Multiple object tracking (MOT) focuses on modeling the relationship of detected objects among consecutive frames and merge them into different trajectories. MOT remains a challenging task as noisy and confusing detection results often hinder the final performance. Furthermore, most existing research are focusing on improving detection algorithms and association strategies. As such, we propose a novel framework that can effectively predict and mask-out the noisy and confusing detection results before associating the objects into trajectories. In particular, we formulate such "bad" detection results as a sequence of events and adopt the spatio-temporal point process to model such events. Traditionally, the occurrence rate in a point process is characterized by an explicitly defined intensity function, which depends on the prior knowledge of some specific tasks. Thus, designing a proper model is expensive and time-consuming, with also limited ability to generalize well. To tackle this problem, we adopt the convolutional recurrent neural network (conv-RNN) to instantiate the point process, where its intensity function is automatically modeled by the training data. Furthermore, we show that our method captures both temporal and spatial evolution, which is essential in modeling events for MOT. Experimental results demonstrate notable improvements in addressing noisy and confusing detection results in MOT data sets. An improved state-of-the-art performance is achieved by incorporating our baseline MOT algorithm with the spatio-temporal point process model.

Video instance segmentation is a challenging task that extends image instance segmentation to the video domain. Existing methods either rely only on single-frame information for the detection and segmentation subproblems or handle tracking as a separate post-processing step, which limit their capability to fully leverage and share useful spatial-temporal information for all the subproblems. In this paper, we propose a novel graph-neural-network (GNN) based method to handle the aforementioned limitation. Specifically, graph nodes representing instance features are used for detection and segmentation while graph edges representing instance relations are used for tracking. Both inter and intra-frame information is effectively propagated and shared via graph updates and all the subproblems (i.e. detection, segmentation and tracking) are jointly optimized in an unified framework. The performance of our method shows great improvement on the YoutubeVIS validation dataset compared to existing methods and achieves 36.5% AP with a ResNet-50 backbone, operating at 22 FPS.

The task of spatial-temporal action detection has attracted increasing researchers. Existing dominant methods solve this problem by relying on short-term information and dense serial-wise detection on each individual frames or clips. Despite their effectiveness, these methods showed inadequate use of long-term information and are prone to inefficiency. In this paper, we propose for the first time, an efficient framework that generates action tube proposals from video streams with a single forward pass in a sparse-to-dense manner. There are two key characteristics in this framework: (1) Both long-term and short-term sampled information are explicitly utilized in our spatio-temporal network, (2) A new dynamic feature sampling module (DTS) is designed to effectively approximate the tube output while keeping the system tractable. We evaluate the efficacy of our model on the UCF101-24, JHMDB-21 and UCFSports benchmark datasets, achieving promising results that are competitive to state-of-the-art methods. The proposed sparse-to-dense strategy rendered our framework about 7.6 times more efficient than the nearest competitor.

Although almost blank subframes (ABS) proposed in heterogeneous networks (HetNet) can enhance the performance of user equipments (UEs) in Pico-cell range expansion (CRE) area, it also significantly degrades the Macro-cell throughput. To address this issue, further-enhanced inter-cell interference coordination (FeICIC) scheme is considered in 3GPP Release 11, where low power ABS (LP-ABS) are adopted for the Macro-cell center region users to improve the Macro-cell throughput. However, LP-ABS power, Pico CRE bias and Pico base station (PBS) density will jointly affect on the system performance, which eventually deteriorates the network energy efficiency (EE) without careful configuration. In this paper, we first deduce the closed-form expression of network EE as a function of PBS density, Pico CRE bias and LP-ABS power based on stochastic geometry model. Then we provide Monte Carlo simulations to verify the accuracy of theoretical derivation of the network EE and analyze the impacts of these parameters on the network EE. The simulation results show that the reasonable PBS density, Pico CRE bias and LP-ABS power can improve the network EE obviously.

