Virtual Sensors

AI-based virtual sensors are the main feature of our product. They are purely software-based and use standardly available signals of standard vehicles.
We use them to measure loads, like forces, torques, stresses, temperatures or electric currents. These loads wear and damage components like powertrain parts, chassis parts, brake systems, tires, the steering assembly or the batteries of electric vehicles.

Based on the loads measured by our virtual sensors we calculate precise usage profiles for the main components of each individual vehicle of a fleet. Comparisons show that the signals of our virtual sensors match with high-quality and fidelity those of the physical sensors. This means that no additional sensor or hardware is necessary to monitor vehicle systems. Moreover, our virtual sensor technology can even replace existing hardware sensors.

analytics

On our computation server – the AI Core – the data generated by the virtual sensors is enriched with meta-data like material characteristics, test rig results from development manufacturing information, weather conditions, etc. Finally, machine learning pipelines are applied, which learn from all these features and can thus exactly predict why, how and when a component will fail. On top, we also identify potential for design optimization and calculate residual value of vehicles based on their usage.

As shown below, all analyzed data can be managed and visualized through a web-based interface. The remaining lifetime predictions for the most important components of each individual vehicle are graphically depicted and the vehicle data is used to calculate possible material, CO2 and cost savings.

Benefits

Through predictive analytics, COMPREDICT Virtual Sensors & Analytics helps you to avoid the failure of your vehicles due to unexpected damage and to optimize the design and construction of your own components.

AI-based failure
prediction

Lighter & safer
components

Optimized development
processes and testing

Precise value rating
for used cars

Would you like to learn even more about our product?

Visit us at Medium.com for detailed publications.

Medium.Com

About Us

COMPREDICT was founded by Dr.-Ing. Rafael Fietzek and Dr.-Ing. Stéphane Foulard in 2016 in Darmstadt and has currently a team of 20 people from 8 different countries. Our focus lies on software-based load monitoring, failure probability calculation and life time prediction using AI and machine learning for mechanical, mechatronic, electric and electronic components . In the field of R&D, we supervise bachelor and master theses and offer internships.

COMPREDICT works worldwide with international automotive manufacturers and suppliers. We participated in the most renowned accelerator programs like Startup Autobahn in Germany in 2017, German Accelerator Tech in Silicon Valley in 2018 and PlugAndPlay Tech in Sunnyvale in 2019. We also won several prizes including the Johann Puch Open Innovation Award from MAGNA in 2016, the main and the special prize for Big Data at the “Gründerwettbewerb – Digitale Innovationen” from BMWi in 2017, and the second prize at the “Digitales Startup des Jahres” from BMWi in 2018. We have also been ranked in the top 50 Startups of 2018. In march 2020 we successfully closed our second fundraising round to support our growth strategy.