Through past and ongoing research and customer projects, we have built up a great expertise in the field of Artificial Intelligence. We are ready to share our knowledge in the AI sector and offer it either in a consulting package or directly in a development service package.
Together with you, we look at your individual use case and develop ideas & solutions to drive your projects forward.
When managing test drives, the automated, in-vehicle assignment and categorization of recorded data is a great benefit. The detection of objects and whole scenarios is increasingly based on methods of artificial intelligence and enables the pre-labelling of data already during the test drive. Therefore, the processing and analysis of data is supported significantly. In addition, the labelled data can also be used for online campaign management.
During validation and development of new driving functions it is very interesting to detect data that differs from regular characteristics (outlier detection). This data indicates anomalies, outliers and can indicate a possible mailfunction in the algorithm. Machine learning methods detect these anomalies, classify them automatically and support the analysis of their origin.
Camera-based object detection and surrounding pose estimation offers a wide range of possibilities for the automation of various situations and targeted navigation. Our deep knowledge in classic and AI-based algorithms can support the successful development and realisation of new ideas, whether it is an ADAS or robot control system project.
For testing automated driving functions, it is necessary to modify target scenarios precisely and reproduce them uniquely. In HiL or SiL processes, there are many use cases, such as the insertion of virtual scenarios (e.g., with generative methods), development of individual SiL algorithms (e.g., GPGPU algorithms for reproducing DSPs), or intelligent management of HiL clusters.
Data Preprocessing
Classification
Image Processing / Object Detection
Regression
Clustering
Feature Extraction
Anomaly detection
Postprocessing
