Multipath LOS Detection for Accurate Wireless Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing localization technologies in multipath propagation environments face challenges in accurately determining the line-of-sight (LOS) path due to reflections from landmarks with unknown locations, leading to biased location estimates and computational intensity, especially in dynamic radio environments like cm and mm wave frequency bands.
Innovation Solution
A method and apparatus utilizing mean-field theory (MFT) for multipath reconstruction and multi-LOS detection, estimating approximate probability density functions for delay, phase, and amplitude of channel components, with flexible delay grid resolution, to improve accuracy and reliability of location estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If the strongest component in power-delay profile is selected as TOA, then the measurement is simpler, but the location estimate becomes biased when NLOS components are stronger
Solution Approach 1:
The patent applies dynamics by making the TOA selection adaptive rather than static. The system dynamically evaluates multiple candidate TOA values and their corresponding locations, then selects the most probable one based on statistical criteria. This resolves the contradiction by maintaining operational simplicity while improving accuracy through adaptive decision-making.
Solution Approach 2:
The patent changes the parameter selection criterion from simply 'strongest power' to a probabilistic evaluation considering multiple parameters including power levels, delay values, and statistical likelihood. This allows the system to distinguish LOS from NLOS components more accurately while maintaining a relatively simple measurement process.
2Measurement precision
If hypothesis testing or machine learning methodology is used for LOS detection, then the detection accuracy improves, but the computational complexity increases significantly
Solution Approach 1:
The patent applies partial action by evaluating only the most relevant candidate TOA values rather than exhaustively analyzing all possible hypotheses. It focuses computational resources on the most probable LOS and NLOS components, achieving good detection accuracy without the full computational burden of comprehensive machine learning approaches.
Solution Approach 2:
The patent uses lightweight statistical tests that are computationally inexpensive compared to full machine learning models. These simplified tests provide sufficient accuracy for LOS detection while being much less demanding in terms of computational resources and implementation complexity.
3Ease of operation
If features are extracted from noisy channel impulse response estimate, then the LOS detection can be performed, but the features are inherently noisy and exhibit misleading dependencies
Solution Approach 1:
The patent applies preliminary action by performing TOA candidate identification and evaluation before final LOS detection. It pre-processes the channel estimates to identify multiple candidate components, then systematically evaluates their likelihood as LOS versus NLOS. This staged approach reduces the impact of noise by distributing the detection task across multiple evaluation steps rather than relying on single noisy features.
Data Source
AI summary
There are provided apparatuses, methods and computer program products. In accordance with an embodiment, there is disclosed a method including receiving positioning reference signals from a positioning signal transmitter; forming a delay search space from the received positioning reference signals to obtain a plurality of channel taps; estimating a noise precision for the channel taps; estimating channel gains for the channel taps; and estimating a probability of line of sight signal for the channel taps.