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

Akshay Hinduja, Yunsik Ohm, Jiahe Liao, Carmel Majidi, Michael KaessIROS 2022 · DOI

TL;DR

This paper introduces a novel, low-cost acoustic localization and communication system for bio-inspired underwater robots. By leveraging MEMS microphones and a constellation of speakers, the system enables robots to determine their position and receive commands, paving the way for more autonomous and affordable underwater exploration.

What Problem Does This Paper Address?

Accurate localization is crucial for autonomous robots, but underwater environments present significant challenges. GPS signals don't penetrate water, and visual methods are hampered by turbidity. Traditional underwater localization systems often rely on bulky, expensive hardware like Doppler velocity logs or high-end inertial navigation systems, making them unsuitable for small, low-power, bio-inspired robots. Existing acoustic methods, while promising, can still be costly and power-intensive, or require complex synchronization. This paper aims to bridge this gap by developing an affordable, low-power acoustic localization and communication solution specifically for these smaller, bio-inspired underwater robots.

Methods

The core of the proposed system is acoustic pseudoranging, a technique inspired by GPS localization. Instead of relying on precise time-of-arrival (TOA) from synchronized clocks, this method uses time-difference of arrival (TDOA) of acoustic signals emitted from a known constellation of four speakers. The robot, equipped with a low-cost MEMS microphone, listens to these signals and estimates its position by solving a system of equations derived from the pseudorange measurements.

Key aspects of the method include:

  • Low-Cost Hardware: Utilizing readily available MEMS microphones and Raspberry Pi microcontrollers for both the robot's receiver and the speaker transmitter.
  • Acoustic Pseudoranging: Estimating the time difference of arrival of chirped acoustic signals from multiple speakers to determine the robot's position and time bias.
  • Matched Filtering: Employing matched filters to process the recorded acoustic signals and extract the most significant peak, which corresponds to the pseudorange estimate. This is crucial for handling potential non-line-of-sight reflections.
  • One-Way Communication: The same acoustic framework is used to transmit simple commands to the robot, enabling basic motion control by differentiating communication chirps from localization chirps.
  • Dilution of Precision (GDOP) Analysis: The paper analyzes the impact of speaker geometry on localization accuracy, a concept borrowed from GPS, to understand the "solvable volume" for reliable estimates.

Datasets

Not discussed in the paper. The experiments were conducted in a controlled water tank environment with a known speaker setup and ground truth measurements.

Key Insights and Ideas

  • Democratizing Underwater Localization: The use of inexpensive MEMS microphones and microcontrollers significantly lowers the barrier to entry for acoustic localization, making it accessible for a wider range of research and applications.
  • Robustness to Non-Line-of-Sight: The paper highlights the importance of selecting the first major peak in the matched filter output, suggesting a practical approach to mitigate issues caused by signal reflections in enclosed spaces.
  • Speaker Geometry Matters: The analysis of GDOP demonstrates that the spatial arrangement of speakers is critical for achieving accurate localization, similar to how satellite geometry affects GPS. This insight is vital for designing effective underwater acoustic positioning systems.
  • Unified Framework for Localization and Communication: The ability to perform both localization and basic command transmission using the same acoustic signals and hardware simplifies the system architecture and reduces complexity for the robot.

Key Results

The paper demonstrates successful localization of a bio-inspired robot in a water tank. While specific quantitative results vary with speaker configurations, the authors report:

  • Mean error of 0.345m with a standard deviation of 0.228m across various trials and configurations.
  • XY RMSE values are considerably lower than total RMSE, indicating that depth estimation is more challenging due to less variation in speaker heights compared to X-Y positions.
  • Configuration 3, with greater variation in speaker positions, generally yielded better localization accuracy (smaller covariance ellipses) compared to Configuration 1.
  • Successful execution of basic directional commands to the robot using acoustic signals.

Limitations and Future Work

The authors acknowledge several limitations and propose future directions:

  • Enclosed Spaces vs. Open Water: The experiments were conducted in a tank, which is prone to multi-path reflections and speaker washout. The system is expected to perform better in larger, open water environments.
  • Limited Communication Protocol: The current communication is basic, relying on differentiating chirp types. Future work aims to implement more robust protocols like frequency shift keying for a wider range of motions and tasks.
  • Filtering and Smoothing: The current localization relies on a direct least-squares solution. Future work will involve implementing filtering and smoothing techniques (e.g., Kalman filters) for more accurate state estimation and optimization.
  • Sensor Fusion: Integrating data from other sensors like IMUs and pressure sensors with the acoustic pseudoranging estimates is a promising avenue for improving state estimation.
  • Open Water Testing: The ultimate goal is to demonstrate the system's efficacy in real-world open water scenarios, comparing it to traditional inertial navigation methods.

Summary updated

BibTeX
@inproceedings{hinduja2022acoustic,
  author = {Akshay Hinduja and Yunsik Ohm and Jiahe Liao and Carmel Majidi and Michael Kaess},
  title = {{Acoustic Localization and Communication Using a MEMS Microphone for Low-cost and Low-power Bio-inspired Underwater Robots}},
  booktitle = {IEEE/RSJ International Conference on Intelligent Robots and Systems},
  year = {2022},
  doi = {10.1109/iros47612.2022.9981355},
}
Figure