WiGAN Radio Map Construction for Indoor Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing WiFi-based indoor localization systems face challenges in constructing accurate radio maps due to labor-intensive manual data collection and vulnerability to temporal and environmental changes, while automatic methods like crowdsourcing and machine learning face limitations in scalability and accuracy.
Innovation Solution
An automatic radio map construction and adaptation scheme using a Gaussian process regression (GPR) initialized Wasserstein generative adversarial network (GAN) to model irregular signal distributions, dividing environments into free and constrained spaces for accurate RSS estimation and fine-grained radio map generation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual RSS data collection and site survey approaches are adopted to construct the radio map, then the localization accuracy is improved, but the construction process becomes extremely labor-intensive and time-consuming
Solution Approach 1:
The system enables automatic radio map construction by having mobile devices self-collect RSS measurements during normal usage without requiring manual site surveys. The radio map is built automatically from crowd-sourced data as users move through the environment, eliminating the need for dedicated calibration teams.
Solution Approach 2:
The system performs preliminary RSS data collection in advance by gathering measurements from multiple mobile devices during their normal operations. This pre-collected data is then used to construct the radio map before actual localization tasks, allowing the system to be ready for immediate use without requiring on-demand manual surveys.
2Measurement precision
If manual RSS data collection and site survey approaches are adopted to construct the radio map, then the localization accuracy is improved, but the system becomes vulnerable to temporal and environmental dynamics
Solution Approach 1:
The radio map is transformed from a static manually-collected dataset into a dynamic structure that continuously adapts to environmental changes. The system automatically updates RSS measurements and radio map entries as new data becomes available from mobile devices, allowing the localization system to adapt to temporal and environmental dynamics such as moving occupants, changed furniture arrangements, or structural modifications.
Solution Approach 2:
The system implements feedback mechanisms where RSS measurements from mobile devices are continuously monitored and used to update and refine the radio map. This feedback loop allows the system to detect and adapt to environmental changes automatically, maintaining localization accuracy despite temporal and environmental dynamics.
3Productivity
If crowdsourcing method is used to automatically collect RSS measurements, then the radio map construction time is reduced, but the localization accuracy decreases due to location uncertainties
Solution Approach 1:
The system introduces GPS coordinates and sensor fusion data as intermediary elements to bridge the gap between crowd-sourced RSS measurements and accurate localization. By using GPS position information and other sensors (accelerometers, gyroscopes) as mediators, the system can associate RSS measurements with more accurate location estimates, reducing the impact of location uncertainties inherent in pure crowdsourcing approaches.
Solution Approach 2:
The system combines multiple data sources and measurement types to create a composite localization solution. By fusing GPS coordinates, accelerometer data, gyroscope information, and RSS measurements into a unified radio map structure, the system achieves both the efficiency of automatic construction and the accuracy needed for reliable localization, overcoming the limitations of any single data source.
4Ease of manufacture
If machine learning methods like GPR are applied to generate interpolated RSS values, then the manual data collection effort is reduced, but the scalability is poor and representational power is restricted
Solution Approach 1:
The system creates a universal radio map construction framework that can handle diverse indoor environments and configurations. By using a standardized data collection approach based on mobile device sensors and GPS that works across different locations and scenarios, the system achieves both ease of construction and scalability, unlike specialized machine learning methods that require environment-specific training and have limited applicability.
Data Source
AI summary
An automatic wireless fine-grained ratio map construction and adaptation system may include a Gaussian process regression (GPR) model constructed with real wireless received signal strength (RSS) measurements collected in a free space to provide coarse RSS estimation in a constrained space, and a generative adversarial network (GAN) to provide fine-grained RSS estimation in the constrained space by using an output of GPR as an input for a generator of GAN, modeling the irregular RSS distributions in complex indoor environments. The system may generate realistic RSS data in the constrained space that has not been manually site-surveyed.