About the project
This project aims to explore the design of an alternative class of microphone arrays, so-called co-prime microphone arrays, for recording and processing speech in adverse environments. This will address a key challenge to improve the performance of modern smart devices in noisy and reverberant environments.
Smart devices have been rapidly advancing over the past few decades, including robots, autonomous vehicles and smart home assistants (e.g. Amazon Echo, Google Home, Microsoft Kinect), which leads to a revolution in our daily lives. A clear recording of speech is highly crucial for human-machine interaction, without which the performance of all smart devices is limited.
In the real world, there is always noise, reverberation and other interfering sounds, and the recording and processing of the target speech is highly challenging. Therefore, a high-quality, high-efficiency and low-cost speech recording and processing method with robustness to adverse environments is urgently needed.
Existing recording approaches use an array of uniformly spaced microphones that are used to obtain a desired output with emphasis on the sound from the target direction and suppression of the sound from all other directions. Jointly processing the multiple microphone recordings to form a single output is termed beamforming and can bring much clearer recordings compared to a single microphone with the presence of noise, reverberation and simultaneous interfering sounds. However, the speech is time-variant and the frequency range of such signals is broad. Thus, the performance of speech recording and processing offered by traditional uniform microphone arrays is not sufficient.
This project aims to investigate co-prime microphone arrays, which have shown great potential in the recording and processing of speech in adverse environments. It is of high value to further develop the geometry, beamforming and speech direction of arrival estimation algorithms using co-prime microphone arrays, which are highly practical for real-world smart products and will boost their development and application.
The project will focus on:
- Investigation of fundamentals of sound propagation, beamforming and array signal processing to evaluate the strength and limits of different microphone arrays.
- Design of new co-prime microphone array geometries and analysis of their properties and performance.
- Implementation of new algorithms for speech processing using the designed microphone array, including direction of arrival estimation, source separation and speech enhancement.
- Evaluation of the designed microphone array and algorithms for achieving real-time technology tested in real-world environments.
