[column parallax_bg=”disabled” parallax_bg_inertia=”-0.2″ extended=”” extended_padding=”1″ background_color=”” background_image=”” background_repeat=”” background_position=”” background_size=”auto” background_attachment=”” hide_bg_lowres=”” background_video=”” vertical_padding_top=”0″ vertical_padding_bottom=”0″ more_link=”” more_text=”” left_border=”transparent” class=”” id=”” title=”” title_type=”single” animation=”none” width=”1/1″ last=”true”] [column_1 width=”1/1″ last=”true” title=”” title_type=”single” animation=”none” implicit=”true”]

New IRF Taskforce on AI for Predictive Maintenance

Working with industry leaders, IRF has convened a new taskforce to explore how AI and computational advances can assist road agencies diagnose functional distresses to their pavements, structures and traffic assets.

To make informed decisions, necessary and relevant information needs to be available, and with the complexity and size of data sets, this task is time consuming, costly and at high risk of human error. Deep learning is a method of training a computer algorithm with large data sets that are populated with existing knowledge, such that the algorithm will continue to teach and correct itself. As new data sets are curated and recursively provided to the algorithm the confidence and knowledge of AI far exceeds the subjective and limited capacity which field experts can maintain.

The practical reality of using AI in this field are still emerging. Nevertheless, with the latest AI and computational advancements, it is already possible to automatically identify functional distresses such as potholes, cracks, edge breaks, bleeding, etc., by deep learning techniques from visual data such as photos and videos. Combined, they will provide the fingerprint of a section which can be used in estimating the future condition, likelihood of failure criteria, and maintenance that is required.


[/column_1] [/column]

Related Posts