Klaus Okkelberg
Klaus Okkelberg is an Electrical Design & Analysis Engineer at Boeing. In his current position, he works at Boeing Phantom Works Space on moving ground-based signal processing algorithms to space, reducing turnaround time, which is crucial for enabling autonomous operations. Klaus also works with artificial intelligence and machine learning (AI/ML), specifically deep neural networks (DNNs), for image-based inspection and pose estimation – the detection of the position and orientation – of unknown satellites. Supervised AI/ML greatly reduces the need for feature engineering, meaning fewer engineers and revisions are needed per project. Using DNN algorithms has sped up development of new image processing functions and reduces the cost of developing and deploying new space-based imaging sensors, which has led to many new business opportunities for Boeing with national space government customers. Klaus has personally developed several key innovations in AI/ML, including data augmentations, training regularizations and high-dimensional space embeddings. These developments greatly reduce generalization error and enable the use of real satellite imagery from different sources without additional pre- or post-processing. Klaus is a graduate of Penn State Schreyer Honors College with a Bachelor’s degree in Electrical Engineering and of the University of New Orleans with a Master’s degree in the same field. At Penn State, his thesis on nonlinear control of marginal oscillators contributed to the more precise detection of explosive materials, providing immense benefit to potential customers such as the Department of Homeland Security and the Transportation Security Administration (TSA). Klaus interned at Xilinx in their signal integrity division. Klaus also has 4 years of PhD coursework and research at Georgia Tech in Electrical and Computer Engineering and intends to pursue a PhD in machine learning.
In addition to his work on AI/ML for Boeing, Klaus served on a team that developed a near real-time FPGA device for increasing the sensitivity of a wide instantaneous-bandwidth radio frequency (RF) receiver assuming no knowledge of the received signals. The team began with an existing algorithm implementation for a smaller bandwidth with the intent to scale it up for the target FPGA platform, but Klaus discovered during the project that it did not meet customer needs due to increased signal distortion and high tuning requirements. He created a hardware-friendly version of the improved algorithm that lowered hardware SWaP (size, weight and power) by over 70% while doubling the effective filter length, leading the team to receive the BDS Engineering Team of the Year and Phantom Works Team of the Year awards in 2020. In the following year, Klaus presented a paper at the Advanced Maui Optical and Space Surveillance Technologies (AMOS) conference on his use of self-supervised learning in DNNs to improve pose estimation of space objects on work for the Air Force Research Laboratory (AFRL). Additionally, Klaus developed a noise reduction system using a DNN for Bayesian signal detection across a wide RF spectrum combined with optimal filtering to maximize the signal-to-noise ratio of the output. The noise reduction algorithm was implemented to run in near real-time in hardware on Xilinx's Versal AI platform, and this proof of concept enabled the PW Space team to assess the feasibility and details of implementing a ML algorithm in a newly available AI accelerator on space qualifiable hardware for future space missions.
Outside of work hours, Klaus is a member of the Boeing Asian and Pacific Association (BAPA) and the Boeing Employees Pride Alliance (BEPA). He is also a member of Out in Science, Technology, Engineering and Mathematics (oSTEM), an organization that supports LGBTQ+ collegiate individuals in STEM fields. Klaus has volunteered at several of their conferences and acted as a role model and mentor for students. He also previously volunteered with Habitat for Humanity to build and repair homes in New Orleans and with Penn State THON to raise money for childhood cancer. Klaus and his wife foster senior dogs through a local organization and have adopted one of them as well.
