Wireless acoustic communication is the typical physical-layer technology for underwater communication; however, video transmissions via acoustic waves are hard to accomplish as the acoustic waves suffer from attenuation, limited bandwidth, Doppler spreading, high propagation delay, high bit error rate, and time-varying channel. For these reasons, state-of-the-art acoustic communication solutions are still mostly focusing on enabling delay-tolerant, low-bandwidth/low-data-rate scalar data transmission or at best low-quality/low-resolution multimedia streaming in the order of few tens of Kilobits per second (Kbps). Hence, the objectives of this research program were: (1) To design novel communication solutions for robust, reliable, and high-data rate underwater multimedia streaming on the order of hundreds of Kbps; (2) To investigate the problem of integrating communication methods available in multiple environments on an innovative software-defined testbed architecture integrating Microelectromechanical (MEMS)-based Acoustic Vector Sensors (AVSs) to enable processing-intensive physical-layer functionalities as software-defined, but executed in hardware that can be reconfigured in real time by the user based on the Quality of Experience (QoE).

By exploiting multiple-antenna arrays and AVSs, in Task 1 a novel physical-layer solution was proposed to boost the data rate for underwater acoustic transmission so as to transfer high-resolution video underwater. By following a novel probabilistic approach, an efficient Medium Access Control (MAC) layer solution was designed to share reliably the space among the steered vehicles by using AVSs so as to reduce the acoustic interference. The quality of multimedia delivery was improved by applying a robust closed-loop hybrid Automatic Repeat Request (ARQ) coding technique based on the estimated angles of arrivals using AVSs. In Task 2, the SEANet G2 acoustic networking platform was modified to investigate the design and fabrication of a new class of miniaturized and integrated AVS arrays based on the Aluminum Nitride (AlN) piezoelectric MEMS technology. Finally, in Task 3, scenarios were defined to validate the ideas proposed in Task 1 using the Naviator, AVSs, and SEANet testbed; Task 2 was evaluated by integrating SEANet to a buoy and the Naviator along with AVSs to build a testbed for conducting video transmission experiments. Tasks 1 and 2 were also integrated by comparing the pros and cons of MEMs AVSs with Commercial Off-The-Shelf (COTS) AVSs and by evaluating the semi-autonomous human-in-the-loop features to enhance user QoE.

Specific achievements/outcomes accomplished during this project are summarized below.

Deep Learning Framework for Link Adaptation in Underwater Wireless Optical Communications. This framework is based on a novel deep-Recurrent Neural Network (RNN)-based architecture. The dataset generation methodology was also described under this framework.

UW-CTSM: Circular Time Shift Modulation for Underwater Acoustic Communications. The key feature of the proposed modulation is the utilization of the high autocorrelation of the Zero-Correlation-Zone (ZCZ) signals.

ASVTuw: Adaptive Scalable Video Transmission in Underwater Acoustic Multicast Networks. The advantages of the ASVTuw include selecting Modulation and Coding Schemes (MCSs) adaptively with Equal Error Protection (EEP) or Unequal Error Protection (UEP) by referring to the Channel State Information (CSI) based on ML, decoding the Scalable Video Coding (SVC) video adaptively according to users' video quality requirements, and saving resources by avoiding transmitting redundant SVC enhancement layers.

Deep Joint Source-Channel Coding for Underwater Image Transmission. We presented our data-driven scheme for Joint Source-Channel Coding (JSCC) in underwater acoustic channel using CNN-based feature extraction and a novel variable-length encoder and decoder design based on Recurrent NNs (RNNs). The variable-length encoder-decoder design has the potential to adapt to changing underwater channel depending on the feedback received from the receiver.

ML-based Joint Doppler Estimation and Compensation in Underwater Acoustic Communications. The proposed solution adopts the concept of the Decision Feedback Equalizer (DFE) model as well as ML-based computation acceleration to save power without sacrificing performance.

Spatial Modulation-based Orthogonal Signal Division Multiplexing for Underwater ACOMMS. The simulation results showed that the SM-OSDM achieved a lower BER than Multiple-Input Multiple-Output (MIMO)-OSDM with the same deployment of transducers and hydrophones and a higher spectral efficiency than Single-Input Single-Output (SISO)-OSDM, especially in high Doppler channels..

Full-Duplex Underwater Acoustic Communications via Self-Interference Cancellation in Space. The proposed technique utilizes underwater Acoustic Vector Sensors (AVSs) and a Phased Array Transducer (PAT) as well as an adaptive protocol. The protocol helps FD UWA communication establish a reliable communication with spatial SIC on both side of the link. By updating DOA from information acquired with AVS, the BeamFormer (BF) helps PAT adjust the steering angle to aim at the other side of the link, therefore transmitting intensive signals with the main lobe in beamforming.

High-resolution Data Acquisition and Joint Source-Channel Coding in Underwater IoT. We proposed a MOSFET-based encoding scheme to realize low-power and low-cost substrate sensors. Further, we proposed a correlation-based Hybrid Automatic Repeat Request (HARQ) to transfer data between digital surface buoys and the fusion center that leverages the correlation in the data to avoid costly packet retransmissions and thereby enable timely reconstruction of the phenomenon. 

Last Modified: 12/31/2022
Modified by: Dario Pompili