Researchers use machine-learning human emotions to ‘drive’ self-driving cars

According to foreign media reports, Florida Atlantic University Engineering and Computer Science Dr. Mehrdad Nojoumian developed a new technology for autonomous driving systems “Adaptive Mood Control in Semi or Fully Autonomous Vehicles” (Adaptive Mood Control in Semi or Fully Autonomous Vehicles), which can be based on Machine Learning Human Emotions Respond to Human Emotions.

With adaptive emotion control technology, the experience of autonomous vehicles interacting with humans can be more convenient, enjoyable, and trustworthy. In addition, the technology could be used in a multitude of autonomous systems, including self-driving cars, self-driving military vehicles, self-driving planes or helicopters, and even social robots.

(Image credit: Florida Atlantic University)

Dr Mehrdad Nojoumian said: “This technology is unique in that, in a collaborative driving environment, operating modes and parameters related to perceived emotion can be exchanged with adjacent vehicles, enabling adaptive emotion control modules in semi- and fully autonomous vehicles. The goal. Human-AI/autonomous interactions have always been the focus of attention at both academia and industry level. More precisely, trust between humans and AI/autonomous technologies is critical in this area and will directly affect these The social acceptance of modern technology."

The technology uses non-invasive sensory solutions in semi-autonomous or fully autonomous vehicles to sense driver and occupant emotions. In addition, the technology can collect information based on facial expressions, sensors in the handle/seat, and thermal cameras in other surveillance equipment. The Adaptive Emotion Control System incorporates real-time machine learning mechanisms that continuously identify driver and occupant emotions over time. The results are then sent to the self-driving car’s software system, allowing the vehicle to respond to the perceived emotion by selecting the appropriate operating mode, such as normal, cautious, or alert driving modes.

Dr Stella Batalama, Dean of the School of Engineering and Computer Science, said: “One of the main problems with fully autonomous or semi-autonomous vehicle technology is the inability to accurately predict the behavior of other autonomous and human-driven vehicles, which is critical for navigating the road correctly. Driving a car is critical. By using machine learning algorithms to learn the behavioral patterns of vehicle drivers and occupants, Professor Nojoumian addresses the above problem.”

Source: Gasgoo

Author: Liu Liting

The Links:   KS621K40A41   AT080TN52-V5