Christoph Doehring

Topics so far: - standard application software development for LiDAR sensors - software development for AI/DNN-based object detection from LiDAR and camera data - linux kernel driver development for GMSL camera on Nvidia Jetson - designing area-optimized hardware for digital signal processing functions using VHDL

Accurate object detection for LiDAR data. Focus on deep neural network architectures (training & inference) as well as their efficient implementation in Xilinx SoCs using Vitis AI toolchain.

Dataflow FPGA accelerators for quantized neural networks (QNNs), fault-tolerance in QNNs, dependable QNN accelerators based on built-in self-tests. Contributed to research paper: https://ieeexplore.ieee.org/document/8875314

Embedded software programming including low-level driver development for ARM Cortex M cores.

Project-driven masters program targeting hardware accelerated 3D object detection for LiDAR sensors. Study courses were chosen to map the project scope, intensifying background in computer vision, parallel processing and research competences.

Computer science study with focus on embedded systems.