Clustering algorithms have a very wide range of applications on data analysis, such as machine learning, mining. However, sets often problems with unbalanced and non-spherical distribution. by fast search find density peaks (DPC) is density-based clustering algorithm which could identify clusters data. In real applications, this its variants are not effective for the division unevenly distributed clusters, because they only use one indicator (the distance neighbor points) to handle inner...

Traditional trajectory clustering algorithms mostly cluster trajectories from the perspective of temporal or spatial. However, with and spatial similarity may be semantically related, ignoring semantic information lead to unreasonable results. Besides, most existing cannot deal overlapping well. In this paper, we propose a based on improved label propagation core structure, which could better measure spaces network. Experimental results demonstrate that proposed algorithm can effectively...

This paper investigates sensor fault problems in time-delay systems with uncertain disturbances. Using the measurement equation, can be translated into state inputs. Subsequently, a cluster of residual generators is designed by employing space geometric method. The corresponding filter parameters are obtained based on approach, then using H <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">∞</sub> optimization technique reduce effects disturbance...

One of the major challenges in target tracking is inaccurate state estimation caused by motion uncertainty target. In this paper, a deep neural network-based method proposed to fit nonlinear mapping relationship between filtering state, measurement, and real achieve effective maneuvering trajectory. First, encoder used mine favorable information radar observation. Next, bidirectional long short-term memory network (BI-LSTM) memorize potential laws motion. Then, attention mechanism introduced...

It is difficult to determine the dynamic Bayesian network (DBN) parameters applied inference of intention in warship air defense environment. This paper considers time variables based on QMAP algorithm that uses both domain knowledge and a small number static samples for parameter learning, then we propose D-QMAP suitable DBN taking into account variables, which utilizes timing relationship between samples. The simulation carried out background defense, results are line with reality,...

In this paper, we propose an algorithm to solve the issue of actuator fault FDI combined system. The is based on space geometry method. Firstly, use division theory and projection operation unobservability subspace system matrix parameters. Then build a residual generator realize decoupling corresponding relation residuals faults in Finally, feasibility effectiveness are shown for by simulations.

This paper investigates sensor fault problems in Markov jump systems with uncertain disturbances. Using the measurement equation, faults can be translated into state inputs. Subsequently, a cluster of residual generators is designed by employing space geometric method. The corresponding filter parameters are obtained based on approach, then using H <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">∞</sub> optimization technique reduce effects...