underwater-acoustic-pseudoranging

Acoustic Localization and Communication Using a MEMS Microphone for Low-cost and Low-power Bio-inspired Underwater Robots

This repository provides the code used to perform acoustic pseudoranging, as well as a few notes on hardware preparation and observations for future iterations. A copy of the paper can be found here.

Note:

This page is still being updated gradually with more useful information. Come again later for an updated version! (10/24/2022)

Dependencies

The system runs on a raspberry pi using Python3. The following non-default packages are required:

• numpy
• librosa
• soundcard
• soundfile
• pyaudio

Hardware setup

Transmitter side-

• Raspberry Pi 4 with a generic USB sound card allows 4 channel or more output to connected speakers.
• Generic dual channel amplifiers are required to interface speakers to the soundcard
• Speakers can also be generic, we used the Lubell UW30 speakers which were the most cost effective, off the shelf option
• Speaker placement can be determined through a GDOP analysis in simulation

Receiver side-

• Raspberry Pi Zero W
• Updating to RT-Preempt Kernel reduces latency in reading inbuilt Real-time clock. Recommended to perform, but not necessary
• I2S ICS43434 microphone is connected. Wiring diagram is attached below.
• 3D printed water tight boxes are used to house computation with the microphone top port on top, covered by a thin adhesive film, 1-2mm thick. The microphone has to be sealed with silicon glue to ensure water tightness.

A cross sectional view of how we prepared the MEMS microphone for waterproofness is shown below.

More details on MEMS microphone waterpoofing using thin membranes and its effects can be found in this helpful pdf link from invensense here.

Future improvements in design

• Using a different microcomputer capabale of using PCM inputs is ideal. These allow for higher sampling rates than I2S, making a wider range of audio signals possible.
• Vesper MEMS systems also has several PCM and analog MEMS microphones which are water resistant.
• Oil bladders are a way to wterproof mics which may allow them to work at deeper depths. Action cameras like GoPros use a waterproof diaphragm for each of their microphones. Applying a similar system would yield more effective waterproofing.

Some future directions for the algorithm

• State estimation using online filters
• 2 microphone system for bearing estimation along with location