Industries
A hub-side RF physical-unclonable authenticator that recognizes transmitters from manufacturing “grain” in their radio chains — without putting keys, PUFs, or preambles on the constrained IoT node.
A configuration studio for embodied IoT agent fleets that captures variability in sensors/actuators/behaviors, learns from evaluative feedback, and reconfigures agent groups when environmental…
A streaming fall-detection service for retirement homes and rehab clinics that scores wearable accelerometer feeds in real time and notifies care staff when irregular activity patterns indicate a…
Delivers beacon-accurate, personalized airport journeys—wayfinding, flight alerts, offers, parking pay—so hubs absorb demand spikes (Olympics-scale) while unlocking partner revenue on a shared…
A human-in-the-loop active learning IDS for wireless IoT that queries analysts only for the most informative unlabeled flows, so teams reach high detection accuracy without labeling oceans of traffic.
A deployment toolchain that packs large CNNs into local quantization regions so resource-constrained IoT controllers run vision and speech models on-device at 8-bit down to 1-bit without shipping…
A network-side deep-autoencoder sentinel that learns each IoT device’s benign traffic snapshots and flags Mirai/BASHLITE-class attack execution in under a second, so enterprises can cut bots off…
A collaborative inference orchestrator that shards live vision DNNs across the cameras and Pi-class hubs already on a LAN, so homes and venues get real-time recognition without shipping video to the…
A low-cost smart-city noise classifier that labels urban sounds (jackhammer, gunshot, street music, horns) on Raspberry Pi–class nodes using MFCC + SVM/KNN — going beyond dB maps to actionable noise…
A software-defined IoT intrusion platform that selects flow features with a swarm (Bat + differential mutation) then classifies attacks with a weight-adapted Random Forest under the SDN controller’s…
Operates an edge information system for intelligent IoV—caching, compute offload, and edge AI—so vehicles meet sub-100 ms perception, HD mapping, and localization needs without exhausting onboard…
A predictive maintenance service for industrial IoT vehicle fleets that scores repair probability from telematics and garage history using hierarchical modified fuzzy SVMs, so operators schedule…
A smart-building indoor localization service that learns BLE fingerprint policies from sparse user-confirmed labels plus abundant unlabeled RSSI, so campuses stop surveying every square meter to…
A lightweight motion-anticipation service for V2X and mobile robots that keeps predicting object trajectories when IoT broadcasts arrive late, drop, or shift, so mission-critical agents do not go…
An online bandit that picks the least-jammed, lowest-security-risk edge server for each IoT offload — with optional peer side-observations — so devices keep computing under jamming without burning…
A battery-first edge vision service that runs deep multiple-object tracking on Jetson-class nodes from the live camera frame only — no cloud round-trip — for outdoor and mobile IoT perception.
Helps enterprises decide cloud vs edge for AI/IoT use cases—with latency/bandwidth/autonomy checklists, adoption-stage roadmaps, and dataset readiness—so real-time OT decisions are not stranded…
Predicts energy, latency, and accuracy trade-offs for image-classification workloads on constrained IoT edge boards so operators pick resolution, algorithm, and phase before battery or harvest…
An activity-embedding service that turns smart-home and smart-office IoT usage logs into device “kinship” vectors, so operators can identify unknown gadgets and recommend true functional…
A human-in-the-loop IoT agent configuration studio that models body/behavior/architecture variability, generates customized agent fleets, and uses feedback-evaluative neuroevolution so municipal and…
An enterprise IoT device-type whitelist that classifies what just joined the network from traffic behavior alone, so security teams can isolate unauthorized gadgets before BYOIoT becomes a…
An edge-first anomaly detection service for wireless sensor fleets that scores reconstructions on the node itself while the cloud retrains compact autoencoders on a configurable schedule.
Runs an IIoT “digital services factory” that takes connected-product ideas from ideation through incubation, fast prototyping, industrialization, and operations with value KPIs—cutting…
A multi-stage cognitive spectrum coordinator that learns which channels favor which IoT devices, predicts how long primary users stay off, and skips sensing for that window — so battery nodes get…
A multi-camera license-plate recognition network that turns every roadside and facility camera into one searchable “universal eye,” so agencies can register a plate once and get location-time alerts…
A phone-side context engine that predicts user context in ~11ms and actuates linked IoT devices even when the cloud is unreachable, while periodically syncing heavier model parameters from a server…
A collaborative learning coordinator for IoT fleets that trains deep models on Gaussian-projected samples so device owners improve shared intelligence without handing the coordinator raw sensor data.
A real-time ECG biometric gate that verifies a living person in about one second from a three-second finger-plate capture, so facilities and clinical systems stop trusting forgeable badges and…
A blockchain-style consensus layer that merges on-device anomaly models only after peer self-attestation, so large fleets of identical industrial IoT nodes keep learning without letting early…
Rates and orchestrates edge-intelligence applications across a six-level cloud–edge–device ladder for DNN training and inference so teams cut WAN cost and privacy leakage without pretending every…
An on-device IoT malware filter that turns executables into grayscale “binary images” and scores them with a shallow CNN, so cameras and DVRs can catch Mirai/Gafgyt-class DDoS bots before they join…
A black-box adversarial testing service that stress-gauges learning-based Android malware detectors for IoT devices by generating feature-preserving adversarial APK samples with near-complete attack…
A model-distillation workbench that siphons a large, multi-class CNN into a mission-specific core that recognizes only the classes an IoT application cares about — accurate enough for the mission,…
A governance workspace for Cyber-Physical Social Machines that inventories IoT-augmented human–platform systems (sensors and actuators in the loop), runs contextual-integrity and security risk…
An analytics placement control plane that assigns each IoT deep-learning workload to device, fog, or cloud based on stream urgency and data gravity, so operators stop treating every sensor feed as a…
A brain-inspired trust engine that scores neuroscience IoT nodes on both behavior and data fidelity, so cloud tele-neuro pipelines stop treating every EEG/ECoG stream as equally reliable.
A binary-weight vision inference stack for milliwatt IoT cameras that streams compact weights instead of feature maps, so battery nodes can run high-resolution detection without hitting the I/O…
An IoT security intelligence workspace that maps each attack surface (device, network, cloud, app) to fit-for-purpose ML/DL monitors so security teams replace brittle encryption-only postures with…
Governs how IoT and edge AI models are trained—human-in-the-loop curricula, crowdsourced fleets, cleansing/curation controls, and continuous improvement baked into the operating model—so IoT data…
Unifies wired/wireless/WAN edge networking with role-based dynamic segmentation, AI assurance, and experience analytics so workplaces, stores, hotels, hospitals, and schools deliver smart digital…
An on-device breathing-acoustic authenticator that runs a quantized RNN on phones, watches, and Pi-class hubs, so IoT unlock stays private and works without a cloud round-trip.