Wireless CSI Fire Detection for Large-Area Location Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing fire detection methods, such as those using physical sensors and camera-based monitoring, face challenges in efficiently covering large and complex areas due to high installation costs, limited coverage, and susceptibility to environmental factors, making continuous and accurate fire detection difficult.
Innovation Solution
A system utilizing wireless signals, specifically channel state information (CSI) from WiFi signals, employs machine learning to infer fire occurrence and location through an artificial intelligence model, generating feature maps and tile maps for precise detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If physical sensors are used for fire detection, then detection accuracy is improved, but installation cost and device complexity increase due to dense sensor deployment requirements
Solution Approach 1:
The patent replaces physical sensor-based fire detection systems with a wireless signal-based detection system using channel state information (CSI) analysis. Instead of deploying dense networks of temperature, smoke, or flame sensors, the system uses existing wireless communication infrastructure to detect fire conditions by analyzing changes in wireless signal characteristics caused by environmental factors associated with fires.
Solution Approach 2:
The patent enables wireless communication signals to serve dual purposes: maintaining communication functionality while simultaneously performing fire detection. The channel state information extracted from wireless signals for communication purposes is repurposed to detect fire conditions, eliminating the need for dedicated detection sensors and infrastructure.
2Area of stationary object
If camera-based monitoring is used for wide area fire detection, then coverage area is improved, but reliability deteriorates due to false detections from environmental factors and limited operational hours
Solution Approach 1:
The patent changes the detection parameters from visual/optical domain (camera-based) to electromagnetic signal domain (wireless CSI analysis). By analyzing channel state information from wireless signals, the system detects fire-induced environmental changes such as temperature variations and atmospheric conditions, which are less susceptible to false detections from visual obstructions like clouds or fog that plague camera systems.
Solution Approach 2:
The patent enables continuous 24/7 fire detection operations using wireless signal analysis, unlike camera-based systems that require manual monitoring and are affected by lighting conditions. The wireless CSI-based detection operates continuously regardless of time of day, weather conditions, or visibility, providing uninterrupted monitoring coverage.
3Area of stationary object
If multiple camera towers are installed to cover complex terrain, then coverage completeness is improved, but installation cost and device complexity increase
Solution Approach 1:
The patent leverages existing wireless communication infrastructure (mobile phones, wireless devices, communication towers) to perform fire detection across complex terrains. Instead of installing dedicated camera towers throughout the terrain, the system utilizes the already-deployed wireless communication network, which naturally covers extensive areas including difficult-to-reach regions.
Solution Approach 2:
The patent introduces channel state information as an intermediary that bridges the gap between wireless communication signals and fire detection. The CSI parameters serve as a mediator that translates environmental conditions (temperature, atmospheric changes) into detectable signal characteristics, enabling fire detection without direct line-of-sight requirements that would necessitate multiple camera towers.
Data Source
AI summary
The present invention relates to a system and method for detecting fire occurrence, and a device therefor. A method for detecting occurrence of a fire according to one aspect of the present invention may include: periodically receiving a wireless signal from a second device; deriving an amplitude value and a phase value for the wireless signal from the wireless signal; detecting whether a fire has occurred and a location of the fire by performing inference of an artificial intelligence model using the amplitude value and the phase value as input data; and notifying an external system when it is detected that the fire has occurred at a specific location.