Researchers at Oxford University’s Department of Computer Science, in collaboration with colleagues from Bogazici University, Turkey, have developed a new artificial intelligence (AI) system to enable autonomous vehicles (AVs) to achieve safer and more reliable mobility, especially under adverse weather and driving scenarios that It was rejected using GPS. The results were published today in The intelligence of nature’s machine.
“The difficulty for assistive vehicles in achieving accurate positioning during a challenge is the difficulty for assistive vehicles to achieve accurate positioning during the challenge,” said Yasin Elmaleoglu, who completed the research as part of his Ph.D. in the Department of Computer Science. bad weather This is one of the main reasons why these experiments have been limited to relatively small trials thus far. For example, weather, such as rain or snow, may cause an AV to detect itself in the wrong lane before turning, or to stop late at an intersection due to inaccurate positioning. “
To get around this problem, Malioglu and his colleagues developed a self-supervised novel deep learning model To estimate ego motion, a critical component of the vehicle’s driving system that estimates the position of the vehicle’s motion relative to objects observed from the vehicle itself. The model combined detailed information from optical sensors (which can be disrupted by adverse conditions) with data from weather immune sources (such as radar), so that the benefits of each can be used under different weather conditions.
The model was trained using several publicly available AV datasets that included data from multiple sensors such as cameras, lidar and radar under various settings, including changing light/darkness levels and precipitation. These were used to create algorithms to reconstruct scene geometry and calculate vehicle position from new data. Under different test situations, the researchers demonstrated that the model demonstrated robust performance in all weather conditions, including rain, fog and snow conditions, as well as day and night.
The team anticipates that this work will bring autonomous vehicles one step closer to safe, seamless, all-weather autonomous driving, and ultimately wider use within communities.
Professor Nikki Trigoni, from the Department of Computer Science at the University of Oxford, who co-led the study, said: “Accurate positioning capability provides the basis for many essential functions of assistive vehicles such as motion planning, prediction, situational awareness and collision avoidance. This study provides an exciting complementary solution to a suite of software sound and image to realize this potential.”
Professor Andrew Markham (Department of Computer Science, University of Oxford), who is also co-author of the study, added, “Estimating the precise location of self-driving vehicles is a milestone in achieving reliable autonomous driving under challenging conditions. This study effectively exploits the complementary aspects of autonomous driving devices. Various sensors to help vehicles navigate challenging everyday scenarios.”
Yasin Elmaleoglu, Powerful positioning based on deep learning for all-weather autonomous driving, The intelligence of nature’s machine (2022). DOI: 10.1038 / s4256-022-00520-5. www.nature.com/articles/s4256-022-00520-5
the quote: New AI enables autonomous vehicles to adapt to challenging weather conditions (2022, September 8) Retrieved September 8, 2022 from https://techxplore.com/news/2022-09-ai-enables-autonomous-vehicles-weather.html
This document is subject to copyright. Notwithstanding any fair dealing for the purpose of private study or research, no part may be reproduced without written permission. The content is provided for informational purposes only.