Uplink Frequency-Bin Compression With Shared Exponent Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing number of antennas in telecommunication systems leads to a data rate problem through interfaces, and the use of virtualized evolved packet core further increases data transfer requirements, necessitating efficient data compression to enhance system performance.
Innovation Solution
The implementation of uplink data compression by transforming time-domain signals into frequency-domain representations, selecting groups of adjacent frequency bins, generating floating point number representations with a shared common exponent, and outputting mantissas and the common exponent, thereby reducing data transmission needs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the number of antennas is increased to improve network capacity and performance, then system performance and network capacity are improved, but data rate through interfaces increases causing data transmission bottlenecks
Solution Approach 1:
The patent extracts only the essential information from full signal samples by identifying and transmitting only the maximum magnitude sample and its position index within each group, rather than transmitting all signal samples. This extraction approach maintains the ability to reconstruct signal characteristics while dramatically reducing data transmission requirements.
Solution Approach 2:
The patent changes the representation parameters of signal data by transforming complete signal samples into a compressed format consisting of maximum magnitude values and their corresponding indices. This parameter transformation allows the system to represent the same information with fewer bits, reducing data rate while maintaining network capacity.
2Adaptability or versatility
If virtualized evolved packet core is implemented to improve network flexibility and resource utilization, then network flexibility is improved, but data transfer requirements increase
Solution Approach 1:
The patent extracts only the critical elements needed for signal processing - the maximum magnitude sample and its index - and transmits only these extracted components through the network. This reduces the data transfer burden on virtualized evolved packet core systems while maintaining the flexibility and resource utilization benefits of virtualization.
3Quantity of substance
If data compression is applied to reduce data transmission needs, then data transmission requirements are reduced, but compression complexity increases
Solution Approach 1:
The patent segments signal samples into groups and applies compression independently to each group by identifying the maximum magnitude sample and its index. This segmentation approach simplifies the compression process by breaking down complex signal data into manageable units that can be processed with simple comparison and indexing operations.
Solution Approach 2:
The compression algorithm uses the signal data itself to generate the compression parameters - the maximum magnitude sample and its index are directly derived from the signal samples without requiring external reference data or complex transformation functions. This self-service approach reduces compression complexity by leveraging the inherent properties of the signal data.
Data Source
Figure 1~2
Figure 3A~4A
Figure 4B~5
AI summary
There is provided a method for uplink data compression, the method comprising: transforming a signal from a time-domain presentation into a frequency-domain presentation, selecting a group of adjacent frequency bins from the frequency-domain presentation, generating a floating point number representation from values of the group of adjacent frequency bins, wherein the floating point number representation comprises a mantissa for each of the frequency bins' values of the group of adjacent frequency bins, and wherein the floating point number representation further comprises a common exponent shared between the mantissas, and outputting the mantissas of the floating point number representation, and the value of the common exponent shared between the mantissas of the floating point number representation.