Jiaxin Sun

Topic: Investigation of non-linear regression models for the identification of time-dependent DC link current in electric and hybrid vehicle applications (interdisciplinary task between electromobility and artificial intelligence, Data Mining Process). Algorithms: Random Forest, Decision Tree, Fully Connected Neural Network, K-Nearest Neighbor. Development of an user interface using machine learning methods (GUI) which simplifies the training process of a data-based model.

Internship in e-Control und e-Drive engineering, tasks: – Searching for best setpoints based on FFT – Generation of pseudo-random signals based on linear feedback slide register – Data exchange between test bench and simulation with MATLAB – Verification of the pseudo-random signal with MATLAB

Specialization area: Information Technology and Communication (Scientific Visualization, Digital Image Processing, Fiber Optic Communication, Electromagnetic Wave and Components)

Specialization area: Information Technology and Communication (Statistical Adaptive Signal Processing, Detection and Pattern Recognition, Deep Learning, Communication, Digital Image processing) – Master’s thesis: Exploring Speech Enhancement using Hybrid Domain Models (Generative Adversarial Networks (GANs), WaveNet, TensorFlow, PyCharm, MATLAB, GPU acceleration, remote server, large data sets) – projects: 1. Recognition of prostate cancer 2. Speech recognition based on Gaussian Mixture Model

Bachelor’s thesis at the Institute of Communication Engineering: Performance analysis of different guard intervals in OFDM based on MATLAB, investigation and comparison of the BER-Performance with three common guard intervals (cyclic prefix, cyclic suffix, zero padding) and their combinations