RESUMO
Incident detection is fundamental for freeway management to reduce non-recurrent congestions and secondary incidents. Recently, machine learning technologies have made considerable progress in the incident detection field, but many still face challenges in uncertainty quantification due to the aleatoric uncertainty of traffic data and the epistemic uncertainty of model deviations. In this study, a Bayesian deep learning method was proposed for freeway incident detection with uncertainty quantification. A convolutional neural network variant was designed on a Bayesian framework, and mechanisms of Bayes by backpropagation and local reparameterization technics were used to update the weight of the proposed model. The predictive uncertainty of the proposed method was modeled jointly by integrating the aleatoric and epistemic uncertainty. The proposed model was tested on the PORTAL dataset and compared with four benchmark models: standard normal deviate, wavelet neural network, long-short term memory neural network, and convolutional neural network. The results show that the proposed model outperforms the baseline methods in terms of accuracy, detection rate and false alarm rate. Perturbation experiments were used to test the robustness of the model against noise. The results indicated that the aleatoric uncertainty of the model remained almost constant under different noise levels. The proposed method may benefit future studies on uncertainty quantification while using machine learning method in incident management and other fields in intelligent transportation systems.
Assuntos
Aprendizado Profundo , Acidentes de Trânsito , Teorema de Bayes , Humanos , Redes Neurais de Computação , IncertezaRESUMO
A series of Fe-doped Na3Mn2-xFex(P2O7)(PO4) (x = 0, 0.2, 0.4) (abbreviated as NMFP-0/NMFP-0.2/NMFP-0.4) compounds have been successfully prepared using the sol-gel method. The Rietveld refinement results indicate that single-phase Na3Mn2-xFex(P2O7)(PO4) with an orthorhombic structure can be obtained. Our results reveal that by controlling the raw materials, the molar ratio of the reactants, the stirring rate of the precursor, the annealing temperature rate, and the reaction time, the proportion of nanowires in the morphology increases as the Fe component rises, and the NMFP-0.4 nanowire-shaped compounds show the best electrochemical activity when used as a cathode material for SIBs. Additionally, its specific capacity is enhanced to â¼126 mA h g-1 in the first cycle when operated at 0.1 C and a working potential window of 1.8-4.3 V (vs. Na/Na+). The material can also be applied in lithium-ion batteries as an anode and achieves â¼600 mA h g-1 specific capacity at a current density of 0.1 C (1 C = 1000 mA g-1) in a working potential window of 0.01-3 V (vs. Li/Li+).
RESUMO
The electronic and magnetic properties of the two-dimensional Ti3C2MXenes have attracted a lot of interests due to its potential applications. In this paper, Ti3C2TxMXenes and Mn-doped Ti3C2TxMXenes are synthesized and investigated. The experimental data shows that Mn2+ions are homogeneously and randomly intercalated between Ti3C2sheets as function terminals, which increase the interlayer distance between Ti3C2sheets and offer a mass of uncoupled magnetic moment. The temperature dependence of the electric resistivity of both samples show similar complex behavior although the resistivity increases dramatically as Mn doping. The inter-flake variable range hopping (VRH) dominates the low temperature electric transport behavior, while the inter-flake thermally activated hopping as well as the metallic intra-flake transport competing with the inter-flake VRH play important roles at high temperature. The increasing resistivity of the Mn-doped sample could be attributed to the increase of the interlayer distance and the enhancement of the localization of the transport electrons after Mn2+ions intercalation.
