System Calibration

Calibrate complex systems faster for key performance requirements

Balance hundreds of input conditions and parameters.  Find the best options faster.  Move forward with confidence. 

BAE Systems

Combine test data and AI models to predict system performance.  Make decisions with fewer tests but more confidence.  Make better tradeoffs in less time. 



Highly-complex, non-linear systems are challenging to design and calibrate to perform under different conditions.  Predicting which combination on input values and parameter settings are required to meet system requirements can quickly exceed the what the human mind can comprehend.  Instead of running more tests and collecting more data, make your decisions through machine-learning models faster and with greater confidence. 

it installation
Calibrate with fewer tests to save time and money
research and collab
Optimize across hundreds of conditions quickly
System calibrtaion
Set coefficients with greater confidence

“Monolith allowed us to understand and optimize the gas meter's behaviour for all operating conditions and optimize meter accuracy under extreme conditions, allowing us to build a superior, more accurate product in a much shorter amount of time.” ​

-Dr. Bas Kastelein, Sr. Director of Product Innovation, Honeywell Process Solutions

Find optimal calibration values faster with machine learning

Finding the proper balance of parameter settings to ensure your dynamic system performs properly under different conditions is a huge challenge.  Whether you are programming automotive engine control modules, smart sensors and measurement devices, or transmission systems - making decisions on parameter settings to meet your performance requirements comes down to an engineer making the final recommendation and putting their name, and reputation, behind it.

Traditionally, engineers rely on gigabytes of test data from prototypes on test tracks, dynamometers, or complex test rigs.  At times, more testing and more data only makes the decision more difficult - with more operating conditions to consider.  Machine learning models offer a smarter path to finding the optimal values for your system.

Engine calibration_ (DONE)

Machine learning models don't replace your testing and data collection efforts, but they make it more efficient.  You start by training models with test data.  Using those models, you can predict performance characteristics under more operating conditions without having to run more tests.  So you can quickly get an idea of how your system performs in different scenarios.

If you need more data to make the best decision, you can again go back to the model for recommendations on which data will give you the most insight into your design.  Rather than guess at which additional tests you should run to learn more, tools like Next Test Recommender give you direction on the best tests needed to address unknowns in your design space.  When you apply machine learning to the problem, you can move much faster with even more confidence. 

Dashboard - Forecast Degradation

Even the most experienced engineers can reinforce their knowledge with machine learning models.  When your working on a complex system, AI models can help you understand which inputs have the greatest influence on the outputs you are optimizing for. 

Balancing performance, efficiency, and safety in any calibration exercise starts with knowing exactly how your system operates and which inputs are most critical.  AI optimization tools can make tradeoffs across hundreds of options to find the best values for your system settings. Although machine learning models are too big to directly embed in controller processors, they help you get to the best option for system coefficient values faster and with more confidence. 

6245870d2708030972ffc879_Smart Meter-2
Honeywell Case Study 


Smarter Energy Measurements Using Predictive Self-Learning Models


Explore how engineers at Honeywell have been able to utilise Monolith self-learning models to reduce product development times by 25% and achieve smart meter product safety certification.


No-Code AI Software


By engineers. For engineers.


  • Avoid wasted tests due to faulty data
  • Build just the right test plan - no more, no less 
  • Understand what drives product performance and failure
  • Calibrate non-linear systems for any condition 
battery health

Trusting test data 


It's vital to understand that testing every possible scenario is not feasible. Over-testing confirms what's already known, while under-testing risks failing certification or missing issues. To optimize testing efforts, identify critical performance components and prioritize tests accordingly.

How We Solve It:

Revolutionised testing 


Using self-learning models that get smarter with every test, Monolith identifies the input parameters, conditions, and ranges that most impact product performance so you do less testing, more learning, and get to market faster. 

Notebook - Manipulator Menu (Apply2) (1)
4 validation cases white paper battery testing
AI for Simplifying Validation Testing


4 applications for AI in validation test


AI has a significant impact on validation testing in engineering product development. You can reduce testing by up to 73% based on battery test research from Stanford, MIT, and Toyota Research Institute. Learn more with Monolith:

Key use cases 

test plan optimize
Test Plan Optimization
test Data Validation
Test Data Validation
root-cause analyiss
Root-Cause Analysis 
forrester report with monolith state of AI in engineering-1
A commissioned study conducted by Forrester Consulting on behalf of Monolith


The State of AI in Engineering  

First-ever study on AI in product development surveys US and European automotive, aerospace and industrial engineering leaders.

Ready to get started?

BAE Systems