Years of Experience in Research and Development

Knowtion develops algorithms and software tools for applications, where heterogeneous data needs to be processed, not directly measurable quantities needs to be determined and behavoir patterns must be predicted. The focus is on customer-specific consulting and solution development, from system analysis and design through implementation and optimization on the target system.

This calculation allows the determination of state variables that are not directly measureable and a qualitative assessment of sensor data and their correction. Thereby, mathematical methods are used which are able to extract value-added information from the sensor data to predict future behavior of the underlying system and to detect outliers and anomalies.

Sensor Fusion

The state of the system is determined by predefined variables such as position, angular position, or speed. In most industrial processes and equipment, it is not possible to acquire the desired information about the system state accurately, completely, and robustly with a single source of information. The solution to many tasks therefore implies the use and combination of different sources of information.

For the unification of the different information, specific algorithms and software products are necessary. Depending on the application we use primarily the Bayesian information methodology, allowing a systematic fusion of the individual data. In this way uncertainties are minimized and an improved result – maximum information – can be obtained from the data. For example, the position and intention of people can be determined from different camera images and additional data.

Application Advanced Analytics: Sensor fusion
Click here to see our webinar about sensor fusion (in German).

Automatic Data Analysis

In many applications, in addition to the above-mentioned system state, a recognition and measurement of characteristic patterns of information and combination of individual states is necessary.

The individual sensor data is analyzed and evaluated in order to generate added-value information that allows a description of the underlying system. Furthermore, appropriate algorithms are able to predict future behavior and detect anomalies and system faults.

Application Machine Learning: Anomaly detection
Click here to see our webinar about anomaly detection (in German).


Knowtion was founded in 2011. Our highly qualified team members have been active for more than eight years in the research and develop- ment of algorithms for sensor fusion and automatic data analysis and have successfully used them in a variety of client projects.


With our expertise we can help you to deduce more information from your data and get a better basis for assessments and forecasts. Your product will be more reliable and accurate, so you gain a competitive advantage over your competitors.


We have profound understanding and expertise in the field of sensor fusion and automatic data analysis. We are familiar with stochastic estimation theory, information fusion, machine learning, neural network and fuzzy logic.


Knowtion UG
home An der RaumFabrik 33c
D-76227 Karlsruhe
speech+49 721 486 995-10

Scroll to top