一种算力自适应调度的智能旁站监管系统和方法

By integrating hardware status perception and scheduling units into an integrated portable side-monitoring robot, and dynamically adjusting thread quotas, the problems of low integration of construction site monitoring equipment and difficulty in balancing computing power and battery life are solved, achieving concurrent timing synchronization of multiple algorithms and long battery life of the equipment.

CN122111683BActive Publication Date: 2026-07-17POWERCHINA ZHONGNAN ENG

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
POWERCHINA ZHONGNAN ENG
Filing Date
2026-04-21
Publication Date
2026-07-17

AI Technical Summary

Technical Problem

Existing construction site monitoring equipment suffers from low integration, a disconnect between computing power and hardware, lag when multiple algorithms run concurrently, and difficulty in balancing computing power and battery life.

Method used

An integrated portable side-station robot is adopted, which integrates a hardware status perception unit, a thread pool management unit, a latency monitoring unit, and a scheduling unit. It collects hardware status parameters in real time and dynamically adjusts thread quotas to ensure that the inference cycle of the algorithm model group is consistent and the computing power utilization rate is within the threshold.

Benefits of technology

It achieves real-time linkage between hardware status and computing power requirements, ensuring that the device is synchronized in the concurrent operation of multiple algorithms, extending the continuous working time in battery-powered mode, and solving the problems of low device integration and difficulty in balancing computing power and battery life.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122111683B_ABST
    Figure CN122111683B_ABST
Patent Text Reader

Abstract

本发明涉及一种算力自适应调度的智能旁站监管系统和方法,硬件状态感知单元实时采集电量、温度和充放电状态数据输入至模式控制单元,模式控制单元根据硬件状态匹配对应的运行模式并设定该模式下全局算力资源的占用率阈值上界,实现了硬件健康状态与算力需求的分级匹配;线程池管理单元为每个算法模型组分配独立的推理线程池以并行处理视频帧数据,时延监测单元实时获取各模型组单帧推理耗时并计算最大差值输入至调度单元,调度单元根据该最大差值与硬件状态参数动态调整各推理线程池的线程配额,在保证多算法并发推理时序同步的前提下,实现算力分配与硬件状态的实时联动,确保在不同工况下均能维持算力与功耗的动态平衡。
Need to check novelty before this filing date? Find Prior Art