FleetEasy represents a paradigm shift in telematics from reactive GPS tracking to proactive Edge-AI Evidence Capture. By utilizing a high-frequency Bayesian sensor-fusion model and on-device model, the platform eliminates false alarm fatigue while ensuring a 95 percent capture rate for legitimate road incidents.
Traditional dashcams rely on rigid G-force thresholds (e.g., Trigger if > 3G). These fail on rough roads or during low-speed impacts. FleetEasy employs a Dynamic Bayesian Update model:
• Likelihood Modeling: The system maintains two simultaneous probability curves—one for normal driving (centered at 1.0G) and one for accidents (centered at 8.0G).
• Risk-Weighted Penalty: We implement a durationFactor (up to 1000x) that scales the accident probability based on impact duration. A 50ms structural impact is mathematically prioritized over a 5ms road-noise spike.
• Zero-Drift Gating: A 1.5G logic gate combined with a natural decay algorithm ensures that the system forgets minor bumps, maintaining a clean slate for the duration of the journey.
FleetEasy operates an intelligent 10-second rotating WebM buffer.
• Pre-Event Context: Upon detection, the system locks the current 10-second buffer, capturing the critical seconds preceding the impact—data that is often lost in traditional systems.
• On-Device Inference: Utilizing a optimized MediaPipe/EfficientDet-Lite model, the system performs sub-second object detection on high-resolution snapshots. This allows for immediate verification of Car, Pedestrian, or Obstacle even in zero-bandwidth environments.
The FleetEasy architecture is built for scale across diverse mobile environments:
• Multi-Plugin Native Core: With 17 sub-projects (including custom-patched Filesystem, Share, and MLKit plugins), the app manages hardware lifecycles with absolute precision, preventing battery drain and sensor zombies.
• Persistent Native Integrity: Critical modifications to the Android SharePlugin and build systems are stabilized via patch-package, ensuring that the enterprise-grade multi-file sharing logic remains consistent across updates.
• Fleet Group Admin : In addition to manager side dashboard accessible from native app and through web interface, the backend utilizes a python-based-FastAPI&Django administrative layer for sophisticated group management, group profiling, and scheduled daily notification, providing managers with a single pane of glass view of their fleet's safety.
• Manager Portal: Direct relay to your fleet manager for instant support during road emergencies. Reduces incident reporting time from hours to seconds. Fleet managers receive a comprehensive evidence bundle immediately via the unified socket relay.
In the event of network failure, FleetEasy acts as a Local Black Box.
• Atomic Bundling: Every incident generates an Evidence Bundle (JSON telemetry + JPEG AI Snapshot + WebM Video).
• Local Persistence: All data is written to the platform-specific Cache directory before being offered for sharing. This ensures that the driver holds the physical evidence locally, regardless of server availability. Conclusion: Reducing the Total Cost of Ownership (TCO) By providing Context + Telemetry + Video in a single sub-second package, FleetEasy reduces insurance claim cycles from weeks to minutes. For fleet owners, this translates to lower premiums, faster exoneration of innocent drivers, and a measurable increase in operational safety.
Reducing the Total Cost of Ownership (TCO) By providing Context + Telemetry + Video in a single sub-second package, FleetEasy reduces insurance claim cycles from weeks to minutes. For fleet owners, this translates to lower premiums, faster exoneration of innocent drivers, and a measurable increase in operational safety.