A SOC and position double prediction based space-time scheduling method for electric vehicles

By adopting a time-based scheduling method for electric vehicles based on dual prediction of SOC and location, and combining the dual prediction model with user preference mining, the problems of forward-looking and personalized electric vehicle charging scheduling are solved, thereby achieving grid load balancing and improved user satisfaction.

CN122134059AActive Publication Date: 2026-06-02HEFEI UNIV OF TECH
View PDF 4 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
HEFEI UNIV OF TECH
Filing Date
2026-04-30
Publication Date
2026-06-02

AI Technical Summary

Technical Problem

Existing electric vehicle charging scheduling methods lack foresight and cannot predict future vehicle movement trends and power consumption, which may lead to charging station saturation. Furthermore, the recommended strategies are too simplistic and cannot meet the personalized needs of users, which can easily cause excessive concentration of charging station load and grid impact.

Method used

We adopt a time-based scheduling method for electric vehicles based on dual prediction of SOC and location. By combining SOC prediction model, location prediction model, user preference mining and regional collaborative decision-making, we can achieve forward-looking and personalized charging scheduling through scene identification, dual prediction, regional charging volume prediction and dynamic saturation threshold calculation.

Benefits of technology

It improves the foresight and accuracy of charging scheduling, provides personalized services, optimizes the utilization of charging resources, ensures grid load balance, and enhances user experience and service efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122134059A_ABST
    Figure CN122134059A_ABST
Patent Text Reader

Abstract

This invention discloses a time-space scheduling method for electric vehicles based on dual prediction of SOC (State of Charge) and location, belonging to the technical field of commercial data processing and intelligent transportation systems. The method collects multi-source data and identifies vehicles in fixed or moving scenarios. Then, a dual prediction mechanism is executed: LSTM is used to predict the SOC change trend, SARIMA is used to predict the spatiotemporal location, and a clustering algorithm is used to predict regional charging volume to dynamically calculate the regional saturation threshold. Simultaneously, a two-level optimization mechanism based on a genetic algorithm (global and individual levels) is employed to deeply mine users' personalized preferences for distance and time. Finally, a two-layer collaborative framework of environment and decision-making unit layers is constructed, combining predicted data and preference parameters to calculate scheduling costs, and conflict resolution is performed under dynamic saturation threshold constraints, supplemented by a rolling optimization mechanism to correct decisions in real time. This invention achieves forward-looking, personalized, and globally balanced charging resource scheduling, effectively alleviating local congestion and significantly improving charging efficiency and user satisfaction.
Need to check novelty before this filing date? Find Prior Art