Main focus of research at the Autonomous Systems Technology Institute is the development of autonomous mobile robot platforms. For example, such systems are to be enabled to explore and navigate in unknown unstructured environments on their own.
Research
Other Topics
CUDA
GPUs are highly scalable processor architectures with potentially thousands of cores and, compared to CPUs, magnitudes higher memory bandwidth.
On the other hand, they have less single-thread performance.
Therefore, GPUs are predestined to process the scalable part of the code, whereas the CPU processes the other part.
For example, the training and inference of Deep Neural Networks benefits from the use of GPUs.
For this purpose, we have our institute's CUDA-cluster. The dominant programming language for Nvidia-GPUs is CUDA.
In the following a fraction of our institute's applications:
- SIFT-Features
- Stereo through Semi-Global-Matching
- Pixelwise segmentation to enhance our online calibration
Machine Learning
Optimization
Recursive Estimation (Filtering)
The goal of filtering is the estimation of a true state of a process with erroneous measurements.
For a efficient implementation it is desirable to use recursive algorithms, with a defined process model for the relation between different time steps. Potential model errors can be tackled with uncertainties.
A prominent recursive filter is the Kalman-Filter which is an optimal estimator for linear, additive and Gaussian process and measurement noise.