The novelty of this WP is to produce important statistical models for the SI, which are at this point virtually unknown for UAC waveforms. In this WP, benefiting from the testing facilities at Newcastle and our industrial partner, we aim to produce accurate models that are key to the deeper understanding required for the design of SIC approaches for UAC and are not currently available in prior research work.

Task A.1 - Modelling of SI for FD systems:In this task, the statistical characteristics of the SI after passive and active cancellation will be modelled. The latter will include analogue and digital cancellation approaches. Despite recent advances in FD transceiver design, the impact of the interactions between cascaded analogue and digital cancellations is still unclear. Therefore, using experimental channels for various arrangements of transmit and receive transducers, we will develop a theoretical framework that provides system models for the linear and nonlinear residual SI components after each stage of cancellation. Probabilistic models of residual SI will be developed too for receiver performance analysis purposes.

Task A.2 - Modelling of hardware imperfections: In this task, hardware imperfections will be modelled and their impact on SIC investigated. We will model the effect of amplitude clipping and phase distortion effects resulting from analog-to-digital converter (ADC) saturation and investigate the effects of dynamic range and bit resolution related ADC limitations along with the effects of the non-flat frequency responses of commercially available transducers and power amplifiers.

Task A.3 – Modelling of time-varying, temporal and spatial effects of SI: In this task, the impact of the dynamics of the UAC channel will be considered under the FD mode assumption. SI will be modelled for different Doppler effects, and bottom and surface reflections will be taken into account as source for waveform interference. One and two-dimensional array configurations will be investigated by varying the positions of receive/transmit transducers in space, and their impact on achieving high natural isolation will be modelled and analysed.

The novelty of this WP is to provide SIC methods that are capable of achieving levels of cancellation up to 100 dB. Opportunities provided by innovative adaptive cancellation techniques developed at York will be explored, to attain a level of SIC in underwater acoustic transmission sufficient for true FD operation of acoustic modems.

Task B.1 – Waveform-domain (WD) cancellation methods: In this task, WD SIC methods that cancel SI before the ADC stage to avoid saturation will be investigated. 

Task B.2 – Active digital SIC methods: In this task, digital-domain (DD) SIC methods applied to residual SI to further improve WD SIC performance will be investigated.

Task B.3 – Beamforming methods: In this task, transmit and receive beamforming approaches will be explored, that can focus the emitted energy to the desired user’s direction while placing spatial nulls on the direction of the local receiver used for SIC.

The novelty of this WP is to design relaying techniques that utilize FD-enabled relays or hybrid HD/FD relay nodes. We will explore the expertise and techniques developed at NU for the EM-based wireless systems to design innovative relaying methods that can extend the communication range of current modems.

Task C.1 – Relaying Strategies: In this task, relay-enabled cooperative UAC will be investigated under the FD transmission mode assumption requiring critical re-consideration and re-design of relay strategies that presently exist in the wireless radio systems. These include: amplify and forward (AF), denoise and forward (DNF), physical-layer network coding (PLNC), compress and forward (CF), and quantize, map and forward (QMF). These methods will need to be integrated with the SIC developed in WP B to cope with the in-band FD interference. Furthermore, the FD principle will be utilized in conjunction with coded orthogonal frequency division multiplex (COFDM) waveforms, thus, the 2-timeslot approach for the uplink and downlink half-duplex transmissions will need to be replaced by a 1-timeslot approach involving the simultaneous transmission of the previously detected OFDM symbols and reception of the currently transmitted end-node OFDM symbols. The simulated and theoretical system performance of such relaying and physical layer techniques combined with FD is virtually unknown.

Task C.2 – MIMO and SIC-aided Relaying: This task will investigate multiple-input multiple-output (MIMO) aided SIC methods including natural isolation, antenna subset selection, spatial cancellation, joint transmitreceive beamforming and precoding. These methods will be investigated by taking into account the developments from WPs A and B, and by extending key findings to the multi-element case. Focus will be placed on the idiosyncrasies of the UAC waveforms which are wideband with respect to multipath delay and Doppler spread, and the underlying hardware imperfections, which are distinct and different from equivalent wireless radio components.

Task C.3 – Impact of CE and Syncronisation Effects: Although channel estimation (CE) and synchronisation methods for the UAC channel have been extensively studied and understood, most studies were conducted under the assumption of HD transmissions. CE will include approaches for COFDM and coded single carrier alternatives designed for the FD transmission mode, along with mechanisms of efficient channel state information (CSI) exchanges between the relay and the end nodes. We will investigate signal synchronisation methods that involve joint estimation and correction of motion-induced Doppler and hardware-related effects, such as frequency and carrier phase off-sets, and symbol and frame timing misalignments. They will be designed in conjunction with SIC and relaying methods and the impact of residual SI on their performance will need to be investigated.

The novelty of this WP is to devise MAC protocols which can exploit full duplex communication in order to maximise the capacity of UAC networks. Extensive understanding of the underwater acoustic communications channel will be used to engender MAC and logical link control methodologies that can effectively support simultaneous bi-directional communication between pairs of nodes.

Task D.1 Single-hop Networks: Centralised MAC protocols will be developed for single-hop communication scenarios, which exploit the capabilities of a central node (for example on a boat or tethered to a buoy) to assign channels and coordinate transmission and reception. The potential of employing strategies designed for geostationary satellite systems will be explored, to overcome the resource reservation constraints imposed by the long propagation delay.

Task D.2 - Multi-hop Networks: Distributed MAC protocols will be developed based on more advanced forms of ALOHA, as a more suitable approach for distributed multi-hop scenarios, with individual nodes determining appropriate frequency channels and transmission times. Reinforcement learning has been shown to provide efficient medium access in terrestrial multi-hop networks, with minimal signalling overheads compared with distributed scheduling.

In this WP, wet tests will be conducted in a highly controlled environment using NU’s acoustic anechoic water tank and a frequency-scaled channel model to obtain initial bit-error rate, outage capacity and throughput performances. Acoustic power amplifiers and transducers will be used to model SI, and the SIC methods will implemented in DSP/FPGA development boards to evaluate the performance of the proposed methods and refine the parameters of the developed SI models. Following this, using NU’s research vessel, full-scale sea trials will be conducted in the North Sea in realistic shallowwater channels exhibiting local reverberation at the transceiver and different transducer array configurations. In addition to testing the developed SIC algorithms in a realistic environment, signals will be acquired for further off-line data analysis and fine-tuning of the SI models and the SIC algorithms. Furthermore, the frequencyscaled simulation model used in the anechoic water tank will be honed. Ultimately,  the functionality of a fully integrated, FD-based, relay-enabled demonstrator transceiver will be tested and its performance will be verified. Our industrial partner, ATLAS ELEKTRONIK UK, will be also profoundly involved in this WP providing insight to practical deployment scenarios, hardware limitations, transducer array configurations and utilised transmission levels.