Optical Signal State Estimation Using Time-Series Constellation Histograms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing optical signal processing apparatuses fail to accurately account for changes in signal points over time, leading to difficulties in understanding the state of optical signals with high precision.
Innovation Solution
An optical signal state estimation apparatus that includes a processor to acquire a constellation of an optical signal, generate time series data by counting signal points in grids, and estimate the signal state using a learned model trained on known state data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional optical signal processing methods are used that divide symbol regions into divided regions and calculate phase noise based on aggregate signal points, then the processing method is simple, but the accuracy of understanding optical signal state deteriorates because changes in signal points over time are not considered
Solution Approach 1:
The constellation plane is divided into multiple grid regions, and signal points are counted separately in each region to generate histogram information. This segmentation allows the system to capture temporal changes in signal point distribution across different regions, thereby improving estimation accuracy while maintaining manageable processing complexity through systematic regional analysis
Solution Approach 2:
The patent introduces time-series histogram data that dynamically tracks the distribution of signal points across constellation grids over time. By using learned models trained on this temporal data, the system adapts to changing signal conditions and captures dynamic variations in optical signal state, significantly improving measurement precision compared to static conventional methods
2Measurement precision
If time series histogram data with learned models is used to estimate optical signal state, then the estimation accuracy improves by considering changes over time, but the device complexity and processing requirements increase
Solution Approach 1:
The learned model automatically processes the time-series histogram data to estimate optical signal state without requiring manual intervention for each measurement. The model self-adjusts based on training data and autonomously performs estimation by analyzing temporal patterns in the histogram information, reducing the need for complex automated control systems while maintaining high precision
Solution Approach 2:
The system performs preliminary training of the learned model using known state optical signal data before actual estimation. This preliminary action prepares the model to automatically handle subsequent estimations with high accuracy, reducing the computational burden during real-time operation and balancing automation requirements with processing efficiency
Data Source
AI summary
At least one processor included in an optical signal state estimation apparatus carries out: an acquisition process for acquiring a constellation of an optical signal; a generation process for generating time series data in which pieces of histogram information are arranged in time series, the histogram information being obtained by counting the number of signal points of the constellation, the signal points being included in each of grids obtained by dividing an in-phase component direction and a quadrature component direction into a specific number of rows and the specific number of columns, respectively; and an estimation process for estimating a state of the optical signal by inputting the time series data to a learned model.