Scientists at Moscow Polytech are developing an innovative software and algorithmic complex to reduce fuel consumption and CO emissions from semi-trailer tractors without interfering with the vehicle's design. This was reported in the press service of the university.
The development is based on a hybrid approach — artificial intelligence and a physical model.
First, the system calculates longitudinal movement and fuel consumption based on physical factors — aerodynamics, rolling resistance, slopes, road profile. The model automatically "adjusts" to a specific tractor based on field experiment data.
The system receives telemetry from the vehicle's on-board network (CAN bus), which allows it to be quickly configured for a specific vehicle. Based on the collected data, the system generates an optimal speed profile for the route, providing the driver with recommendations and explanations of the influencing factors.
First, the physical model calculates, and then the model trained on field experiment data carefully corrects it. If the conditions, for example, the mass, have changed, a quick recalibration takes only a few minutes.
Experts plan to test key components of the system on a "trip reproduction" stand using real telemetry logs. This will allow testing the effectiveness of the new development.
Earlier www1.ru reported that a new coating to increase engine efficiency and reduce fuel consumption was developed at SUSU.
