一种基于神经形态计算的混沌时间序列预测装置
By using a chaotic time series prediction device based on neuromorphic computing, which employs mask signal modulation and weighted summation of pre-trained weight matrices, the problems of high computational difficulty and high power consumption in existing technologies are solved, and efficient chaotic time series prediction is achieved.
CN117494819BActive Publication Date: 2026-07-17XIDIAN UNIV
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- XIDIAN UNIV
- Filing Date
- 2023-11-16
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
Existing chaotic time series prediction models based on artificial neural networks suffer from problems such as high computational difficulty, high system power consumption, and high hardware implementation difficulty.
Method used
A chaotic time series prediction device based on neuromorphic computing is adopted, which includes an input unit, a reservoir unit and an output unit. Chaotic time series prediction is achieved by mask signal modulation, delayed signal mapping and weighted summation of pre-trained weight matrices.
Benefits of technology
This reduces the difficulty of hardware implementation and computational load, improves the efficiency and accuracy of the prediction device, and reduces electrical domain memory consumption and energy consumption.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN117494819B_ABST
Abstract
本发明提供的基于神经形态计算的混沌时间序列预测装置，包括：输入单元、储备池单元以及输出单元；输入单元，用于产生掩膜信号并获取当前时刻的混沌时间序列信号；将掩膜信号和混沌时间序列信号进行信号调制得到当前时刻的注入信号；储备池单元，用于发射激光以及将激光作用于自身的延迟回路从而产生延迟信号，并在延迟信号的作用下利用时分复用机制将当前时刻的注入信号进行映射，得到高维状态矩阵；输出单元，利用预训练的权重矩阵对高维状态矩阵进行加权求和，得到下一时刻的混沌时间序列预测信号。本发明中由于仅利用输出单元的预训练的权重矩阵，便可得到预测结果，减少了预测装置的训练难度和计算量，大大降低了预测装置的硬件实现难度。
Need to check novelty before this filing date? Find Prior Art