This paper studies the energy efficiency (EE) maximization for an orthogonal frequency division multiple access (OFDMA) downlink network aided by a relay station (RS) with subcarrier pairing. A highly flexible transmission protocol is considered, where each transmission is executed in two time slots. Every subcarrier in each slot can either be used in direct mode or be paired with a subcarrier in another slot to operate in relay mode. The resource allocation (RA) in such a network is highly complicated, because it has to determine the operation mode of subcarriers, the assignment of subcarriers to users, and the power allocation of the base station and RS. We first propose a mathematical description of the RA strategy. Then, a RA algorithm is derived to find the globally optimum RA to maximize the EE. Finally, we present extensive numerical results to show the impact of minimum required rate of the network, the user number, and the relay position on the maximum EE of the network.

Time allocation optimisation for a multi‐antenna wireless energy and information transfer system is studied in this study. The terminal first harvests the wireless energy from the base station (BS) during phase I, and then uses the harvested energy to power its information transmission during phase II. Realistic channel state information (CSI) assumptions are made where the CSI at the BS is obtained via training and feedback and is subject to channel estimation error and quantisation error. Therefore, in addition to the tradeoff between information and power transfer, there is also a tradeoff in determining the overheads of training and feedback. To gain insight into this problem, the authors characterise the asymptotic behaviour of the parameters in the large system limit where the number of transmit antennas N t and transmission block length L tot tend to infinity with fixed ratio L tot / N t . The optimal overheads of training and feedback within each phase are first derived to maximise the harvested power and the information transmission rate, respectively. The optimised time allocation between phase I and phase II is then obtained by maximising the bounds of the effective average information transmission rate. Finally, numerical results are presented to verify the proposed theorems.

Energy‐efficient wireless communication is important for wireless devices with a limited battery life and cannot be recharged. In this study, a bit allocation algorithm to minimise the total energy consumption for transmitting a bit successfully is proposed for a two‐way orthogonal frequency‐division multiplexing relay system, whilst considering the constraints of quality‐of‐service and total transmit power. Unlike existing bit allocation schemes, which maximise the energy efficiency (EE) by measuring ‘bits‐per‐Joule’ with fixed bidirectional total bit rates constraint and no power limitation, their scheme adapts the bidirectional total bit rates and their allocation on each subcarrier with a total transmit power constraint. To do so, they propose an idea to decompose the optimisation problem. The problem is solved in two general steps. The first step allocates the bit rates on each subcarrier when the total bit rate of each user is fixed. In the second step, the Lagrangian multipliers are used as the optimisation variants, and the dimension of the variant optimisation is reduced from 2 N to 2, where N is the number of subcarriers. They also prove that the optimal point is on the bounds of the feasible region, thus the optimal solution could be searched through the bounds.

In this paper, a novel resource allocation (RA) strategy is designed for the downlink of orthogonal frequency division multiple access networks employing practical modulation and coding under quality of service constraints and retransmission techniques. Compared with previous works, two basic concepts are combined together, namely: 1) taking the goodput (GP) as performance metric and 2) ensuring maximum fairness among users. Thus, the resulting RA maximizes the GP of the worst users, optimizing subcarrier allocation (SA), per-subcarrier power allocation (PA), and adaptation of modulation and coding (AMC) of the active users, yielding a nonlinear nonconvex mixed optimization problem (OP). The intrinsic demanding difficulty of the OP is tackled by iteratively and optimally solving the AMC, PA, and SA subproblems, devoting special care to the difficult nonlinear combinatorial SA-OP. First, the optimal (yet computationally complex) solution is found by applying the branch and bound method to the optimal SA solution found in the relaxed domain, and accordingly, it is taken as benchmark. Then, an innovative suboptimal yet efficient solution based on the metaheuristic ant colony optimization (ACO) framework is derived. The proposed RA strategy is corroborated by comprehensive simulations, showing improved performance even at the cost of affordable numerical complexity.

