Beam Domain Indoor Localization via Deep Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing fingerprint indoor localization technologies face challenges with accuracy due to the weakening effect and multipath effect, particularly when using received signal strength indication (RSSI), and require high system complexity when using channel state information (CSI).
Innovation Solution
A beam domain based localization system that utilizes a wireless transceiver and a computer device to obtain a beam selection result, transforming it into a beam domain received power map for accurate mobile device location, employing a deep learning model with an autoencoder for feature extraction and training, allowing for accurate localization with low system complexity using a single wireless transceiver.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If RSSI information is used for indoor localization, then system complexity is low, but localization accuracy is easily affected by weakening effect and multipath effect
Solution Approach 1:
The patent transforms the localization approach by changing from using RSSI values directly to using beam selection results and beam domain received power maps. This parameter transformation enables the system to maintain low complexity while improving accuracy by exploiting spatial beamforming characteristics instead of relying on signal strength alone.
Solution Approach 2:
The patent introduces a new dimension for localization by using beam domain information (beam selection results and beam domain received power maps) instead of traditional signal strength measurements. This dimensional shift from scalar RSSI values to spatial beam domain representations allows the system to achieve better accuracy without proportionally increasing system complexity.
2Measurement precision
If CSI information is used for indoor localization, then localization accuracy is improved, but system complexity increases
Solution Approach 1:
The patent extracts only the necessary beam selection results and beam domain received power information from the full CSI data, rather than processing the entire CSI matrix. This extraction approach allows the system to achieve improved localization accuracy while avoiding the high complexity of full CSI processing by focusing only on the most relevant beam domain parameters.
3Adaptability or versatility
If multiple base stations or wireless access points are used for RSSI-based localization, then localization service coverage is improved, but system complexity increases
Solution Approach 1:
The patent enables each wireless transceiver to independently provide accurate localization services using beam domain information from its own beam selection results. This self-service capability allows single transceiver localization, eliminating the need for complex multi-transceiver coordination while maintaining service coverage through the inherent spatial information in beamforming.
Data Source
AI summary
The present disclosure provides a beam domain based localization system, and the beam domain based localization system includes a wireless transceiver and a computer device. The computer device electrically connected to the wireless transceiver, and the computer device configured for: obtaining a beam selection result associated with a mobile device through the wireless transceiver; locating the mobile device according to the beam selection result.