Wireless Positioning Feature Library Construction Using Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current wireless positioning technologies using rasterization for constructing positioning feature libraries suffer from location data indicating area ranges rather than points, leading to positioning track distortion and reduced accuracy due to noise and environmental changes.
Innovation Solution
A method that performs clustering and merging of training data based on similarity within a preset range to obtain sample data, which better reflects real locations and tracks, thereby improving positioning accuracy and resisting noise and GPS positioning errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If rasterization policy is used to process positioning data, then noise impact is resisted and data stability is improved, but location data indicates area ranges rather than point data causing positioning track distortion
Solution Approach 1:
The patent segments the positioning data processing into two distinct phases: data collection phase using rasterization for stability, and data storage/retrieval phase using precise point coordinates for accuracy. The sample data set stores original point coordinates rather than rasterized area data, allowing precise positioning without track distortion while maintaining the noise resistance benefits of rasterization during collection.
Solution Approach 2:
The patent merges the advantages of both rasterization and point-based positioning by combining them in different phases. Rasterization is applied during data collection to resist noise, while point coordinates are preserved in the sample data set for accurate positioning. This merging allows the system to achieve both data stability and positioning precision simultaneously.
2Ease of manufacture
If rasterization is applied to location data, then data processing becomes simpler and more robust to environmental changes, but the moving track of UE becomes distorted and less accurate
Solution Approach 1:
The patent segments the data representation into rasterized form for processing simplicity and point coordinates for track accuracy. The sample data set preserves the original point coordinates separately from the rasterized data, enabling accurate track representation while maintaining simple raster-based processing for noise resistance.
Solution Approach 2:
The patent introduces an intermediary mechanism where the sample data set acts as a bridge between rasterized data and precise positioning. The sample data stores point coordinates that mediate between the simplified raster representation and the requirement for accurate track visualization, allowing both simplicity and precision to coexist.
3Measurement precision
If clustering and merging is performed on training data based on similarity, then positioning accuracy is improved and track distortion is reduced, but data processing complexity increases
Solution Approach 1:
The patent performs clustering and merging operations during the offline training phase before actual positioning operations. By pre-processing the training data to create a sample data set with clustered samples, the system reduces positioning accuracy requirements during real-time operation. This preliminary action transfers processing complexity from the time-critical positioning phase to the non-time-critical training phase.
Solution Approach 2:
The patent implements dynamic sampling where the number and distribution of samples in the sample data set are adaptively determined based on the characteristics of the training data. The clustering process dynamically identifies regions requiring higher sampling density, optimizing the balance between positioning accuracy and processing complexity based on actual data distribution patterns.
Data Source
AI summary
A method for constructing a wireless positioning feature library includes: obtaining a wireless positioning training data set, where the wireless training data set includes a plurality of pieces of training data; performing clustering and merging on the plurality of pieces of training data one by one; and determining wireless positioning feature data based on at least one piece of sample data obtained after the merging.