behave more responsibly in ethically ambiguous situations, such as congestioned intersections or near schools, “said Professor Yang.

“Our framework is designed to be flexible and adaptable to comply with different social regulations and standards. For example, while some countries grant the protection of vulnerable road users, others put more emphasis on the effectiveness of the traffic flow.

“In addition, the legal interpretations of accident responsibility vary from one jurisdiction to another. Our system can adjust the weights, allowing AVS to drive as residents and make global deployment as possible.”

This pioneering study was carried out in collaboration with the University of Sciences and Technologies of Hong Kong (Guangzhou), Southeast University, Beijing Institute of Technology, Tsinghua University, Tongji University and the University of Washington.

As a next step, the research team develops a large -scale set of data representing various regional driving schemes and social expectations. They are also under discussion with potential employees to support future integration and test efforts.