Artificial intelligence
We turn raw sensor data into decisions — on the board, in the field or on your own server. Without bringing in a separate software house: the same team that designs the hardware.
- 01
Decisions at the edge
Sensor data processed on the board itself, with the decision made in place. In our lighting controller the movement decision derived from the radar signal is produced on the board and never leaves for the cloud.
- 02
Signal processing
Filtering, feature extraction and classification of raw signals such as radar, current and vibration. Because it happens under the same roof as the hardware design, sensor selection and algorithm are optimised together.
- 03
Insight from field data
Modelling the data gathered from IoT nodes. In the Smart Bee Colony Monitoring system, temperature, humidity and sound data is interpreted by an AI model and delivered to the beekeeper's mobile app as colony health.
- 04
Image and video analysis
Image analysis integrated with network cameras; used in devices that connect to ONVIF ecosystems and in our unmanned aerial vehicle projects.
- 05
Predictive maintenance
Deviation detection in time series data: catching the change in behaviour before the failure arrives.
- 06
In-house knowledge assistant
An assistant that works over approved technical documents and cites its source for every answer. It turns minutes of searching into seconds.
Our principles
- DATA STAYS INModels run on your own server; customer data is not sent to cloud services
- CITES SOURCESEvery answer names the document it rests on; unverifiable output is not produced
- HUMAN APPROVALThe system proposes, a person decides on anything critical, and the decision is recorded
- ROLE-BASED ACCESSEveryone sees only what they are authorised to see; access is logged
- MEASURABLEAccuracy is measured regularly and improvement is shown in numbers
You have data but no return from it?
Tell us what you are collecting and we will assess together what can be drawn from it.