Cooperatively Mapping of the Underwater Acoustic Channel by Robot Swarms

Abstract

This paper presents a novel approach that employs coordinated robot swarms to create a spatial map for parameters of underwater acoustic communication channels. Understanding spatial variations of acoustic channel parameters is crucial to underwater networking with mixed static and mobile communication nodes. We envision using multiple marine robots maintained in a formation to collect simultaneous measurements of communication signals transmitted by a sound source. The data collected is processed by a cooperative filtering algorithm to generate a local map of a class of channel parameters. The motion of the robots will then be planned by an explorability based method to facilitate map construction under the influence from uncertainty in channel parameters. Our work integrates robotic mapping methods with underwater acoustic communication channel modeling and identification algorithms. We have also combined simulation tools from robotic map making and acoustic channel simulations, which may be valuable for future field experiments.

Model-based learning of underwater acoustic communication performance for marine robots

“…In addition, the GLS mean estimation facilitates a more accurate determination of the covariance function. To this end, we employ the detrended measurements (23) and iterate the covariance training. The training is terminated when,…”

“…A distributed kriging methodology was used in [22] to estimate coverage holes in large-scale wireless sensor networks. The authors in [23] developed a cooperative robust algorithm to compose a spatial map of underwater acoustic communication signals and channel parameters using an H ∞ filter and ordinary kriging. In [24], the acoustic communication performance of micro autonomous underwater vehicles (AUVs) was assessed with field trials.…”

Future Trends in Marine Robotics [TC Spotlight]

“…However, as reliable commercial acoustic modems have become more accessible and easier to use, a confluence between marine robotics and underwater networking has emerged as mobile maritime networks. Even though mobile communication underwater is still unreliable, new applications and ideas have been formulated, such as using a swarm of robots to map underwater acoustic fields [17], or using robots to collect data from underwater sensor networks [18]. This marriage between two previously separate areas is mutually beneficial.…”

Underwater acoustic intensity field reconstruction by kriged compressive sensing

“…Geostatistical kriging estimators take spatial correlation structures into consideration, producing both interpolated mean field and also the field of variance [17]. For underwater acoustic communication channels, kriging techniques have been applied to construct spatial maps for data collected by boats or robot swarms [10,22]. The limitation of kriging is that it does not prevent overfitting, where high frequency spatial fluctuations due to uncertainty in the measurements may affect the performance of the map construction.…”