Received signal strength (RSS) based source localization method is popular due to its simplicity and low cost. However, this method is highly dependent on the propagation model whose parameters are not easy to be captured in practice. Although source localization problem under uncertain propagation parameters has been wildly investigated, most existing works are devoted to the single source localization case. Nevertheless, multiple co-channel sources may transmit simultaneously in practice. In this paper, we estimate the locations of multiple co-channel sources from the superimposed RSS observations, while jointly inferring the parametric propagation model. Specifically, we first model the multiple sources localization (MSL) problem as being parameterized by the unknown source locations and propagation parameters, and then reformulate it as a joint parametric dictionary learning (PDL) and sparse signal recovery (SSR) problem, which is solved under the framework of sparse Bayesian learning with iterative parametric dictionary approximation. Moreover, a fast iterative update strategy is developed for the proposed method to reduce the complexity from $\mathcal{O}(MN^2)$ to $\mathcal{O}(MN)$ for high-dimensional large-scale problems, and the Cramèr-Rao lower bound (CRLB) is derived to analyze the theoretical estimation error bound. Finally, some important properties of the assumed MSL model and the proposed algorithms, as well as the future research directions are discussed. Comparing with the state-of-the-art sparsity-based MSL algorithms as well as CRLB, extensive simulations show the importance of jointly inferring the source locations and the propagation parameters, and highlight the effectiveness and superiority of the proposed methods.

This letter addresses the localization problem of a directional source in non-line-of-sight (NLOS) environments. To estimate both the position and the orientation of the source, we propose an iterative algorithm that skillfully adopts the received-signal-strength (RSS) and time-of-arrival (TOA) measurements. The algorithm is derived by converting the non-convex estimating problem into a generalized trust region subproblem (GTRS). The converted problem could be efficiently solved by a bisection process. Considering that the directivity of the target could greatly affect its RSS measurements in different directions, a preliminary estimate for the target’s location is conducted using only the TOA measurements, based on which we then estimate the orientation of the source and the NLOS biases. The alternating process of the algorithm can converge quickly. Finally, the excellent performance of the proposed algorithm is demonstrated through simulations.

The vehicular ad hoc network (VANET) constitutes a key technology for realizing intelligent transportation services. However, VANET is characterized by diverse message types, complex security attributes of communication nodes, and rapid network topology changes. In this case, how to ensure safe, efficient, convenient, and comfortable message services for users has become a challenge that should not be ignored. To improve the flexibility of routing matching multiple message types in VANET, this paper proposes a secure intelligent message forwarding strategy based on deep reinforcement learning (DRL). The key supporting elements of the model in the strategy are reasonably designed in combination with the scenario, and sufficient training of the model is carried out by deep Q networks (DQN). In the strategy, the state space is composed of the distance between candidate and destination nodes, the security attribute of candidate nodes and the type of message to be sent. The node can adaptively select the routing scheme according to the complex state space. Simulation and analysis show that the proposed strategy has the advantages of fast convergence, well generalization ability, high transmission security, and low network delay. The strategy has flexible and rich service patterns and provides flexible security for VANET message services.

Flying ad hoc networks (FANETs), which are composed of autonomous flying vehicles, constitute an important supplement to satellite networks and terrestrial networks, and they are indispensable for many scenarios including emergency communication. Unfortunately, the routing therein is largely affected by rapid topology changes, frequent disconnection of links, and a high vehicle mobility. In this paper, an intelligent routing strategy based on deep reinforcement learning (DRL) is proposed, which is decentralized and takes into account the status of symmetrical nodes in two hops. In order to perceive the local dynamics of the network as comprehensively as possible, the location, moving speed, load degree, and link quality of the nodes are considered into the setting process of state elements in the method. The nodes can select the neighboring node adaptively according to the Q values calculated by the model obtained through the training of Deep Q-Networks. The simulation and analysis show that the proposed method possesses good convergence characteristics and has obviously better performance compared with several common methods.