Cloud side-end cooperative computing resource scheduling method and system based on Internet of Vehicles

Abstract

The invention provides a cloud side end cooperative computing resource scheduling method and system based on the Internet of Vehicles. The method comprises the steps of collecting road image data, performing model training on a large-scale central cloud platform to obtain a cloud AI neural network model, and pushing the trained cloud AI neural network model to a user vehicle terminal; sensor data are uploaded to an edge cloud server through a laser radar, a camera and other sensors arranged on the roadside, data collected by a vehicle-mounted sensor are synchronously uploaded to the edge cloud server under the agreement of a vehicle user, and the edge cloud server can adopt the computing power of a city computing center for training. An edge AI neural network optimization model is obtained, a local edge AI neural network optimization model is added on the basis that an AI model is generated by a cloud platform, model optimization is carried out by adopting a method of combining edge federated learning with neural network model parameter fine tuning, and the optimized algorithm model is preferentially pushed to a vehicle with a local registration place.

Images