Partial-band noise jamming is an important countermeasure against frequency-hopping spread spectrum technology in wireless sensor networks, and the jamming resource allocation (JRA) problem involved is a high-dimensional combinatorial optimization and also an NP-hard problem. Moreover, the users can dynamically alter their communication status, such as changing the channel spectrum distributions of hopping sets, posing additional difficulties for JRA. This paper develops two methods to address above challenges. First, a deep reinforcement learning (DRL)-based method is proposed for efficient JRA optimization, with the jamming scheme of each jamming node decided sequentially by the policy neural network, and its parameters are updated through trust region policy optimization (TRPO) within a trust region to ensure stable and fast convergence. Second, we propose a meta-TRPO based method to improve the generalization capability of the policy network. After the meta-training process, it can update the meta policy network with just a few fine-tuning steps and quickly obtain a task-specific policy for the fresh task. Extensive simulation results show the proposed DRL-based method converges faster than other DRL methods. Additionally, the proposed meta-TRPO based method can rapidly adapt to unseen jamming tasks with only a small quantity of training trajectories.

The advancement of deep learning (DL) techniques has led to significant progress in Automatic Modulation Classification (AMC). However, most existing DL-based AMC methods require massive training samples, which are difficult to obtain in non-cooperative scenarios. The identification of modulation types under small sample conditions has become an increasingly urgent problem. In this paper, we present a novel few-shot AMC model named the Spatial Temporal Transductive Modulation Classifier (STTMC), which comprises two modules: a feature extraction module and a graph network module. The former is responsible for extracting diverse features through a spatiotemporal parallel network, while the latter facilitates transductive decision-making through a graph network that uses a closed-form solution. Notably, STTMC classifies a group of test signals simultaneously to increase stability of few-shot model with an episode training strategy. Experimental results on the RadioML.2018.01A and RadioML.2016.10A datasets demonstrate that the proposed method perform well in 3way-Kshot, 5way-Kshot and 10way-Kshot configurations. In particular, STTMC outperforms other existing AMC methods by a large margin.

Specific emitter identification (SEI) aims to distinguish different emitter individuals based on the subtle differences in the signals. The technology can be widely applied to various wireless communication systems. Existing individual identification methods often ignore the radio frequency (RF) fingerprint information carried by the waveforms and thus can be interfered by unreliable RF features. To alleviate these issues, we devise the deep bidirectional long short-term memory (DBi-LSTM) and the one-dimensional residual convolution network with dilated convolution and squeeze-and-excitation block (Conv-OrdsNet). We exploit the combination of DBi-LSTM and Conv-OrdsNet (CoBONet) to extract temporal structure features directly from baseband in-phase and quadrature (I/Q) samples. The proposed network is able to capture the fine-grained details of signals and combine different information extracted by two networks. Moreover, we propose a data augmentation method to solve the interference of unreliable features by randomly changing the values of noise, power, frequency offset (FO), and phase offset (PO). It is worth noting that our method only needs a short slice of a steady-state signal without complex preprocessing, which reduces the cost of acquisition and calculation. Extensive experiments show that our method can effectively extract reliable RF fingerprinting features from I/Q samples and the classification results are better than most existing methods.

Jamming against frequency-hopping spread spectrum (FHSS) in wireless sensor networks (WSNs) has been primarily investigated with the follower jamming mode. However, implementing follower jamming in practical applications encounters manifold challenges, such as stringent requirements on hardware performance, difficulties in attaining accurate synchronization with signals. Diverging from existing works, in this paper, we propose a novel partial-band noise jamming (PBNJ) decision-making algorithm based on asynchronous deep reinforcement learning, which can allocate central jamming frequency and bandwidth more efficiently in FHSS jamming. Firstly, we model the problem of allocating jamming resource of PBNJ to disrupt the FHSS communication in WSNs as a Markov decision process (MDP). Next, considering the interrelationship among decisions made by different jamming nodes (JNs), we construct a multi-step decision framework in a time-division manner, and the Long Short-Term Memory (LSTM) network is leveraged to fully extract decision features from historical data, capturing correlations between jamming strategies of the deployed JNs, and guides future jamming decisions and enhances collaboration among different JNs in jamming resources allocation. Furthermore, to accelerate the convergence, we adopt the asynchronous advantage actor-critic (A3C) algorithm to optimize the allocation of central jamming frequency and bandwidth of JNs, utilizing the architecture of multi-threaded parallel training, and update the actor network and critic network in an asynchronous gradient descent manner. Simulation results show that the proposed LSTM-A3C algorithm converges fast and outperforms various baselines in terms of the convergence speed, jamming success rate and the total reward.

This letter studies the problem of dynamic jamming power allocation (JPA) with incomplete sensing information in dynamic and unknown environments. Most existing studies assume that jammers have perfect sensing information, without the consideration of the reliability of sensors, leading to poor performance when facing incomplete sensing information. In response, we propose a robust intelligent JPA scheme and adopt a two-stage algorithm namely “offline training and online deploy" to guide the training and deployment process. The approach utilizes the observed values from spatially distributed sensor units as inputs and employs deep reinforcement learning (DRL) for jamming power decision-making. To handle the missing observation data, we designed a data completion module based on the generative adversarial networks (GAN) framework. In addition, we introduce the priority experience replay mechanism (PERM) and opponent modeling (OM) in the decision model to enhance the learning efficiency and decision accuracy of the network. Simulation results show that the proposed approach can achieve efficient jamming with incomplete information, and outperform conventional DRL-based approaches.

In the automatic driving 3D object detection task, the multi-modal fusion method can realize the complementarity of different modal information, especially the fusion algorithm of LiDAR point cloud and camera image has been widely used. Nowadays, most point cloud-image fusion methods employ external projection matrix to achieve data alignment. However, it's not reliable in practical environments, and whether the data can be accurately aligned will directly affect the final detection performance. To alleviate this problem, we propose a novel multi-modality fusion 3D object detection algorithm based on Transformer named GATR. In our method, we give up external projection matrices to achieve the alignment of two modalities but use the attention mechanism to adaptively fuse the features of the two modes. We propose the Fusion Encoder, which uses the detection results of the image as queries to match the range image of point cloud to obtain the fusion features. The Guided Aggregation Decoder is used to output prediction and classification. We also design methods such as spatial weight graphs and category embeddings in the decoder to accelerate model convergence. Experiments on the KITTI and nuScenes datasets demonstrate that our proposed algorithm achieves a competitive level compared to some state-of-the-art algorithms.