Oversampling is a promising preprocessing technique for imbalanced datasets which generates new minority instances to balance the dataset. However, improper generated minority instances, i.e., noise instances, may interfere the learning of the classifier and impact it negatively. Given this, in this paper, we propose a simple and effective oversampling approach known as ASN-SMOTE based on the k-nearest neighbors and the synthetic minority oversampling technology (SMOTE). ASN-SMOTE first filters noise in the minority class by determining whether the nearest neighbor of each minority instance belongs to the minority or majority class. After that, ASN-SMOTE uses the nearest majority instance of each minority instance to effectively perceive the decision boundary, inside which the qualified minority instances are selected adaptively for each minority instance by the proposed adaptive neighbor selection scheme to synthesize new minority instance. To substantiate the effectiveness, ASN-SMOTE has been applied to three different classifiers and comprehensive experiments have been conducted on 24 imbalanced benchmark datasets. ASN-SMOTE is also extensively compared with nine notable oversampling algorithms. The results show that ASN-SMOTE achieves the best results in the majority of datasets. The ASN-SMOTE implementation is available at: https://www.github.com/yixinkai123/ASN-SMOTE/.

In this study, a novel prediction model for the number of vacant parking spaces after a specific period of time is proposed based on support vector regression (SVR) with fruit fly optimisation algorithm (FOA). In the proposed model, the SVR parameters are initialised as the fruit fly population, and FOA is utilised to search the optimal parameters for SVR. Sufficient experiments within various scenarios, i.e. predicting the vacant parking space availability in parking lots with various capacities after various periods of time, have been conducted to verify the effectiveness of the proposed FOA‐SVR prediction model. Three other commonly used prediction models, i.e. backpropagation neural network (NN), extreme learning machine and wavelet NN, are used as the comparison models. The experimental results show that the proposed FOA‐SVR method has higher accuracy and stability in all the prediction scenarios.

To cope with the tremendous growth in mobile data traffic, ultra-dense networks (UDNs) will be deployed in the future cellular network. Meanwhile, the enormous amount of available spectrum in millimeter-wave (mmWave) bands can be also exploited to increase the wireless link capacity. In accordance with these trends, this article introduces a novel hierarchical software-defined networking (SDN) architecture to facilitate the deployment of UDNs with both microwave base stations and mmWave access points. Aimed at reducing the cost of computational complexity and signaling overhead, the novel SDN architecture employs a semi-distributed design approach by configuring two levels of controllers: the centralized superior SDN controller and the localized subordinate SDN controller. Some practical challenges posed by the characteristics of mmWave transmissions, such as high propagation loss, channel blockage, and directional beamforming, are also discussed. To deal with such challenges, we then propose several important network operations for the hierarchical SDN, including dynamic subordinate SDN management, mobility management, and user-centric design. We finally provide simulation results to demonstrate the performance advantages of our proposals.

Integrated sensing and communication (ISAC) has emerged as a pivotal technology for next-generation mobile networks to embed sensing function on communication waveforms. A major challenge in ISAC is the effective integration of sensing and communication functions. Addressing this, this paper introduces a dual-functional waveform design that builds on the existing orthogonal frequency division multiplexing (OFDM) waveform. Unlike prior approaches that generally sacrifice communication performance to enhance sensing performance, our design contains a null-space sensing precoder that utilizes the null space of the communication channel to project additional sensing signals, thus improving the sensing functionality of the OFDM waveform without degrading any communication performance. We formulate a waveform optimization problem aimed at maximizing the sensing performance under the null-space sensing precoder and then propose a majorization-minimization (MM)-based waveform design algorithm. Additionally, to meet the real-time communication requirement in practice, we analyze the intrinsic characteristics of the high-performance sensing waveform and then develop a low-complexity waveform design algorithm. Simulation results show that the proposed MM-based algorithm can dramatically improve sensing performance without incurring any additional sensing power and degrading the communication performance. Furthermore, the low-complexity algorithm achieves substantial improvements in the sensing performance with much reduced computational complexity.