Indoor Location Estimation Using Magnetic Field Maps and Neural Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing indoor location-based services face performance issues due to temporal instability and signal attenuation caused by indoor obstructions, leading to localization errors of up to 10 meters in large environments, and existing solutions require expensive equipment or infrastructure.
Innovation Solution
A method using a recurrent neural network (RNN) to estimate user location based on changes in the indoor magnetic field, generating a magnetic field map and learning data to create a location estimation model, which can be implemented on a portable device like a smartphone without the need for expensive hardware like access points or beacons.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If RF-based localization techniques (trilateration, fingerprinting, TDoA) are used, then indoor location-based services can be provided, but localization accuracy deteriorates to errors of 2-10 meters due to signal attenuation, reflection, and diffraction caused by indoor obstructions
Solution Approach 1:
The patent replaces RF-based localization systems with a magnetic field-based localization system. Instead of using radio waves that are susceptible to attenuation and reflection by indoor obstructions, the invention utilizes the Earth's magnetic field and local magnetic anomalies created by building structures. Magnetic field sensors (磁力传感器) detect magnetic field strength and direction, which remain stable and unaffected by typical indoor obstacles, thereby resolving the contradiction between providing localization services and suffering from signal degradation.
2Reliability
If expensive infrastructure equipment (APs, beacons, ultrasound, camera, LED, laser sensors) is deployed to improve localization performance, then localization accuracy improves, but system cost increases
Solution Approach 1:
The patent enables portable terminals (smartphones, tablets) to perform self-service localization using built-in magnetic field sensors. The system leverages the Earth's magnetic field and local magnetic features created by building structures, eliminating the need for expensive external infrastructure such as access points, beacons, or specialized sensors. The portable terminal independently collects magnetic field data, processes it through neural network algorithms, and determines location without requiring costly deployed infrastructure.
Solution Approach 2:
The invention replaces expensive, permanent infrastructure equipment with inexpensive, portable devices that users already possess. Instead of deploying costly access points, beacons, or specialized sensors throughout a building, the system uses the magnetic field sensors already present in smartphones and tablets. This approach transforms the localization system from a capital-intensive infrastructure project to an affordable consumer application.
3Speed
If pedestrian dead reckoning (PDR) using IMU is used to improve localization, then movement tracking is enhanced, but overall localization accuracy remains insufficient
Solution Approach 1:
The patent introduces magnetic field data as an intermediary that bridges the gap between PDR's movement tracking capability and accurate location determination. While PDR effectively tracks movement and direction changes, it accumulates errors over time and cannot determine absolute position. The invention uses magnetic field sensors to detect local magnetic anomalies created by building structures (steel beams, elevators, electrical equipment) as unique magnetic fingerprints. These magnetic field characteristics serve as an intermediary reference that corrects PDR's accumulated errors and provides accurate absolute location information.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enhances indoor localization accuracy and reduces costs by leveraging changes in the magnetic field to estimate user location with high precision, even in environments without expensive infrastructure, achieving localization errors of minimal 0 m and average 1.256 m.
Implementation Method 1
a magnetic field sensor part configured to measure magnetic field values as the magnetic field values change according to the movement of a user within the indoor space
Data Source
AI summary
A device that generates a location estimation model is provided. The location estimation model generator device includes a map generator part configured to generate a magnetic field map, which includes magnetic field values corresponding respectively to the coordinates of an indoor space; a data generator part configured to generate learning data by implementing the magnetic field map; and a learning part configured to generate a location estimation model by artificial neural network (ANN) learning implementing the learning data.