The target tracking of nonlinear maneuvering radar in dense clutter environments is still an important but difficult problem to be solved effectively. Traditional solutions often rely on motion models and prior distributions. This paper presents a novel improved architecture of Kalman filter based on a recursive neural network, which combines the sequence learning of recurrent neural networks with the precise prediction of Kalman filter in an end‐to‐end manner. We employ three LSTM networks to model nonlinear motion equation, motion noise, and measurement noise, respectively, and learn their long‐term dependence from a large amount of training data. They are then applied to the prediction and update process of Kalman filter to calculate the estimated target state. Our approach is able to address the tracking problem of nonlinear maneuvering radar target online end‐to‐end and does not require the motion models and prior distributions. Experimental results show that our method is more effective and faster than the traditional methods and more accurate than the method with LSTM network alone.

As the limited communication spectrum can not meet the demand of the exponential growth of intelligent connected devices in the internet of things(IoT) and typical smart city applications, in this paper, we propose a tractable spectrum sensing method based on Rao detection over non-Gaussian noise, such as generalized Gaussian noise(GGN), Gaussian mixture noise(GMN) and symmetric alpha-stable distribution ( $\text{S}\alpha \text{S}$ ) noise, multi-path fading channels environment to alleviate the issue of spectrum scarcity. In this method, there are unknown parameters in the multi-path fading channels. When the probability density function (P.D.F.) of non-Gaussian noise has a closed-form expression, the spectrum sensing method based on Rao detection is used. Otherwise the P.D.F. for $\text{S}\alpha \text{S}$ noise is estimated firstly by using non-parametric kernel estimation method, which addresses the issue that $\text{S}\alpha \text{S}$ noise has no closed-form P.D.F. expression, and then the performance of spectrum sensing is derived based on the theory of Rao detection in multi-path fading channels over typical smart city applications. Simulation results show that the accuracy of estimated P.D.F. for $\text{S}\alpha \text{S}$ noise and the performance of spectrum sensing under different $\alpha $ values over indoor, outdoor, and vehicle fading channels environment.

Signal-to-noise ratio (SNR) estimation is essential in the unmanned aerial vehicle (UAV) orthogonal frequency division multiplexing (OFDM) system for getting accurate channel estimation. In this paper, we propose a novel non-data-aided (NDA) SNR estimation method for UAV OFDM system to overcome the carrier interference caused by the frequency offset. First, an absolute value series is achieved which is based on the sampled received sequence, where each sampling point is validated by the data length apart. Second, by dividing absolute value series into the different series according to the total length of symbol, we obtain an output series by stacking each part. Third, the root mean squares of noise power and total power are estimated by utilizing the maximum and minimum platform in the characteristic curve of the output series after the wavelet denoising. Simulation results show that the proposed method performs better than other methods, especially in the low synchronization precision, and it has low computation complexity.

This study addresses the challenge of the precise positioning of Unmanned Aerial Vehicles (UAVs) in enclosed environments, particularly the inadequacy of traditional positioning methods in complex and dynamic settings. The study proposes a fusion model, Incremental Kalman Filter-Particle Filter (IKF-PF), based on dynamic environment adaptation, combining Kalman Filter (KF) and Particle Filter (PF). The proposed model can enhance the UAVs’ positioning accuracy and robustness in confined environments while considering computational efficiency. This model dynamically adjusts filtering strategies, employing different sampling methods in line-of-sight (LOS) and non-line-of-sight (NLOS) environments. Meanwhile, it integrates the state estimation capability of KF with the probability density approximation ability of PF to achieve precise positioning. Experiments are conducted in the Robot Operating System (ROS) environment, using Gazebo to simulate real flight scenarios, and are validated with public datasets and real flight data. The results show that in LOS environments, the proposed IKF-PF fusion algorithm achieves a mean positioning error of 0.19 meters, significantly outperforming single-sensor methods and other fusion algorithms. In NLOS environments, it maintains a mean error of 0.21 meters, demonstrating good robustness. Moreover, the algorithm exhibits high stability across multiple experiments, with a mean standard deviation of 0.04 meters. Despite its higher computational complexity, the IKF-PF algorithm maintains satisfactory positioning accuracy under mild to moderate disturbances in dynamic environment adaptability tests. The KF and PF fusion model proposed here effectively improves the positioning accuracy and robustness of UAVs in enclosed environments, particularly showing remarkable advantages in complex and dynamic settings. This method provides an effective solution for the autonomous navigation of UAVs in Global Positioning System-restricted or complex environments. This study offers new insights into the precise positioning of UAVs in enclosed and complex environments, advancing the application of multi-sensor fusion technology in UAV autonomous flight.