System for realizing automatic iteration of prediction model based on data driving

Abstract

The embodiment of the invention discloses a system for realizing automatic iteration of a prediction model based on data driving, the system comprises a processor and a cloud end, the processor labels behavior data of each road participant based on the observed motion track of each road participant to obtain corresponding labeling information, automatic labeling is realized. Then, the screened first behavior data and corresponding labeling information are sent to a cloud end, the cloud end extracts a training sample to train an initial network model to obtain a target network model when determining that a preset automatic triggering condition is met, automatic triggering model training is achieved, then evaluation is automatically triggered, and when an evaluation result meets a model updating requirement, the target network model is sent to the processor, automatic deployment of the target network model is achieved, and the processor automatically updates the prediction model to the target network model after receiving the target network model, so that no manual participation is needed for data annotation, model training, model evaluation or model updating, and the automation degree is high.

Description

technical field [0001] The invention relates to the technical field of intelligent driving, in particular to a data-driven system for realizing automatic iteration of a prediction model. Background technique [0002] In the autonomous driving scenario, predicting the trajectory of the road participants in advance is beneficial to their own safe driving. For prediction, as the complexity of the road scene continues to increase, the prediction model needs to be updated from time to time. [0003] The current update method of the prediction model is: the developer performs offline labeling on a large number of road participants' behavior data to obtain labeling information, and then the developer performs model training and evaluation based on the road participant's behavior data and corresponding labeling information to obtain new predictions model, and then developers update the new predictive model to the unmanned vehicle through network transmission or hard disk connection...

AI Technical Summary

Problems solved by technology

[0004] It can be seen that the update method of the above prediction model requires manual data labeling and manual triggering of model training and model update, which is highly dependent on people, resulting in high labor costs and low automation

Images