During my internship at Maruti Suzuki India Limited in the Infotainment Department, I developed a real-time audio processing system aimed at capturing clear audio and detecting speech location in a 3-row seater car using Raspberry Pi. Leveraging MCP3008 ADC with SPI communication, I implemented moving average and Butterworth bandpass filters, upsampled data, and utilized FFT for signal analysis and microphone selection, achieving high-quality audio playback.Key Contributions:1. Audio Signal Processing: Developed a system to capture and process audio signals using MCP3008 ADC via SPI communication.2. Noise Reduction: Implemented moving average and Butterworth bandpass filters for noise reduction and signal clarity.3. Data Upsampling and Analysis: Upsampled data and applied FFT for accurate signal analysis and microphone selection.4. Microphone Positioning: Determined proximate positions for six microphones around the car by capturing audio at various locations with different AC configurations (low, medium, high).5. SNR Optimization: Used Audacity software to calculate SNR values from recorded audio data to optimize microphone placement at each seat based on comparative analysis.6. Algorithm Research: Researched and experimented with various neural network based speech separation algorithms, including narrowband, fullband, and subband algorithms with mask estimation modules, to separate speech from a mixture of audio signals and detect the speaker's location.This project significantly enhanced the audio capture quality and accuracy of speech location detection within the car, contributing to a more efficient and clear audio system.

• Conducted doubt sessions and debugged code for students enrolled in the course.• Assisted students with their assignments and various projects.• Mentored 124 students and received a TA rating of 4.85.