Introduction

The project focuses on a Real Time Direction of Arrival (DOA) estimation system using dual-channel audio input from laptop microphones. The system captures sound signals from two microphones and estimates the direction of the source by computing the time delay between the signals using cross-correlation. The estimated delay is then used to determine the angle of arrival, which is visualized through a simple software interface..

The motivation behind this system lies in understanding and implementing the fundamental principles of signal-based localization. In many applications such as acoustic tracking, surveillance, and communication systems, identifying the direction of a signal source is essential. However, practical implementation involves challenges such as signal synchronization, noise, and computational complexity, especially when attempting real-time processing on limited hardware resources.

This project addresses these challenges by adopting a simple and intuitive correlation-based approach and building a complete processing pipeline involving signal acquisition, delay estimation, and visualization. While the system is not fully real-time due to time constraints, it successfully demonstrates accurate angle estimation in a controlled setup and establishes a foundation for future improvements involving real-time execution and hardware acceleration.
 

Problem Statement

The project aims to solve the challenge of implementing Direction of Arrival (DOA) estimation using simple hardware and signal processing techniques.

The core problems include:

• Capturing synchronized dual-channel audio signals using laptop microphones
• Accurately estimating time delay between signals using cross-correlation
• Handling noise and signal distortion affecting correlation accuracy
• Resolving ambiguity caused by multiple correlation peaks for periodic signals
• Converting time delay into angle of arrival based on microphone geometry
• Designing a complete pipeline integrating signal acquisition, processing, and visualization
• Managing computational and interfacing hardware related constraints preventing real-time execution

Methodology

The system is implemented as a pipeline consisting of signal acquisition, delay estimation, hardware pipeline design, and system integration.

Signal Acquisition

• Audio signals are captured using dual-channel laptop microphones
• Data is recorded using Python scripts.
• Signals are preprocessed to ensure synchronization and consistency
• A reduced number of samples is used to improve processing efficiency

Delay Estimation (Algorithm Design)

• Cross-correlation is used to estimate the time delay between the two signals
• The lag corresponding to the maximum correlation peak is identified
• Signal constraints (e.g., single-cycle inputs) are applied to avoid multiple peaks
• The estimated delay is used to compute the angle of arrival

Hardware Design (Verilog Implementation)

• A pipelined cross-correlation module is designed for efficient computation
• The implementation supports both positive and negative lag detection
• Fixed-point arithmetic is used for hardware compatibility
• SystemVerilog code and testbench are developed for verification
• Simulation is performed using Icarus Verilog

System Integration 

• Python scripts are developed for data capture and storage
• Verilog simulation outputs are processed using Python
• A complete pipeline is built: data capture → processing → visualization
• Execution is automated through a command-line interface

Results

Successfully simulated the complete end to end pipeline for DOA estimation for single frequency (1000 Hz) source. The least count measurable by our model is 12.85 degrees. The code developed is synthesizable on hardware. Real time viewing of angle works with workable delay which will be considerably reduced if used an actual FPGA. As the whole model works on icarus verilog simulation, it is not able to achieve a very high speed.
 

Conclusion

This project successfully demonstrates a functional Direction of Arrival (DOA) estimation system using dual-channel laptop microphones and a cross-correlation-based approach. By integrating signal acquisition, delay estimation, hardware-oriented design, and visualization into a unified pipeline, the system validates the feasibility of implementing DOA estimation using minimal and accessible hardware.

Although real-time performance is limited due to simulation constraints, the design achieves reliable angle estimation in controlled conditions and produces synthesizable hardware logic suitable for FPGA deployment. Overall, the project establishes a strong foundation for transitioning toward real-time, scalable, and more robust DOA systems.
 

Future Work

Wide band frequency DOA estimation, Scaling array to 2 and 3 Dimensions for better accuracy and precise measurement of the signal.