Drift-Aware Multi-Model Anomaly Detection system for multivariate industrial sensor data using Autoencoders and Correlation-Based Detection. The system detects long-term sensor drift in the UCI Gas Sensor Array Drift Dataset by combining temporal reconstruction errors with cross-sensor behavioral analysis.
Ultra-low-latency, drift-resilient bio-inspired olfactory engine in #![no_std] Rust. Zero-heap, 1.8µs inference, Modbus RTU & LoRaWAN telemetry, verified on 13,910 physical sensor drift measurements.
Multiclass gas classification with Logistic Regression from scratch, including temporal evaluation and distribution shift analysis.
Protocol-aware raw-signal benchmarking for electronic-nose drift and transfer
To associate your repository with the sensor-drift topic, visit your repo's landing page and select "manage topics."