SensorGPT: Generative Pre-Trained Transformer Sensor Package

Offentlig prosjektsammendrag

The world is covered with 70 mil km of roads - connecting people and keeping our society moving. Building and maintaining these roads are a source of 400 MT of CO2 emissions annually, while they at the same time suffer from a huge refurbishment lag. In addition to the high emissions and environmental degradation caused by traditional road maintenance using bitumen and cement, the removal and disposal of old road material results in large amounts of waste. Furthermore, the transportation of material to and from the construction site often requires the closure of lanes or the complete shutdown of roads, leading to traffic disruptions and inconvenience for motorists. The goal of the “SensorGPT” innovation project is to develop and pilot a novel retrofit sensor package to allow for a next generation road crusher - able to leverage sensory data to optimize input and make road rehabilitation cheaper, faster and more sustainable. There are no machines available today that can carry out any kind of (near) real-time monitoring of environmental variables to adjust the mix design as part of their road crushing, rehabilitation or stabilization processes. The foreseen solution will take collection and utilization of road data to a whole new level through a combination of advanced moisture/temperature sensors, light detection and ranging (LiDAR) for mapping and surveying above ground structures, ground penetrating radar (GPR) for detection and location of underground objects, computer vision, and 5G connection.