Channel state information (CSI) feedback is important for multiple-input multiple-output (MIMO) wireless systems to achieve their capacity gain in frequency division duplex (FDD) mode. For massive MIMO systems, CSI feedback may consume too much bandwidth and degrade spectrum efficiency. This work proposes a learning-based CSI feedback framework based on limited feedback and bi-directional reciprocal channel characteristics. The massive MIMO base station exploits the available uplink CSI to help recovering the unknown downlink CSI from low rate user feedback. We propose two deep learning architectures, DualNet-MAG and DualNet-ABS, to significantly reduce the CSI feedback payload based on the multipath reciprocity. DualNet-MAG and DualNet-ABS can exploit the bi-directional correlation of the magnitude and the absolute value of real/imaginary parts of the CSI coefficients, respectively. Experiment results demonstrate that our architectures bring an obvious improvement compared with the downlink-based architecture.

Channel state information (CSI) reporting is important for multiple-input multiple-output (MIMO) wireless transceivers to achieve high capacity and energy efficiency in frequency division duplex (FDD) mode. CSI reporting for massive MIMO systems could consume large bandwidth and degrade spectrum efficiency. Deep learning (DL)-based CSI reporting integrated with channel characteristics has demonstrated success in improving CSI compression and recovery. To further improve the encoding efficiency of CSI feedback, we develop an efficient DL-based compression framework CQNet to jointly tackle CSI compression, codeword quantization, and recovery under the bandwidth constraint. CQNet is directly compatible with other DL-based CSI feedback works for further enhancement. We propose a more efficient quantization scheme in the radial coordinate by introducing a novel magnitude-adaptive phase quantization framework. Compared with traditional CSI reporting, CQNet demonstrates superior CSI feedback efficiency and better CSI reconstruction accuracy.

The development of Internet of Vehicles (IoVs) introduces more intrusions to the vulnerabilities of in-vehicle network. It is important to detect in-vehicle network anomaly for the purpose of driving safety. The previous studies either focus on one-dimension anomaly behavior of in-vehicle network or need to look into the semantics of in-vehicle network, which is neither effective nor efficient for vehicle pilot. In this paper, we propose a literal multi-dimensional anomaly detection approach using the distributed Long-Short-Term-Memory (LSTM) framework for in-vehicle network, especially the Control Area Network (CAN). The proposed approach only needs the literal binary CAN message instead of revealing the semantics of CAN message. To enhance the accuracy and efficiency of detection, it detects anomaly from both time and data dimension simultaneously by exploiting multi-task LSTM neural network on mobile edge. The extensive evaluation results show that the proposed anomaly detection achieves a satisfying accuracy of 90%. The detection speed is as fast as 0.61 milliseconds on mobile edge.

Massive multiple-input multiple-output (MIMO) wireless communication technology with the characteristics of hyperconnectivity is an ideal channel to connect the industrial Internet of Things (IIoT) and the cyber physical system (CPS). It provides stable and reliable connectivity from the data center to distributed user terminals and the IIoT. However, traditional massive MIMO suffers from high power consumption and fabrication cost. The design of energy-efficient massive MIMO technology is essential for larger scale industrial deployments. We design three types of non-linear RF chain structures, which not only reduce the power consumption of massive MIMO systems but also save fabrication costs. Information theoretic analysis demonstrates the power efficiency performance of our non-linear system design. Our non-linear MIMO system designs can increase the power efficiency by up to 2.3 times compared with the traditional MIMO system. We have demonstrated that our systems can achieve the same uplink rate as traditional MIMO by increasing the number of receiving antennas but with less overall power consumption. We also proposed an algorithm to overcome the problem of low computational efficiency due to high-dimensional integration when calculating the uplink achievable rate of non-linear MIMO. Moreover, we reveal that when the skew-normal distribution is used as signaling, the non-linear MIMO systems can achieve better performance than the Gaussian distribution.

People counting is one of the hottest issues in sensing applications. Impulse radio ultrawideband radar has been extensively adopted to count people because it provides a device-free solution without illumination and privacy concerns. However, current solutions have limited performances in congested environments due to signal superpositions and obstructions. In this letter, a hybrid feature extraction method based on the curvelet transform and the distance bin is proposed. First, 2-D radar matrix features are extracted at multiple scales and multiple angles by applying the curvelet transform. Then, the distance bin concept is introduced by dividing each row of the matrix into several bins along the propagating distance to select features. A radar signal data set is constructed for three density scenarios, including people randomly walking in a constrained area at densities of three and four persons per square meter and people in a queue with an average between-person distance of 10 cm. The number of people in the data set scenarios varies from 0 to 20. Four classifiers--a decision tree, an AdaBoost classifier, a random forest, and a neural network--are compared to validate the hybrid features. The random forest achieves the highest accuracy of above 97% in the three density scenarios. To further investigate the reliability of the hybrid features, they are compared with three other features: cluster features, activity features, and features extracted by a convolutional neural network. The comparison results reveal that the proposed hybrid features are stable, and their performance is substantially more effective than that of the others.

Deep learning (DL)-based channel state information (CSI) feedback improves the capacity and energy efficiency of massive multiple-input multiple-output (MIMO) systems in frequency division duplexing mode. However, multiple neural networks with different lengths of feedback overhead are required by time-varying bandwidth resources. The storage space required at the user equipment (UE) and the base station (BS) for these models increases linearly with the number of models. In this paper, we propose a DL-based changeable-rate framework with novel quantization scheme to improve the efficiency and feasibility of CSI feedback systems. This framework can reutilize all the network layers to achieve overhead-changeable CSI feedback to optimize the storage efficiency at the UE and the BS sides. Designed quantizer in this framework can avoid the normalization and gradient problems faced by traditional quantization schemes. Specifically, we propose two DL-based changeable-rate CSI feedback networks CH-CsiNetPro and CH-DualNetSph by introducing a feedback overhead control unit. Then, a pluggable quantization block (PQB) is developed to further improve the encoding efficiency of CSI feedback in an end-to-end way. Compared with existing CSI feedback methods, the proposed framework saves the storage space by about 50% with changeable-rate scheme and improves the encoding efficiency with the quantization module.