{"paper":{"title":"Fully Autonomous Z-Score-Based TinyML Anomaly Detection on Resource-Constrained MCUs Using Power Side-Channel Data","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"A Z-score TinyML system trains and detects appliance anomalies entirely on microcontrollers using power side-channel data with perfect accuracy and minimal resources.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Abdulrahman Albaiz, Fathi Amsaad","submitted_at":"2026-03-28T20:15:50Z","abstract_excerpt":"This paper presents a fully autonomous Tiny Machine Learning (TinyML) Z-Score-based anomaly detection system deployed on a low-power microcontroller for real-time monitoring of appliance behavior using power side-channel data. Unlike existing Internet of Things (IoT) anomaly detection approaches that rely on offline training or cloud-assisted analytics, the proposed system performs both model training and inference directly on a resource-constrained microcontroller without external computation or connectivity. The system continuously samples current consumption, computes Root Mean Square (RMS)"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Results demonstrate perfect detection performance, with Precision and Recall of 1.00, inference latencies on the order of tens of microseconds, and a total memory footprint of approximately 3.3 KB SRAM and 63 KB Flash.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That power side-channel RMS values under controlled anomaly conditions in the 14-day mini-fridge dataset are representative of real-world anomalies and that Z-score thresholds derived from the training phase will generalize without overfitting or missing subtle deviations.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"A Z-score anomaly detector trained and inferred fully on an STM32 microcontroller using power side-channel RMS data achieves perfect precision and recall on a 14-day fridge dataset with low memory and latency.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A Z-score TinyML system trains and detects appliance anomalies entirely on microcontrollers using power side-channel data with perfect accuracy and minimal resources.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"f91ed98faf4b2034ccf5c6c19bc815f227d38cb96395c16a7df2dc0aa6a63e46"},"source":{"id":"2604.08581","kind":"arxiv","version":1},"verdict":{"id":"1e41ad32-a0c2-4f1a-bd24-05389cb2d413","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T22:05:30.422912Z","strongest_claim":"Results demonstrate perfect detection performance, with Precision and Recall of 1.00, inference latencies on the order of tens of microseconds, and a total memory footprint of approximately 3.3 KB SRAM and 63 KB Flash.","one_line_summary":"A Z-score anomaly detector trained and inferred fully on an STM32 microcontroller using power side-channel RMS data achieves perfect precision and recall on a 14-day fridge dataset with low memory and latency.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That power side-channel RMS values under controlled anomaly conditions in the 14-day mini-fridge dataset are representative of real-world anomalies and that Z-score thresholds derived from the training phase will generalize without overfitting or missing subtle deviations.","pith_extraction_headline":"A Z-score TinyML system trains and detects appliance anomalies entirely on microcontrollers using power side-channel data with perfect accuracy and minimal resources."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.08581/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":15,"sample":[{"doi":"","year":2023,"title":"A Comprehensive Survey on TinyML","work_id":"612c03a2-e5b9-4a1e-9970-d23b577d88f0","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2020,"title":"TinyML -Enabled Frugal Smart Objects: Challenges and Opportunities","work_id":"5bf738a1-3d32-497a-9a2d-dbdfa3cb01ae","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2003,"title":"Benchmarking TinyML Systems: Challenges and Direction","work_id":"922e79c4-b1db-4146-9f0d-6904f483ea1d","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Anomaly Detection in Smart Environments: A Comprehensive Survey","work_id":"49ab43d5-6541-4f12-8ced-6d5316474b79","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2011,"title":"Nonintrusive Appliance Load Monitoring: Review and Outlook","work_id":"c5e627f4-dc54-4021-bffe-82d46b35c360","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":15,"snapshot_sha256":"1d32ffb9c9d22e948bcc7aeaed598d0b66c783059f53c50701c12cf883c18851","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}