Cortex Platform
An industrial IoT platform — from measurement to decision, from the field to the fleet.
Cortex Platform is the name of Venu Electronics’ industrial monitoring product family: the whole formed by hardware installed on site and software running in the cloud. This page explains how the components connect and why the system is built the way it is.
The platform’s components
Data passes through four components in order. The technical detail for each lives on its own product page.
- 01
Sense probes
HardwareTakes the measurement in the field. Mounted inside the panel or on the machine, reading gas, thermal, vibration and air quality channels. Signal processing begins on the probe itself — the FFT and envelope analysis happen here.
Sense probes → - 02
IQ Panel
Hardware + embedded softwareWhere the decision is made. Collects data from every probe over the CAN-FD field bus, time-synchronises it, and processes it on an embedded AI engine to produce a risk score. This step requires no cloud connection.
IQ Panel → - 03
Communication module
HardwareGets the result off site. WiFi, cellular or LoRaWAN — one is chosen to match the site infrastructure and fitted into the IQ Panel expansion slot. Field-swappable.
Communication modules → - 04
Cortex Cloud
SoftwareLearns at fleet scale. Processes risk scores from every panel with deep learning, predictive maintenance and forecasting models, making visible across hundreds of devices what no single device could see.
Cortex Cloud →
The architectural principle: decide at the edge, learn in the cloud
The edge decides quickly and independently
The IQ Panel measures deviation from the monitored asset’s reference signature and raises an alarm when needed, with no cloud connection. A communication outage does not stop alarm generation, and the dry-contact relay outputs operate independently of connectivity. This meets the requirement for a network-independent safety function in critical infrastructure.
The cloud uses time and scale
Cortex Cloud models trends spanning months, comparisons between similar assets, and failure patterns shared across the fleet. The leading signature of a failure mode is learned from assets that have already experienced it.
The loop closes back to the field
Models trained in the cloud are deployed to IQ Panels in ONNX format. Knowledge learned across the fleet improves the inference engine running at the edge — with no need to remove the device or visit the site. The two layers do not replace one another; they feed one another.
Per-device modelling
Every panel and every device is modelled separately. During commissioning the system learns the monitored asset’s own operating conditions, load profile and thermal behaviour. Because the risk score is computed as a deviation from that reference signature, equipment operating in a steady state does not produce a high score, and no manual threshold definition is needed on site. The same software runs on equipment with very different load and speed characteristics without individual tuning.
Field bus topology
Probes connect to the IQ Panel over a galvanically isolated CAN-FD field bus. A single bus supports a daisy chain of up to 32 nodes; no star topology and no dedicated run per probe. This markedly reduces cabling cost, particularly in retrofit installations. CAN-FD runs at 5 Mbit/s and is designed for industrial noise environments.
Integration and commissioning
- Probes are installed without cutting power and with no mechanical modification to the panel enclosure. Commissioning does not halt production.
- The field bus is a daisy chain. A single cable run serves every probe in a panel or along a machine line.
- The IQ Panel learns a reference signature for each monitored asset during the commissioning window. The system collects data throughout that period and scoring begins once learning is complete.
- SCADA integration is done through a Modbus TCP register map; Modbus RTU and MQTT/TLS are also supported. Every probe type reports into a single consistent register model, so there is no per-sensor driver development and no middleware layer is required.
- Firmware and ONNX model files are updated over the network; no site visit is required.
Technical foundations
Sensor fusion
Gas, thermal, electrical and mechanical channels are combined into a single risk function — evaluated together rather than one at a time.
Deterministic acquisition
DMA-based, gap-free sampling on the ThreadX real-time operating system. Short transient events cannot fall between sample windows and be missed.
Spectral analysis at the edge
The FFT is performed on the probe; the raw waveform never reaches the bus. Bus bandwidth stays constant regardless of machine speed.
Frequently asked questions
What is the difference between Cortex Platform and Cortex Cloud?
Cortex Platform is the name of the entire product family: the IQ Panel, the Sense probes, the communication modules and Cortex Cloud together. Cortex Cloud is that family’s cloud software product — one of the platform’s four components. In short: the Platform is the whole, the Cloud is its software part.
Does Cortex work without a cloud connection?
Yes. Risk scoring and alarm generation run at the edge on the IQ Panel. A communication outage does not stop alarm generation, and the relay outputs operate independently of connectivity. The cloud connection belongs to the fleet-scale learning and reporting layer.
Does installation halt production?
No. Probes are installed without cutting power and with no mechanical modification to the panel enclosure. The facility carries on operating normally while the system collects data during commissioning.
Will it integrate with our existing SCADA system?
Yes. Integration is done through a Modbus TCP register map; Modbus RTU and MQTT/TLS are also supported. Because every probe type reports into a single consistent register model, there is no per-sensor driver development.
How is the system updated in the field?
Firmware and ONNX model files are updated over the network. There is no need to remove the device or visit the site.