VibraAI_Core

VibraAI_Core — Offline AI Engine for Telematics

VibraAI_Core is a small offline AI engine written in pure C. It runs on typical ARM Cortex-M MCUs with sub-millisecond inference and requires only a few kilobytes of memory.

Engine overview

Every 2-second window, the engine processes:

The engine outputs:

Footprint

Flash usage: 20 to 30 KB
RAM usage: 5 to 8 KB
Inference latency: less than 1 millisecond
Dependencies: none (pure ANSI C)

Integration (C API)

```c int vibra_init(VibraContext *ctx, const VibraConfig *cfg);

int vibra_infer_window( VibraContext *ctx, const VibraAccelSample *samples, float avg_speed_mps, long long timestamp_ms, VibraMetrics *out_metrics ); Contact If you want an evaluation SDK or integration support, you can contact:

morgan5962@live.fr