MACHINE LEARNING FOR ELECTRONIC SENSORS

What we do

Our machine intelligence technology predicts unexpected technical-hardware problems

Machine Intelligence

Machine Intelligence

Machine intelligence is a hybrid of machine learning and artificial intelligence. Selecting the best components to create low-power consumption edge computing. This can detect 53% of internal faults that are hidden from sight and improve system performance by up to 77%.

Business

Machine learning sensors reduces the cost of maintenance by identifying problems before they become expensive and take a system off-line. It also helps with managing the scheduling of product servicing, calibration and warranties.

Engineering

We have developed machine technology that can bolt-on to existing systems for remote condition monitoring. Cambridge Micrologic has developed machine intelligence reduced learning algorithms. Integration is simplified using the enhanced serial bus and small data sets.

INDUSTRY 4.0 IoT

Predict technical-hardware failures

Equipment Calibration retention

Active maintenance schedule

Increase productivity

Smart warranties

Reduce costs

How does it work

Bolt-On

Using ESB™, machine intelligence can be added to any existing microprocessor system with only five wires. Bolt-on technology behaves as a parallel processor that listens to data streams, analyses, then reports as meta-data.

Integration

Four chips combine to make machine intelligence simple for circuit integration. Also machine learning can be embedded in existing customer chips.

Product Development

Technical file for electronics, software, product approvals and design notes. Over twenty-five years of electronics and software design experience to aid product design.

Licensing

Machine intelligence learning is available for commercial deployment via a licencing agreement.

DOWNLOADS

White papers, technical notes and use studies

RUTHERFORDIUM™

Development machine learning platform
Dual ARM core M4-3 for parallel processing

Development Machine Intelligence Platform 


Two ARM cores for machine learning processes. An ARM Cortex M4 for board-level control and machine learning processes. Its ARM Cortex M3 is open for the user to program. The four-port connector is for common third-party daughterboards. This board is size 112x53mm and has onboard debug tools.

This is a development platform. It is easy to use. Either as a platform or a reference design. Many third-party daughterboards can be plugged in. Can be programmed using any ARM software suite. Early-stage product designs start here. The application includes IoT machine learning, monitoring and security.

The four-port connector is for common third-party daughterboards. Ports are ‘power’, ‘analogue’ and two ‘digital’. I2C and SPI combine for an enhance serial bus. Daughterboards can be sensors for IoT, displays and co-processors.

 

Features and benefits:

1W (5Vdc, 200mA) operation
Power monitoring and management
USB debug interface
JTAG and USB program ports
Enhanced serial bus
On-board accelerometer and temperature sensors
SD card to store data and machine learning
Multiple LED for user and debug status
Machine learning ARM core Cortex M3
User-programmable ARM core Cortex M4
Nano-amp real-time clock with backup power
Four-port connector for third-party plug-in daughterboards
In-product anti-counterfeit technology

Development Machine Intelligence Platform 


Two ARM cores for machine learning processes. An ARM Cortex M4 for board-level control and machine learning processes. Its ARM Cortex M3 is open for the user to program. The four-port connector is for common third-party daughterboards. This board is size 112x53mm and has onboard debug tools.
This is a development platform. It is easy to use. Either as a platform or a reference design. Many third-party daughterboards can be plugged in. Can be programmed using any ARM software suite. Early-stage product designs start here. The application includes IoT machine learning, monitoring and security.
The four-port connector is for common third-party daughterboards. Ports are ‘power’, ‘analogue’ and two ‘digital’. I2C and SPI combine for an enhance serial bus. Daughterboards can be sensors for IoT, displays and co-processors.

More

NEXT GENERATION TECHNOLOGY

 

Schematic design and programming guides

Low-power consumption electronics

Simplified third-party integration

Electromagnetic compliance

EINSTEINIUM™

DIN rail enclosed for fast deployment
Programmable machine learning controller quad ARM core

PMLC Platform

 

Four ARM cores for machine learning processes. An ARM Cortex M7-M4-M3-M3 for advanced sensor machine learning processes. It is deployment-ready and can fit into DIN rail mount enclosures.

