Synthetic Aperture Radar (SAR) imagery has been widely used in many maritime applications due to its high resolution, wide coverage, and real-time monitoring characteristics. Nevertheless, the size of SAR images is significantly large for real-time application. In recent years, High-Performance Computing (HPC)-related methods have been used to improve the precision and detection rate of SAR imagery analysis. In this paper, motivated by the state-of-the-art real time object detection You Only Look Once version 3 (YOLOv3), an enhanced GPU-based deep learning method has been proposed, namely Accelereated-YOLOv3 (A-YOLOv3), to detect ships from the SAR images. A-YOLOv3 aims to reduce the computational time with relatively competitive detection accuracy by constructing a new architecture with less layers and channels. The proposed A-YOLOv3 architecture achieves Average Precision (AP) of 97.4% on the Expand Diversified SAR Ship Detection Dataset (EDSSDD).

This study develops a point of familiarity (POF)-based navigation system for people with mild cognitive impairment (PMCI) based on wandering detection. The proposed system has four main functionalities: optimal path planning, POF-based navigation, wandering detection, and remote route tracking. An improved A*algorithm which introduced the θ_WD approach is employed to plan an optimal path for the PMCI. As any wandering behavior is detected, the PMCI who holds a mobile phone will receive the pof data sent by the remote monitoring system for guiding to the point of familiarity (POF) to wait for rescue. The travelling status of the PMCI will be recorded by the remote route tracking unit. Experiment results show that the proposed system achieves the accuracy of 91.7% for wandering detection in the normal GPS environments. Furthermore, the POF-based navigation saves about 19% of the traveling time as compared to the MAP-based navigation approach. Therefore, the system has great potential in helping the PMCI to go home safely.

Using the multilevel concatenation, long block codes can be constructed from shorter component codes, resulting in much less decoding complexity. The component codes can also be constructed from multilevel concatenation so that a hierarchical multilevel concatenated structure is built. The decoding complexity of the hierarchical scheme can be further decreased.

ABSTRACT Orthogonal space–time block codes provide full diversity with a very simple decoding scheme. However, they do not provide much coding gain. For a given space–time block code, we combine several component codes in conjunction with set partitioning of the expanded signal constellation according to the coding gain distance (CGD) criterion. By providing proper interlevel coding between adjacent blocks, we can design an orthogonal space–time block code with high rate, large coding gain, and low decoding complexity. The error performance of an example code is compared with some codes in computer simulation. These codes are compared based on the situation of the same transmission rate, space diversity order, and state complexity of decoding trellis. Copyright © 2010 John Wiley & Sons, Ltd.

A new block coded modulation scheme with interblock memory is proposed. The proposed code construction is based on the use of single parity check codes to concatenate a set of K coded blocks. Simulation results show that the proposed scheme does attain a more efficient utilization of the available bandwidth and power.

Block-coded modulation with interblock memory (BCMIM) is a variation of block-coded modulation (BCM) which is designed for multilevel coding. By providing interblock memory between adjacent blocks, the coding rate of a BCMIM scheme can be increased without decreasing the minimum squared Euclidean distance (MSED) as compared to the original BCM. In an early version of BCMIM, interblock coding is provided only between the first two coding levels of adjacent blocks. In this paper, we design BCMIM with a more general form for which interblock coding can be introduced among many coding levels. In this way, we can further increase the coding rate of BCMIM without decreasing the MSED. We provide many examples to show the advantages of BCMIM with the general form. Most of the examples are designed based on multidimensional signal sets, since a multidimensional signal set can provide more coding levels than a two-dimensional (2-D) signal set.

Proposes a coded modulation scheme with interblock memory. The proposed scheme has a multi-stage coding structure. However, at some coding stage, the coding designs for every two adjacent blocks are related. Compared to coded modulation schemes for which each block is independent of the others, the introduction of interblock coding provides additional flexibility in the design. In this way, for the proposed scheme, one can achieve a large minimum Euclidean distance while the number of nearest neighbors is not too large. The authors also design a multi-stage decoding algorithm, which is optimum in each stage and has low decoding complexity. Three specific examples are provided to demonstrate the advantages of the proposed scheme.>

Proposes a new trellis coded modulation scheme which is constructed by combining a convolutional code and a variation of block coded modulation called block coded modulation with interblock memory. In the proposed scheme, each coded multidimensional signal point has a kind of two-fold dependency on other coded multidimensional signal points. In addition to the dependency on other coded multidimensional signal points described by the trellis of a convolutional code, each coded multidimensional signal point has another form of dependency on one previously coded multidimensional signal point. The additional dependency comes from the design of block coded modulation with interblock memory.>