Wireless Link Interference Identification Using Weighted Time Intervals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In densely populated wireless local area networks (WLANs), existing frequency channel allocation schemes fail to effectively reduce interference between access points, leading to degraded user experience due to the correlation between interference and parameters like background noise and Received Signal Strength Indication (RSSI) not being well understood or deterministic.
Innovation Solution
A computer-implemented method identifies interfering wireless links by generating data on performance loss and medium usage metrics, using weighted time intervals and regression analysis to determine the hindrance caused by each link, employing LASSO regression to select relevant inputs and improve prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If Auto Channel Selection (ACS) algorithms are used to allocate frequency channels based on background noise, beacon messages, and RSSI, then frequency channel allocation is automated, but interference between access points is not significantly reduced because the correlation between interference and these parameters is not well known
Solution Approach 1:
The system implements feedback by continuously monitoring actual interference levels and using this information to identify bad neighbor access points. The interference measurement feedback loop enables the system to adaptively adjust channel allocation based on real-world performance rather than relying solely on predetermined parameters like RSSI and beacon messages.
Solution Approach 2:
The patent replaces traditional mechanical/statistical correlation methods with machine learning algorithms that can discover complex, non-linear relationships between various parameters and interference. Instead of relying on known deterministic correlations, the system uses training data to learn optimal channel allocation strategies that account for the actual deterministic nature of interference patterns in dense environments.
2Object-affected harmful factors
If non-overlapping frequency channels are allocated to neighboring access points, then interference is reduced when possible, but this solution cannot be implemented in densely populated environments where an access point has more neighboring access points than available frequency channels
Solution Approach 1:
The system dynamically adapts channel allocation based on the actual number of neighboring access points and available channels. Rather than relying on static non-overlapping channel assignment, the system uses machine learning models that can optimize channel selection in real-time, adjusting to dense environments where the number of neighbors exceeds available non-overlapping channels by identifying which specific neighbors cause the most interference.
Solution Approach 2:
The patent changes the approach from fixed channel allocation based on simple neighbor counts to dynamic parameter optimization using machine learning. The system evaluates multiple parameters including actual interference measurements, spatial relationships, and traffic patterns to determine optimal channel assignments, allowing adaptation to dense environments where traditional parameter-based methods fail.
3Stability of the object's composition
If traditional frequency channel allocation schemes are used in densely populated WLANs, then existing infrastructure can be maintained, but user experience degrades due to interference that cannot be effectively reduced
Solution Approach 1:
The system performs preliminary training and analysis to build machine learning models that predict interference patterns before actual channel allocation decisions are made. By pre-processing data and identifying bad neighbor relationships in advance, the system can make informed channel allocation decisions that prevent interference issues rather than reacting to them, thereby maintaining infrastructure stability while improving user experience.
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
The present disclosure relates to a computer implemented method for identifying, among a set of wireless links established between at least two wireless end-points, at least a first wireless link causing interferences on at least a second wireless link belonging to said set of wireless links, said method comprising: - Generating at least a set of data enabling the identification of the first wireless link by allocating measurements of an amount of performance loss of a wireless link due to interferences, and measurements of a use of a wireless medium by either said first wireless link or said second wireless link, to a given time interval of a set of time intervals in a weighted manner based on a time overlap between a time interval during which said measurements are performed and at least a given time interval of said set of time intervals, - identifying, based on said at least set of data, at least said first wireless link within said set of wireless links.