Probabilistic Storage Algorithm for Spectrum Measurements
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for storing and analyzing time-domain spectrum measurements in wireless technology face challenges due to the large volume of data required, which hinders efficient storage and analysis, especially in unlicensed frequency bands where wireless coexistence is necessary, and existing data compression techniques fail to effectively reduce storage volume while maintaining accurate channel utilization estimates.
Innovation Solution
The implementation of probabilistic efficient storage algorithms (PESA) that split the dynamic range of monitoring equipment into power bins, assign Gaussian components of a Gaussian mixture model to each bin, identify activity windows, and store the power averages and sample counts, allowing for significant reduction in storage volume while maintaining accurate channel utilization estimates through reconstruction of time-domain signals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If time-domain spectrum measurements are stored in full detail for accurate analysis, then measurement precision is improved, but storage volume increases substantially
Solution Approach 1:
The patent segments the time-domain signal into activity windows (where signal power exceeds threshold) and inactivity windows (where signal power is below threshold). Only activity windows are compressed and stored with reduced detail, while inactivity windows are represented by simple counts. This segmentation allows the system to maintain measurement precision for relevant signal portions while dramatically reducing storage volume by approximately 99%.
2Quantity of substance
If data compression is applied to reduce storage volume, then storage efficiency is improved, but measurement precision deteriorates
Solution Approach 1:
The patent changes the representation parameters from storing individual time-domain sample values to storing compressed activity window data structures containing: activity window start time, activity window duration, and power level statistics (mean and standard deviation). This parameter transformation enables 99% storage volume reduction while preserving sufficient information for accurate channel utilization estimation through probabilistic reconstruction.
3Measurement precision
If all time-domain samples are retained for analysis, then analysis accuracy is improved, but processing time increases
Solution Approach 1:
The patent extracts only the essential characteristics from the full time-domain signal for storage and analysis purposes. Specifically, it extracts activity window boundaries, durations, and power statistics, discarding the redundant individual sample values. This extraction process maintains spectrum analysis accuracy by preserving the key features needed for channel utilization calculation while reducing processing time through the smaller compressed data set.
Data Source
AI summary
Methods include compressing a plurality of time domain samples with a processor and memory by providing the plurality time domain samples and a plurality of power bins, identifying an activity window corresponding to a sequence of the time domain samples that are above a selected power threshold, determining a power average for the activity window, assigning the power average to one of the power bins having a range that includes the power average, and storing the assigned power bin and number of time domain samples of the activity window. Related decompression methods that can estimate a radio frequency power over time from the compressed power window data, as well as systems employing compression and/or decompression methods are also disclosed. Selected examples employ Gaussian mixture models and Bayesian responsibility functions.