Abstract Due to the explosive growth of communication devices and vehicles, the ever‐increasing communication data traffic is a challenge for mobile communication in the Internet of Vehicles (IoV). A practical solution is to offload the traffic from cellular networks to Wireless Local Area Network (WLAN), that is, WiFi, where the spectrum is license‐free. However, the coordination of unlicensed spectrum severely reduces the utilization rate on the spectrum. In this case, unlicensed networks assisted access is proposed to facilitate long‐term evolution technology in unlicensed spectrum (LTE‐U). To improve the data rate for LTE‐U and WLAN users, an allocation optimization scheme for the resources of the licensed and unlicensed spectrum is proposed. Accordingly, the data rate of LTE‐U users and WLAN users can be maximized by adjusting the transmit power and the frequency occupancy time ratio of the unlicensed spectrum for the small cell and WLAN users. Numerical results show that the proposed hybrid LTE networks with WLAN can improve the communication efficiency for the served users.

Object detection is an active study area in the field of computer vision and image understanding. In this paper, we propose an active annotation algorithm by addressing the detection of numerous and scattered objects in a view, e.g., hundreds of cells in microscopy images. In particular, object detection is implemented by classifying pixels into specific classes with graph-based semi-supervised learning and grouping neighboring pixels with the same label. Sample or seed selection is conducted based on a novel annotation criterion that minimizes the expected prediction error. The most informative samples are therefore annotated actively, which are subsequently propagated to the unlabeled samples via a pairwise affinity graph. Experimental results conducted on two real world datasets validate that our proposed scheme quickly reaches high quality results and reduces human efforts significantly.

In this paper, we propose a generalized analog network coding (GANC) scheme for a non-orthogonal multiple-access relay channel (MARC), where two sources transmit their information simultaneously to the destination with the help of a relay. In the GANC scheme, the relay receives interfered signals from the two sources and generates signals to be transmitted with a relay function. We focus on the design of the optimal relay function to achieve the minimum pair-wise error probability (PEP) of the system. Specifically, we first covert the relay function optimization problem to a transformation matrix (TM) design problem by presenting the received complex signals as signal matrices composed of real and imaginary parts. Then, we propose an optimization criteria, i.e.,maximizing the minimal squared Euclidean distance(MMSED), to improve the PEP performance, since the PEP is determined by the Euclidean distance of the received constellation at destination. Next, we prove that the MMSED can be equivalently converted to a convex problem by introducing an intermediate matrix. We solve this convex problem by using the Lagrangian method and obtain the closed-form expression of the optimal TM. We further improve the PEP performance by optimizing transmission power of the two sources. Simulation results show that the proposed GANC scheme has a better PEP performance compared to other alternative schemes.

Power scheduling is an important issue for energy harvesting systems. In this work, we study the power control policy for minimizing the weighted sum of the outage probabilities under a set of predetermined transmission rates over a finite horizon. This problem is challenging in that the objective function is non-convex. To make the analysis tractable, we apply the approximation at high signal-to-noise ratios and obtain a near-optimal offline solution. In the case of infinite battery capacity, we demonstrate that the allocated power has a piecewise structure, i.e., each power scheduling cycle should be divided into disjoint segments and the normalized power should remain constant within each segment. An iterative algorithm is developed to obtain the power solution. In the case of finite battery capacity, we show that the piecewise structure still holds true, and we develop a divide-and-conquer algorithm to recursively solve the power allocation problem. Finally, we obtain a simple online power control policy that is fairly robust to prediction errors of the harvested energy. Simulations demonstrate that the proposed power solution has better performance than other strategies such as best-effort, fixed-ratio and random allocation.

In this paper we propose a new power adaptive network coding (PANC) strategy for a non-orthogonal multiple-access relay channel (MARC), where two sources transmit their information simultaneously to the destination with the help of a relay. In contrast to the conventional XOR-based network coding (CXNC), the relay in PANC generates network coded symbols by considering the coefficients of the source-to-relay channels, and forwards each symbol with a pre-optimized power level. Next, we obtain the optimal power level by decomposing it as a multiplication of a power scaling factor and a power adaptation factor. We prove that with the power scaling factor at the relay, our PANC scheme can achieve a full diversity gain, i.e., an order of two diversity gain, while the CXNC can achieve only an order of one diversity gain. In addition, we optimize the power adaptation factor at the relay to minimize the SPER at the destination by considering of the relationship between SPER and minimum Euclidean distance of the received constellation, resulting in an improved coding gain. Simulation results show that the PANC scheme with power adaptation optimizations and power scaling factor design can achieve a full diversity, and obtain a much higher coding gain than other network coding schemes.

