With a wide scope to explore and harness the oceanic sources of interest, the field of underwater wireless sensor networks (UWSNs) is attracting a growing interest of researchers. Owing to the real-time remote data monitoring requirements, underwater acoustic sensor networks (UASNs) emerged as a preferred network to a great extent. In UASN, the limited availability and non-rechargeability of energy resources along with the relative inaccessibility of deployed sensor nodes for energy replenishments necessitated the evolution of several energy optimization techniques. Clustering is one such technique that increases system scalability and reduces energy consumption. Besides clustering, coverage and connectivity are two significant properties that decide the proper detection and communication of events of interest in UWSN due to unstable underwater environment. Underwater communication is also possible with non-acoustic communication techniques like radio frequency, magnetic induction, and underwater free-space optics. In this paper, we surveyed clustering, coverage, and connectivity issues of UASN and qualitatively compared their performance. Particularly, the impact of these non-conventional communication techniques on clustering, coverage, and connectivity aspects is demonstrated. Additionally, we highlighted some key open issues related to the UWSN. This paper provides a broad view of existing algorithms of clustering, coverage, and connectivity based on acoustic communication. It also provides a useful guidance to the researchers in UWSN from various other communication techniques' perspective.

With a wide scope for exploration and research, underwater wireless sensor network (UWSN) is a fast growing research area in current scenario. UWSNs need energy efficient designing approach, because underwater sensor nodes are battery driven. Also the deployed batteries cannot be easily recharged by non-conventional energy resources, like solar energies. Clustering is an effective technique to design an energy efficient UWSNs. Due to the sparse deployment of nodes and dynamic nature of the channel, the clustering characteristics of UWSNs are different from those of terrestrial wireless sensor networks. In this paper, we focused on optimal clustering for UWSNs which are compliant with any one of the acoustic, free space optical (FSO) and electromagnetic (EM) wave-based communication techniques. Besides, we proposed an energy dissipation model of sensor node for FSO and EM wave-based communication and compared with the contemporary energy dissipation model for acoustic-based communication. In particular, the suitability of the above three techniques for underwater communication is investigated and their performance is compared on the basis of energy consumption and optimal clustering.

The electroencephalogram (EEG)-based brain-computer interface (BCI) has potential applications in neuroscience and rehabilitation. It benefits a person with neurological impairment to communicate their thoughts to the external world without using any appendages. Decoding imagined speech had limited success, mainly because neural signals are weak and more variable than overt speech, hence challenging to decode by machine-learning (ML)-based algorithms. In recent years, deep learning with convolutional neural networks (CNN) has transformed computer vision and can perform pattern recognition better than traditional ML-based algorithms. The objective of this paper is to design a smoothed pseudo-Wigner-Ville distribution (SPWVD) and CNN-based automatic imagined speech recognition (AISR) system to recognize imagined words. This paper uses a publically available 64-channel EEG dataset, collected from 15 healthy subjects for three categories: long words, short words, and vowels. The EEG signals were transformed into time-frequency representation (TFR) using SPWVD, which are utilized as an input to the CNN such that the EEG dataset was identified and classified into binary and multiclass categories. Additionally, the CNN model was optimized using a recently developed Keras-tuner library to achieve optimal performance. The performance of the SPWVD-CNN-driven AISR system is evaluated using seven performance evaluation metrics: accuracy, recall, precision, Mathew’s correlation coefficient, Cohen’s kappa (κ), F1-score, and area-under-the-curve. It is found that the proposed system achieved the maximum classification accuracies of 94.82%, 94.26%, 94.68%, and 84.50% for long words, short-long words, short words, and vowels, respectively. The proposed AISR strengthens the possibility of using imagined speech recognition as a future BCI application.

The magnetic induction (MI)-based communications assume the applications, wherein standard electromagnetic (EM) wave and acoustic communications are not effective. The EM and acoustic-based communications do have high power loss, high propagation delays, very low data rates, and environment-dependent channel behavior. MI communications heavily depend upon the time varying magnetic field to convey information between transmitter and receiver. MI-based communications for non-conventional media (e.g., underwater or underground) exhibit several unique and promising qualities, such as constant channel behavior, negligible propagation delay, and large communications range. This paper surveys the current state-of-the-art magnetic induction-based communications for non-conventional media applications. The widely used communications techniques for non-conventional media are qualitatively compared. In particular, the advantages and the disadvantages of different communications techniques are analyzed in terms of different performance metrics. In addition, we highlight some important open issues related to the design of MI for non-conventional media. This survey aims to provide a useful guidance to the researchers in this area.

Brain-computer interface (BCI) systems are intended to provide a means of communication for both the healthy and those suffering from neurological disorders. Imagined speech conveys users intentions. This paper investigates the feasibility of spectral characteristics of the electroencephalogram (EEG) signals involved in imagined speech recognition. Eleven subjects were recruited to perform the speech imagination task. This paper analyses the spectral features for binary and multiclass classification of imagined words in six different frequency bands. 1D EEG signals were converted into time frequency representation plots using smoothed pseudo Wigner-Ville distribution and classified using a convolutional neural network. Additionally, the analysis was performed for subject-dependent, subject-independent, and leave-one-subject-out (LOSO) approaches along with the all data approach. The proposed method achieved promising results in the Gamma band with a binary classification accuracy of 82.04±2.45%, 81.66±4.93%, 78.97±3.12%, and 81.04±3.08% in all data, subject-dependent, subject-independent, and LOSO approaches, respectively, and multiclass classification accuracy of 51.44±3.55%, 50.20±1.35%, 49.93±1.72%, and 50.42±2.18% in all data, subject-dependent, subject-independent, and LOSO approaches, respectively. Finally, the multiclass scalability in decoding the imagined words is investigated by increasing the number of classes from two to fifteen. The study’s findings demonstrate that EEG-based imagined speech recognition using spectral analysis has the potential to be an effective tool for speech recognition in practical BCI applications. The contribution of this paper lies in developing an EEG-based automatic imagined speech recognition system that offers high accuracy and reliability while also providing a non-invasive method for speech recognition.