This paper considers a multi-cell OFDMA downlink system with several decode-and-forward (DF) relay stations (RSs) aiding the base station (BS) transmissions. The opportunistic DF protocol proposed in [1] is applied. The problem considered is the maximization of the system sum rate with a total power constraint in each cell. An iterative low-complexity resource allocation (RA) algorithm is proposed to optimize mode selection (decision whether relaying should be used or not and which relay), subcarrier assignment (MSSA) and power allocation (PA) alternatively. During the MSSA stage, instead of the original objective function, a lower bound is maximized so that the problem is decoupled into subproblems which can be solved in linear time. During the PA stage, an algorithm based on single condensation and Lagrange duality PA (SC-LDPA) is designed to optimize PA with the tentative MSSA results. Through numerical experiments, the convergence of the low-complexity algorithm (LCA) as well as its benefit compared with a centralized algorithm (CA) are illustrated.

A novel subcarrier-pair based opportunistic DF protocol is proposed for cooperative downlink OFDMA transmission aided by a decode-and-forward (DF) relay. Specifically, user message bits are transmitted in two consecutive equal-duration time slots. A subcarrier in the first slot can be paired with a subcarrier in the second slot for the DF relay-aided transmission to a user. In particular, the source and the relay can transmit simultaneously to implement beamforming at the subcarrier in the second slot for the relay-aided transmission. Each unpaired subcarrier in either the first or second slot is used by the source for direct transmission to a user without the relay's assistance. The sum rate maximized resource allocation (RA) problem is addressed for this protocol under a total power constraint. It is shown that the novel protocol leads to a maximum sum rate greater than or equal to that for a benchmark one, which does not allow the source to implement beamforming at the subcarrier in the second slot for the relay-aided transmission. Then, a polynomial-complexity RA algorithm is developed to find an (at least approximately) optimum resource allocation (i.e., source/relay power, subcarrier pairing and assignment to users) for either the proposed or benchmark protocol. Numerical experiments illustrate that the novel protocol can lead to a greater sum rate than the benchmark one.

This paper investigates the optimum energy efficiency (EE) and the corresponding spectral efficiency (SE) for a communication link operating over a flat-fading channel. The EE is evaluated by the total energy consumption for transmitting per message bit. Three channel cases are considered, namely static channel with channel state information available at transmitter (CSIT), fast-varying (FV) channel with channel distribution information available at transmitter (CDIT), and FV channel with CSIT. The link's circuit power is modeled as ρ+κφ(R) Watt, where ρ>0 and κ≥0 are two constants and φ(R) is a general increasing and convex function of the transmission rate R≥0. For all the three channel cases, the tradeoff between the EE and SE is studied. It is shown that the EE improves strictly as the SE increases from 0 to the optimum SE, and then strictly degrades as the SE increases beyond the optimum SE. The impact of κ, ρ and other system parameters on the optimum EE and corresponding SE is investigated to obtain insight. Some of the important and interesting results for all the channel cases include: (1) when κ increases the SE corresponding to the optimum EE should keep unchanged if φ(R)=R, but reduced if φ(R) is strictly convex of R; (2) when the rate-independent circuit power ρ increases, the SE corresponding to the optimum EE has to be increased. A polynomial-complexity algorithm is developed with the bisection method to find the optimum SE. The insight is corroborated and the optimum EE for the three cases are compared by simulation results.

This paper contains two parts. The first part presents a novel framework for the successive convex approximation (SCA) method to solve a general optimization problem, as well as its properties. This framework starts with making change of variables (COV), motivated by the fact that it might be easier to construct convex approximations for the problem after making the COV. Furthermore, a general method is proposed to construct a convex upper bound approximation (CUBA) for a nonconvex function that satisfies tightness and differentiation conditions. Moreover, a way is introduced to generalize that CUBA by incorporating a convex function. These methods lead to plenty of degrees of freedom for using the SCA method to solve a problem. The second part revisits state-of-the-art dynamic spectrum management (DSM) algorithms, namely the successive convex approximations for low-complexity (SCALE) algorithm, the convex approximation for distributed spectrum balancing (CA-DSB) algorithm and the difference-of-convex-functions algorithm based DSM (DCA-DSM) method, to show how they can be derived from the SCA and CUBA construction methods. Numerical experiments are shown to compare them.

