FireLaneGuard

What it does

This design monitors fire lane occupation in real time and intervenes proactively via multi-modal sensing and graded warnings, solving the problems of high false alarm rate, slow response and high cost in traditional solutions.

Your inspiration

Why we decided to solve this problem Inspired by the urgent need to address the perennial occupation of fire lanes in old communities, coupled with the inefficiency and high cost of traditional monitoring, we were driven by the statistic that 80% of fire casualties are related to blocked lanes - a critical issue endangering public safety. Inspiration for the solution Inspired by industry pain points (60% manual inspection miss rate, 30% pure vision false alarms). Borrowed Dyson's "minimal tech solution" philosophy, using geomagnetic-vision fusion. Enabled by edge computing for low-latency response in weak networks.

How it works

The device works in 3 steps: A buried sensor detects metal objects (cars, e-bikes) and wakes up the camera. It decides locally in 0.5s to first warn with lights, then directional voice, finally alerting property and fire departments. It learns from mistakes monthly, works without WiFi, and costs 1/10 of traditional systems.

Design process

Initial idea Problem identified: 70% fire lane occupation in old communities, high false alarms and costs in traditional monitors. First plan: Pure vision AI camera, but night-time false alarms reached 35% (e.g., mistaking light reflections for vehicles). Prototype 1: Geomagnetic + vision fusion Tech used: Buried geomagnetic sensor + normal camera (cost: ¥300/point). Improvement: Solved 60% metal detection, but 40% missed non-metal obstacles (e.g., cartons). Feedback: False alarms reduced to 12%, but omnidirectional alarms caused noise complaints. Prototype 2: Graded warnings & edge computing Upgrades: 3-level warnings (vision→directional voice→property linkage), noise <50dB beyond 5m; Edge computing reduced response time from 10s to 0.5s. Test result: Occupation resolution rate increased to 85%, complaints down 70%.

How it is different

This design stands out with its multi-modal sensing fusion (geomagnetic + lightweight AI vision), edge computing for local decision-making, and three-level graded warnings, forming a low-cost yet highly reliable solution. Unlike competitors, it uses a 10-yuan geomagnetic sensor to address 73% of metal occupation detection with a mere 1.5% false alarm rate, achieves "community-specific" adaptation through edge-based evolutionary learning, and integrates directional silent warnings (noise <50dB beyond 5m) with a legal evidence chain. Delivering a 90% occupation resolution rate at 1/10 the cost of traditional solutions, it truly balances precision, affordability, and user experience.

Future plans

Next, we will optimize multi-modal algorithms, integrate smoke sensors, complete IP68 certification for extreme environments, expand pilots to 100 communities nationwide, establish industry standards with property associations, reduce costs to ¥150/point via mass production, and launch a "hardware+SaaS" model for annual fees; the future goal is to build an end-to-end "monitoring-warning-dispatching" solution, reduce occupation rates to <5% in partner communities by 2028, promote intelligent law enforcement standards, and adapt to overseas needs as a Chinese solution for global grassroots safety management.

Awards