Probabilistic Traffic Database Generation for Multi-Layer Radio Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating a space-related traffic database in multi-layer radio networks are overly simplified and unrealistic due to the discrete allocation of radio cells, which does not accurately represent the complex layer interactions and coverage areas.
Innovation Solution
A method that subdivides a region into area elements and assigns probability-based allocation to each radio cell, using land usage classes and group-specific coefficients to minimize the difference between measured and predicted traffic, allowing for a more realistic and accurate traffic prediction across multiple layers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a discrete best-server model is used to allocate radio cells to area elements, then the method is simple and computationally efficient, but the model is overly simplified and unrealistic for multi-layer radio networks
Solution Approach 1:
The patent transforms the discrete allocation parameter into a continuous probabilistic parameter. Instead of assigning each area element to exactly one radio cell (discrete), the patent uses assignment probabilities that can take any value between 0 and 1, allowing area elements to be partially assigned to multiple radio cells simultaneously. This parameter change enables the model to capture the complex layer interactions in multi-layer radio networks while maintaining computational tractability through the use of continuous optimization methods.
2Reliability
If a continuous radio cell model with assignment probabilities is used, then the model is more realistic and accurate, but the complexity of generating the traffic database increases
Solution Approach 1:
The patent segments the radio network into distinct layers, where each layer represents a separate radio network with its own set of radio cells. By dividing the network into layers, the patent can apply the continuous assignment probability model to each layer independently and then combine the results. This segmentation reduces the overall complexity by breaking down the multi-layer problem into manageable single-layer subproblems that can be solved using standard optimization techniques.
Solution Approach 2:
The patent introduces a new dimension to the problem by adding the layer dimension. Instead of treating the radio network as a flat two-dimensional space, the patent extends it to three dimensions by incorporating the layer dimension. This allows the model to capture vertical layer interactions while maintaining the horizontal spatial relationships, providing a more comprehensive yet computationally manageable representation of multi-layer radio networks.
3Productivity
If discrete allocation of radio cells is used, then computational efficiency is maintained, but prediction accuracy deteriorates in multi-layer networks
Solution Approach 1:
The patent creates a simplified copy or representation of the complex multi-layer network interactions through the use of assignment probabilities. Instead of explicitly modeling all possible interactions between multiple layers and radio cells, the patent uses probabilistic assignment values that capture the essential characteristics of these interactions in a compressed form. This copying approach maintains computational efficiency while improving prediction accuracy by encoding complex relationships in a manageable format.
Data Source
AI summary
The present invention relates to a method for generating a space-related traffic database for a radio network which comprises a plurality of radio cells which can each be assigned to a group of radio cells, with a region to be mapped being subdivided into area elements by a grid and each of the area elements being allocated a respective assignment probability for each of the radio cells which supply the area element and a land-use class from amongst a finite group of land-use classes by generation being achieved by a minimization process in which the distance between the measured traffic of a radio cell and the predicted traffic of the radio cell is minimized for each radio cell, with the traffic, which is to be predicted, of a respective radio cell being set equal to the sum of area elements which are weighted by land use class-specific and group-specific coefficients, which area elements of the respective radio cell for a respective land-use class and for a respective group of radio cells are produced from the assignment probabilities of the area elements to the radio cells included in the process, and the coefficients are determined by the minimization process and assigned to the corresponding radio cell. The present invention also relates to a computer program and a corresponding computer-program product.