This paper considers a multi-cell relay aided orthogonal frequency division multiple access (OFDMA) downlink system, in which all stations are coordinated by a central controller for resource allocation (RA). The decode-and-forward (DF) protocol with selection relaying (SR) is applied. The problem considered is the maximization of the weighted sum of per cell min-rate (WSMR) with a total power constraint in each cell. An iterative RA algorithm is proposed to optimize mode selection (decision whether the relay should help or not), subcarrier assignment (MSSA) and power allocation (PA) alternatively. Each iteration is composed of the MSSA stage and the PA stage. During the MSSA stage, instead of the original objective function, a lower bound is maximized leading to lower complexity. The lower complexity problem is decoupled into mixed integer linear programs (MILP) that can easily be solved by typical MILP solvers. During the PA stage, an algorithm based on single condensation and geometric programming PA (SC-GPPA) is designed to optimize PA with the tentative MSSA results. The convergence of the proposed RA algorithm is proven. Finally, the performance of the RA algorithm and the benefit of using SR are illustrated through numerical experiments.

This study proposes a novel scheme for locating an asynchronous sensor with synchronised beacons using time-of-arrival (TOA) estimates. In this scheme, a reference beacon initiates the localisation by emitting a ranging signal, and the sensor position is estimated by measuring the length of the signal propagation path from the reference beacon to every beacon via the sensor. The associated Cramér–Rao bound (CRB) is derived, and a localisation algorithm, which does not need a priori knowledge about the TOA estimation accuracy, is proposed based on the constrained least square (LS) principle. It is illustrated by numerical experiments that the proposed scheme features a lower CRB compared to the conventional time-difference-of-arrival (TDOA)-based scheme, and the performance of the proposed algorithm is close to the CRB, especially when the sensor is in the centre area surrounded by the beacons.

In relay-aided wireless transmission systems, one of the key issues is how to manage the energy resource at the source and each individual relay, to optimize a certain performance metric. This paper addresses the sum rate maximized resource allocation (RA) problem in an orthogonal frequency division modulation (OFDM) transmission system assisted by multiple decode-and-forward (DF) relays, subject to the individual sum power constraints of the source and the relays. In particular, the transmission at each subcarrier can be in either the direct mode without any relay assisting, or the relay-aided mode with one or several relays assisting. We propose two RA algorithms which optimize the assignment of transmission mode and source power for every subcarrier, as well as the assisting relays and the power allocation to them for every {relay-aided} subcarrier. First, it is shown that the considered RA problem has zero Lagrangian duality gap when there is a big number of subcarriers. In this case, a duality based algorithm that finds a globally optimum RA is developed. Most interestingly, the sensitivity analysis in convex optimization theory is used to derive a closed-form optimum solution to a related convex optimization problem, for which the method based on the Karush-Kuhn-Tucker (KKT) conditions is not applicable. Second, a coordinate-ascent based iterative algorithm, which finds a suboptimum RA but is always applicable regardless of the duality gap of the RA problem, is developed. The effectiveness of these algorithms has been illustrated by numerical experiments.

This paper considers the downlink of a cellular orthogonal frequency division multi-access (OFDMA) system, in which multiple base stations (BS) are coordinated by a centralized resource allocation algorithm. We address the problem of maximizing the weighted sum min-rate (WSMR) subject to a total power constraint at each BS, in terms of jointly optimizing coordinated BSs' subcarrier and power allocation. In particular, this problem leads to a resource allocation that guarantees similar rates to users in each cell. An iterative algorithm is proposed to optimize subcarrier and power allocation alternatively, so that the WSMR keeps increasing until convergence. In each iteration, the subcarrier allocation is updated by solving a mixed integer linear programming problem for each cell, while the power allocation is updated by solving a successive set of convex optimization problems with an algorithm based on Karush-KuhnTucker conditions. The effectiveness of the algorithm is illustrated by numerical experiments.