ZDS 2023 Management of IoT TinyML Devices.pdf

1、#EMBEDDEDOSSUMMITManagement of IoT TinyML DevicesMieszko Mieruski,AVSystemMMierunskiTinyML IntroductionAnomaly Detection IoT device demoLwM2M overviewCombining LwM2M with TinyMLAgenda#EMBEDDEDOSSUMMITTinyML introductionWhat is TinyML?Key FeaturesLow LatencyPower EfficiencyPrivacy and securityAnomaly

2、 DetectionPredictive MaintenanceEnergy ManagementGesture RecognitionEnvironmental MonitoringHealth Monitoring and WearablesUse cases#EMBEDDEDOSSUMMITAnomaly DetectionIoT demoStandardOperation#EMBEDDEDOSSUMMITLwM2M OverviewLwM2M ArchitectureLwM2M Data Model#EMBEDDEDOSSUMMITLwM2M&TinyMLAccelerometer O

3、bjectML Model ObjectPattern Detector ObjectAnomaly Detector ObjectClassifier ObjectAnomaly Analyzer ObjectTinyML Objects-SummaryObject NameIDDescriptionPattern Detector33650This object is used to report the pattern detected by the ML-based classification algorithms and to count the number of times i

4、t has been detected.Anomaly Detector33651This object is used to report the anomaly detected by the ML-based algorithms and to count the number of times it has been detected.Classifier33652This object is used to report the results of the ML-based classification algorithm.Anomaly Analyzer33653This object is used to report the sensor data that are significantly different from the data in the training dataset.ML Model33654This object is used to report the meta information of the ML model used by the device.https:/ and referencesThank you for your attention!Q&A