This is a complete platform. It is easy to use. Either as a platform, or a reference design ARM core interfacing is a register control system with handshaking. On-board there are accelerometers, temperature and GPS. Data is output by Bluetooth or SD card. The enhanced serial bus allows extra sensors to be added.


The board can be a reference design or deployed in a DIN enclosure. We provide technical support for design and many third-party forums are available. programming is by any ARM software suite.

 

Features and benefits:

2.5W (5Vdc, 500mA) operation
Power monitoring and management
USB debug interface
JTAG program ports
Enhanced serial bus and differential enhanced serial bus
On-board dual-accelerometer, temperature GPS sensors
SD card to store data and machine learning
Bluetooth 4.0 for mobile interface
Multiple LED for user and debug status
System administration control ARM core Cortex M3
Machine learning ARM core Cortex M3 and M4
User programmable ARM core Cortex M7
Nano-amp real-time clock with backup power
Four-port connector for third-party plug-in daughter boards In-product anti-counterfeit technology.

PMLC Platform

Four ARM cores for machine learning processes. An ARM Cortex M7-M4-M3-M3 for advanced sensor machine learning processes. It is deployment-ready and can fit into DIN rail mount enclosures.

This is a complete platform. It is easy to use. Either as a platform, or a reference design ARM core interfacing is a register control system with handshaking. On-board there are accelerometers, temperature and GPS. Data is output by Bluetooth or SD card. The enhanced serial bus allows extra sensors to be added.


The board can be a reference design or deployed in a DIN enclosure. We provide technical support for design and many third-party forums are available. programming is by any ARM software suite.

More

MACHINE INTELLIGENCE

Specialised solutions for temperature and vibration sensors

BARYON™

Machine intelligence in integrated circuits

Machine Learning Chips

 

Simplify bolt-on machine intelligence to existing hardware. Easy to use chips for circuit designers. Tailored for use. They are the machine learning heart of customer designed systems.

Machine intelligence is fragmented into four components:
Data Cleansing to remove errors and bias.
Machine Learning algorithms.
Performance Analysis to verify results.
Algorithm Adaptation to improve learning.

Enhanced serial bus™ (ESB™) is a five-wire communication bus in chips. It is for adding multiple I2C/SPI devices. ESB™ reduces microprocessor signals, improves board connectivity to sensors and improves EMC.

Machine learning, ESB™ and debug features are programmable to a wide range of chips and footprints.

Machine Learning Chips

Simplify bolt-on machine intelligence to existing hardware. Easy to use chips for circuit designers. Tailored for use. They are the machine learning heart of customer designed systems.

Machine intelligence is fragmented into four components:
Data Cleansing to remove errors and bias.
Machine Learning algorithms.
Performance Analysis to verify results.
Algorithm Adaptation to improve learning.

More

NEWS AND UPDATES

Follow us on social media

Market trends and analysis

Engineering design notes and products

HADRON™

Machine Intelligence Algorithms

 

For any machine learning system to be of use, it must generate a conclusion. Hadron™ does this in real-time.

Hadron™ analysis two data streams relative to each other as if they are monochromatic photographs. Hadron™ does not require datasets to be correctly aligned to analyse. It infers the relationship in X and Y-axis.

Hadron™ can be applied to any two streams of datasets and without the two having to be perfectly aligned. 

 

Features and benefits:

Diverging and converging trends analysis
Unstable data
Product moment for linearity
Regression coefficients
Fuzzy-logic
Automated conclusions
Time-domain analysis
Frequency-domain analysis
Multi-dimensional cross-referencing

Machine Intelligence Algorithms

For any machine learning system to be of use, it must generate a conclusion. Hadron™ does this in real-time.

Hadron™ analysis two data streams relative to each other as if they are monochromatic photographs. Hadron™ does not require datasets to be correctly aligned to analyse. It infers the relationship in X and Y-axis.

Hadron™ can be applied to any two streams of datasets and without the two having to be perfectly aligned. 

More

Get in touch

Address

Cambridge Micrologic Limited
Eagle Lab
28 Chesterton Road
Cambridge
CB4 3AZ
Great Britain

Contact

hello@micrologic.io
+44 759 2650 643

Info