In this paper we propose a novel power adaptive network coding (PANC) for a non-orthogonal multiple-access relay channel (MARC), where two sources transmit their information simultaneously to the destination with the help of a relay. In contrast to the conventional XOR-based network coding (CXNC), the relay in PANC generates network coded symbols by considering the coefficients of the source-to-relay channels, and forwards each symbol with a pre-optimized power level. Specifically, by defining a symbol pair as two symbols from the two sources, we first derive the expression of symbol pair error rate (SPER) for the system. Noting that deriving the exact SPER are complex due to the irregularity of the decision regions caused by random channel coefficients, we propose a coordinate transform (CT) method on the received constellation to simplify the derivations of the SPER. Next, we obtain the optimal power level by decomposing it as a multiplication of a power scaling factor and a power adaptation factor. We prove that with the power scaling factor at the relay, our PANC scheme can achieve a full diversity gain, i.e., an order of two diversity gain, while the CXNC can achieve only an order of one diversity gain. In addition, we optimize the power adaptation factor at the relay to minimize the SPER at the destination by considering of the relationship between SPER and minimum Euclidean distance of the received constellation, resulting in an improved coding gain. Simulation results show that (1) the SPER derived based on our CT method can well approximate the exact SPER with a much lower complexity; (2) the PANC scheme with power adaptation optimizations and power scaling factor design can achieve a full diversity, and obtain a much higher coding gain than other network coding schemes.

In this paper, we study the power scheduling problem to optimize the outage probability with both equal and unequal transmission rates under energy harvesting constraints. In particular, we first model the average outage probability minimization problem over finite horizon, and then reformulate the problem by minimizing its upper bound, which converts the non-convex problem to a more efficient convex optimization problem. Next, we show the optimal power scheduling solution is related to the global minimum of an energy-and-time function which is defined as arithmetic average of harvested energy from current time slot to the arbitrary time slot in the future before the end of transmission. Based on such properties, we propose a Min-Average strategy which presents the optimal performance in upper bound of outage probability for equal and non-equal transmission rates. Simulation results show that, our power scheduling scheme achieves better outage performance compared to other alternative strategies, e.g., best-effort, fixed-saving and random saving strategies.

Over the past decades, a wide attention has been paid to crowd control and management in the intelligent video surveillance area. Among the tasks for automatic surveillance video analysis, crowd motion modeling lays a crucial foundation for numerous subsequent analysis but encounters many unsolved challenges due to occlusions among pedestrians, complicated motion patterns in crowded scenarios, etc. Addressing the unsolved challenges, the authors propose a novel spatio-temporal viscous fluid field to model crowd motion patterns by exploring both appearance of crowd behaviors and interaction among pedestrians. Large-scale crowd events are hereby recognized based on characteristics of the fluid field. First, a spatio-temporal variation matrix is proposed to measure the local fluctuation of video signals in both spatial and temporal domains. After that, eigenvalue analysis is applied on the matrix to extract the principal fluctuations resulting in an abstract fluid field. Interaction force is then explored based on shear force in viscous fluid, incorporating with the fluctuations to characterize motion properties of a crowd. The authors then construct a codebook by clustering neighboring pixels with similar spatio-temporal features, and consequently, crowd behaviors are recognized using the latent Dirichlet allocation model. The convincing results obtained from the experiments on published datasets demonstrate that the proposed method obtains high-quality results for large-scale crowd behavior perception in terms of both robustness and effectiveness.

Over the past decades, a wide attention has been paid to crowd control and management in intelligent video surveillance area. In this paper, the authors propose a novel spatiotemporal viscous fluid field to recognize large-scale crowd event with respect to both appearance and driven factor of crowd behavior. Firstly, a spatiotemporal variation matrix is proposed to exploit motion property of a crowd. In particular, the paper exploits characteristics of the matrix with eigenvalue decomposition algorithm and constructs an abstract fluid field to model the crowd motion pattern, which is denoted by spatiotemporal fluid field. Secondly, the paper proposes a spatiotemporal force field to exploit the interaction force between the pedestrians. Furthermore, the fluid and force field constructs a spatiotemporal viscous fluid field. Thirdly, after generating feature with bag of word model, the authors utilize latent Dirichlet allocation model to recognize crowd behavior. The experiments on PETS2009 and UMN datasets show that the proposed method has a better performance for large-scale crowd behavior perception in both robustness and effectiveness comparing with the conventional methods.

This paper considers a multiple-access relay channel (MARC) with two sources, one relay and one destination, where the relay decides what it transmits to the destination according to the outage condition of source-relay links. The outage probability and the approximate bit error rate (BER) are derived, which are shown to be in tight match with Monte-Carlo simulation results. Simulation results reveal that adaptive DF strategy yields better performance than the traditional fixed DF strategies.

Over the past decade, a wide attention has been paid to the crowd control and management in intelligent video surveillance area. This paper proposes a sparse spatiotemporal local binary pattern (SST-LBP) descriptor to extract the dynamic texture of the walking crowd with the application to crowd density estimation. Firstly, the sparse selected location is extracted, which is notably variant in temporal domain and scale invariant in spatial domain. Afterwards, considering the spatial and temporal symmetry, the authors propose a sparse spatiotemporal local binary pattern algorithm and utilize its statistical property to describe the crowd feature. Finally, the crowd features are classified into a range of density levels by adopting support vector machine. The experiments on real video show that the proposed SST-LBP method is effective and robust on the large-scale crowd density estimation. Compared with the other methods, the proposed method does not base on the premise that the background should be extracted perfectly, which is too complicated to implement in real surveillance.