To eliminate the influence of excitation on wavelet packet frequency band energy spectrum (ES), ES is acquired via decomposition a virtual impulse response function. Based ES, character vector and damage eigenvector (DES) are created. Additionally, two identification indexes, ratio standard deviation variation coefficient, proposed. index, an updated method for retaining wall structures advanced. The state can be diagnosed through DES, location detected index trend surface, intensity...

To diagnose damages within the retaining wall structure, Hilbert marginal energy spectrum was acquired via Hilbert–Huang transformation of virtual impulse response functions responses to under ambient excitations. Based on spectrum, damage feature vector created. On basis a identification index proposed. and index, state detected by locations were diagnosed trend surface, intensity identified quantitative relationship between intensity. this, diagnosis method for structures verify...

To warn of the stability retaining wall structures with damage, a simplified mechanical model and finite element this wall-soil coupling system are established. Via updating, baseline is acquired. A damage alarming index ERSD (Energy Ratio Standard Deviation) proposed via wavelet packet analysis virtual impulse response function dynamic responses to model. The internal relationships among index, earth pressure, analyzed. Then, method for walls advanced. verify feasibility validity method,...

Cross-Camera vehicle tracking(CCVT) is an significant field in intellective transportation systems and intelligent visual perception smart city. It challenging due to the large variation of same within various cameras. We propose a cross-camera multi-target tracking framework towards traffic flow parameter estimation, which mainly divided into three steps:(i) multi-object single camera, (ii) multi-camera trajectory association and, (iii) estimation. Among them, we design encoder that...

To identify the damage within retaining wall structures, Hilbert–Huang Transforms of impulse response function and virtual were performed. The Hilbert marginal energy ratio spectrums acquired. reflect information effectively, those bands with stronger sensitivity extracted via threshold value ε 0 . Then, feature bands, which more sensitive to walls, selected by considering contribution residual band identification. Based on vector, reflects variations caused damage, was created. two...

Damage will appear in the wind-break wall under effects of many factors. The may be destroyed when damage is accumulated to some extent, which cause accidents. To identify within structure, Wavelet Packet frequency bands energy ratio spectrum analysis virtual impulse response function responses was performed excitation. Based on sensitivity subfrequency bands, a further updated proposed. reflect information sensitively, feature were more sensitive damage, selected via threshold value ε0....

In order to solve the problems of slow response and inaccurate control traditional ramp metering algorithm, we propose an intelligent model based on deep reinforcement learning. By building a neural network designing learning Double-Deep Q Network, on-ramp traffic flow can be realized real-time is ensured. The environment state space designed key parameters, incentive mechanism system established part data measure value action. JavaScript simulation platform built, update mode detectors...