Time-Series Compression Circuit Using Parallel Polygonal Approximation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional time series data compression techniques using polygonal line approximation have limitations in throughput and power efficiency due to long calculation times and sequential processing, which hinder high-speed data transmission and power saving in IoT applications.
Innovation Solution
The time series data compressing apparatus employs a polygonal line approximating unit with outer product circuits that perform parallel calculations, reducing the number of division operations and increasing throughput, and a coding unit that applies variable-length coding based on optimal coding tables for high compression rates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional polygonal line approximation is used for time series data compression, then compression functionality is achieved, but throughput is low and power consumption is high due to long calculation times and sequential processing
Solution Approach 1:
The patent divides the calculation process into multiple independent segments that can be executed in parallel. Specifically, the slope calculation is segmented into multiple partial slope calculations (first partial slope, second partial slope, third partial slope) that are computed simultaneously using separate calculation units, thereby reducing total calculation time and increasing throughput.
Solution Approach 2:
The patent transitions from sequential single-dimensional processing to parallel multi-dimensional processing by introducing multiple calculation units operating simultaneously. Each calculation unit processes different aspects of the polygonal line approximation in parallel, effectively adding a temporal dimension to the processing architecture and achieving higher throughput.
2Loss of energy
If conventional polygonal line approximation with sequential processing is used, then compression is achieved, but power consumption increases due to long calculation times
Solution Approach 1:
By segmenting the calculation into parallel operations across multiple units, the total execution time is reduced, which directly lowers energy consumption. The patent implements this by having multiple calculation units perform slope calculations simultaneously rather than sequentially, reducing the active processing time and associated power usage.
Solution Approach 2:
The patent employs periodic action through its processing architecture where data is processed in discrete blocks or segments through the parallel calculation units. This periodic batch processing approach allows for efficient resource utilization and reduced power consumption compared to continuous sequential processing.
3Loss of information
If high compression rate is achieved through complex coding, then compression efficiency improves, but device complexity increases
Solution Approach 1:
The patent segments the compression system into distinct functional units: multiple calculation units for slope computation, a determination unit for vertex identification, and a coding unit for data compression. This modular segmentation allows each unit to be optimized independently, achieving high compression rates without proportionally increasing overall device complexity.
Solution Approach 2:
The patent implements universal calculation units that can perform multiple functions - calculating different types of slopes (first, second, third partial slopes) and handling various aspects of polygonal line approximation. This multi-functionality reduces the total number of dedicated circuits needed, thereby controlling device complexity while maintaining high compression efficiency.
Data Source
AI summary
A data compressing apparatus includes a polygonal line approximating circuit receiving first time series and outputs second time series by performing a polygonal line approximation process on the first time series. The polygonal line approximating circuit includes a first multiplier and a second multiplier performing multiplication having a first value calculated based on a difference between a time component of first data and a time component of second data in the time series data as input, and a third multiplier and a fourth multiplier performing multiplication having a second value calculated based on a difference between a sensor component of the first data and a sensor component of the second data as input.