This week, I focused on preparing our Hybrid LxFMS ANC algorithm for eventual integration with the embedded hardware. Our MATLAB simulation has already demonstrated promising noise attenuation, particularly in the lower-frequency range, so I began examining how the existing algorithm can be translated into an implementation suitable for real-time execution on the STM32. This involved reviewing the filtering operations, adaptive coefficient updates, and secondary-path modeling required by the algorithm while considering the computational and memory limitations of the embedded platform.
I also continued investigating the interaction between the physical acoustic configuration and the ANC feedback loop. Because the effectiveness of the algorithm depends heavily on the relative positions of the reference microphones, error microphones, and speakers, I looked further into how these components should be arranged within our wearable design. In particular, I considered how microphone placement and acoustic propagation delays could affect the ability of our system to generate anti-noise before unwanted sound reaches the user’s ears.
Alongside the algorithm work, I continued thinking about the physical construction of the wearable, including the integration of the speakers, microphones, and electronics into a comfortable sleep-mask form factor. My work is currently on schedule, although the final acoustic configuration will require experimental validation once the hardware is assembled. Next week, I plan to continue preparing the ANC implementation for embedded execution and begin developing a more concrete testing procedure for evaluating attenuation across different frequencies and microphone configurations.