Scheduling Decision-Making Methods and Systems for Multiple AI Models in In-Vehicle Intelligent Cockpits

By collecting environmental perception data and AI model status information in the in-vehicle intelligent cockpit, performing extrapolation and correlation analysis, and generating power supply and resource allocation schemes, the problem of rigid resource allocation is solved, and the system's adaptability and operational stability in complex scenarios are improved.

CN121764688BActive Publication Date: 2026-05-26BEIJING DAFANG YUNTU TECH CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
BEIJING DAFANG YUNTU TECH CO LTD
Filing Date
2026-03-02
Publication Date
2026-05-26

AI Technical Summary

Technical Problem

Existing technologies lack in-depth analysis of the dynamic changes in the in-vehicle intelligent cockpit environment, resulting in rigid resource allocation and an inability to accurately predict the impact of hardware temperature rise on model performance. This leads to problems such as local overheating, model response delay, or surge in system energy consumption in complex or sudden scenarios.

Method used

By collecting environmental perception data and the operational status information of AI models, extrapolation analysis and time-series correlation analysis are performed to generate target correlation patterns. Performance prediction models are used to predict load changes and performance degradation, and power supply output parameters and resource allocation schemes are dynamically generated to achieve unified scheduling and decision-making for AI models in the cockpit.

Benefits of technology

It enables accurate identification of hardware temperature change trends and early identification of potential thermal risks, establishes an intrinsic link between the environment and model operation, improves the system's adaptability and operational robustness in complex and ever-changing scenarios, avoids resource allocation imbalance, and ensures the stable operation of key functions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121764688B_ABST
    Figure CN121764688B_ABST
Patent Text Reader

Abstract

This application provides a scheduling and decision-making method and system for multiple AI models in an in-vehicle intelligent cockpit, relating to the field of AI model technology. The method includes: collecting environmental perception data and operational status information of each AI model within the in-vehicle intelligent cockpit; performing extrapolation and analysis on the environmental perception data to obtain the temperature change trends of various hardware components within the in-vehicle intelligent cockpit; performing correlation analysis on the environmental perception data and operational status information to generate target correlation patterns; using a pre-built performance prediction model to predict the load change parameters, performance degradation parameters, and parameter fluctuation range of each AI model within different time periods, obtaining the operational prediction results of each AI model; and generating target power output parameters and target resource allocation schemes for the in-vehicle intelligent cockpit based on the results of the above steps, thereby enabling unified scheduling and decision-making for each AI model within the in-vehicle intelligent cockpit, improving the response efficiency and operational stability of the in-vehicle intelligent cockpit in complex scenarios.
Need to check novelty before this filing date? Find Prior Art