In the summer of 2023, during my second year at university, I was selected for the prestigious Imperial College Center for Cold Matter Research Program to work alongside Professor Ben Sauer. I was tasked with reducing the internal magnetic noise within our ultracold electric-dipole-moment apparatus, which offers a direct window into the unknown dimensions of our universe. Initially, I decided to use independent component analysis to decompose the internal noise, which yielded suboptimal results due to the non-linear mixing of the noise sources. 

​

Drawing inspiration from noise-canceling headphones, I worked on developing a dynamic noise compensation system with a much higher spatial complexity. The system incorporates six compensation coils and up to thirty magnetic field detectors, from which a proportionality factor matrix was deduced. The key of my method lies in the utilization of the optimized Moore-Penrose Pseudoinverse of the proportionality matrix. On top of that I have also implemented a PID control software to ensure the stability of the magnetic field. The completed prototype could reduce the noise floor by a factor of up to 3.1 at 0.1 Hertz, which will make our measurements some of the most accurate in the world.