New technology improves self-driving cars during sensor failures
Imagine a self-driving car suddenly losing one of its key sensors. What happens then? Tiago Cortinhal, who has completed his PhD in signals and signal engineering with a focus on sensor recovery at Halmstad University, has dedicated his research to this issue. His findings could change the way we view autonomous vehicles and their safety.
Tiago Cortinhal’s research aims to make self-driving cars more reliable and secure, even when unexpected problems arise.
“My thesis focuses on how to recover missing sensor data to prevent the car from stopping”, says Tiago Cortinhal.
This work could be key to preparing self-driving cars for real-world traffic situations. His research has led to two surprising discoveries. Firstly, he has demonstrated that synthetic data generated by AI models can be invaluable for training and improving self-driving systems. Secondly, he has found that semantic segmentation, where each pixel in an image is classified to identify objects like roads and pedestrians, can serve as an effective intermediary in the training process for these complex networks.
These findings highlight the importance of preparing autonomous vehicles for real scenarios, including potential sensor failures. Tiago Cortinhal believes we can make this technology more robust and reliable in everyday situations by addressing these challenges.
Text and video: Anna-Frida Agardson
More information
Research at the School of Information Technology
På svenska
Ny teknik förbättrar självkörande bilar vid sensorfel
External link.
