Existing robust filters under generalized nonstationary noises conditions are difficult to choose suitable prior parameters, this brief proposed a Pearson-type VII distribution with adaptive parameters selection based interacting multiple model (IMM) Kalman filter (IMMPVII KF). In the model conditional filtering process, both the one-step prediction and the measurement likelihood are modeled as Pearson-type VII distributions. They are decomposed into Gaussian-Gamma Hierarchies (GGH), which are then matched to the time-varying heavy-tailed properties of the noises by pre-selecting the sets of shape and rate parameters and the variational Bayesian (VB) technique. Finally, a new model probability update method for filter under non-Gaussian conditions is derived. Simulation results show that the filter proposed in this brief has better robustness and adaptability to generalized non-stationary noises than the existing filters.

Aerospace vehicle navigation systems are equipped with multi-source redundant navigation sensors. According to the characteristics of the above navigation system configuration, building a resilient navigation framework to improve the accuracy and robustness of the navigation system has become an urgent problem to be solved. In the existing integrated navigation methods, redundant information is only used for backup. So, it cannot use the redundant navigation information to improve the accuracy of the navigation system. In this paper, a resilient multi-source fusion integrated navigation method based on comprehensive information evaluation has been proposed by combining of qualitative analysis and quantitative analysis in information theory. Firstly, this paper proposes a multi-layer evaluation framework of redundant information and carries out quantitative analysis of redundant information with the information disorder analysis theory to improve the reliability of the navigation system. Secondly, a navigation output effectiveness evaluation system has been established to analyze the output of heterogeneous navigation subsystems qualitatively to improve the fusion accuracy. Finally, through the mutual correction of multi-level information evaluation results, the error decoupling between the output parameters of heterogeneous navigation sensors has been realized to improve the robustness of the system. The experimental results show that the method proposed in this paper can adaptively allocate and adjust the weight of navigation information at all levels, realize the “non-stop” work of the navigation system and enhance the resilient of the navigation architecture. The navigation accuracy is improved compared with the existing multi-source fusion algorithm, which reflects the reliability and robustness of this algorithm.

Deep neural networks have been widely used in detection tasks based on optical remote sensing images. However, in recent studies, deep neural networks have been shown to be vulnerable to adversarial examples. Adversarial examples are threatening in both the digital and physical domains. Specifically, they make it possible for adversarial examples to attack aerial remote sensing detection. To defend against adversarial attacks on aerial remote sensing detection, we propose a cascaded adversarial defense framework, which locates the adversarial patch according to its high frequency and saliency information in the gradient domain and removes it directly. The original image semantic and texture information is then restored by the image inpainting method. When combined with the random erasing algorithm, the robustness of detection is further improved. Our method is the first attempt to defend against adversarial examples in remote sensing detection. The experimental results show that our method is very effective in defending against real-world adversarial attacks. In particular, when using the YOLOv3 and YOLOv4 algorithms for robust detection of single-class targets, the AP60 of YOLOv3 and YOLOv4 only drop by 2.11% and 2.17%, respectively, under the adversarial example.

In this paper, a multi-prior mixture (MPM) distribution and arithmetic average (AA) fusion-based Student’s t filter (MPMAASTF) is proposed for the problem of non-stationary heavy-tailed noises filtering with only rough prior information available. Firstly, a MPM distribution is proposed to model the joint probability density function (PDF) for one-step prediction and measure likelihood, then the Kullback-Leibler Divergence (KLD) minimization technique and Jensen’s inequality are used to derive approximate posterior PDFs for states and parameters. A finite number of Student’s t PDF fusion algorithm based on AA fusion is developed to guarantee the closure of the filtering posterior PDF, and finally existing high precision moment matching technique is embedded. The proposed filter improves the filtering accuracy of non-stationary heavy-tailed noises by adaptive learning of the degrees of freedom (dof), the one-step prediction error scale matrix and the measurement noise scale matrix. Simulations validate the effectiveness and superiority of the proposed method.