Machine-learning technology powers many aspects of modern society. Compared to the conventional machine learning techniques that were limited in processing natural data in the raw form, deep learning allows computational models to learn representations of data with multiple levels of abstraction. In this study, an improved deep learning model is proposed to explore the complex interactions among roadways, traffic, environmental elements, and traffic crashes. The proposed model includes two modules, an unsupervised feature learning module to identify functional network between the explanatory variables and the feature representations and a supervised fine tuning module to perform traffic crash prediction. To address the unobserved heterogeneity issues in the traffic crash prediction, a multivariate negative binomial (MVNB) model is embedding into the supervised fine tuning module as a regression layer. The proposed model was applied to the dataset that was collected from Knox County in Tennessee to validate the performances. The results indicate that the feature learning module identifies relational information between the explanatory variables and the feature representations, which reduces the dimensionality of the input and preserves the original information. The proposed model that includes the MVNB regression layer in the supervised fine tuning module can better account for differential distribution patterns in traffic crashes across injury severities and provides superior traffic crash predictions. The findings suggest that the proposed model is a superior alternative for traffic crash predictions and the average accuracy of the prediction that was measured by RMSD can be improved by 84.58% and 158.27% compared to the deep learning model without the regression layer and the SVM model, respectively.

Flight control technology of hypersonic vehicles is very difficult because hypersonic vehicles have many features, such as strong nonlinearity, variable coupling, parameter variation, and uncertainties. In this paper, based on the generic hypersonic vehicle longitudinal model, two flight control methods based on back-stepping and dynamic inversion control respectively are simulated and analyzed. For the back-stepping based control, the back-stepping method is introduced to design control inputs for subsystems of the flight model. For the dynamic inversion based control, an inversion system is obtained by using feed-back linearization, and then controllers are designed by using linear control methods. Simulation of the hypersonic vehicle longitudinal model and two control methods are carried out in this paper. The tracking control performances of velocity and altitude of the hypersonic vehicle are verified. Then the two control methods are analyzed and evaluated in terms of control performances and engineering practicability. The back-stepping based control has high response speed and precision, and it is very suitable for on-line control. The dynamic inversion based control has a stable dynamic performance and better portability.

An integrated predictive iterative learning control (IPILC) scheme for batch process is designed from a two-dimensional (2D) system point of view. The integrated control framework combines batch-wise ILC and time-wise model predictive control (MPC), referred as 2D-IPILC. In the trajectory tracking problem of batch process, the predictive model can be obtained based on the system response using 2D theory. The control profile in the current batch is updated by MPC, using a quadratic objective function defined over time horizon. The major advantages of the proposed design scheme are shown in the better tracking performance as well as faster convergence speed by taking into account the time-wise feedback control with-in the current batch. The simulation results demonstrate the effectiveness of the proposed scheme.

Global climate change poses a critical challenge to sustainable urban development. The construction of low-carbon transportation systems is therefore a core strategy for enhancing the sustainability of mega-city road networks. Combining the characteristics of urban road traffic networks, this paper establishes a method for vehicle trip segmentation and carbon emission estimation based on GPS trajectory data (5699 vehicles, Beijing, September 2019) and the COPERT emission model, analyzing the spatiotemporal distribution characteristics of vehicle emissions. By incorporating the Life Cycle Assessment (LCA) emissions of electric vehicles, this study proposes carbon reduction strategies based on stochastic selection and ranking-based optimization from two dimensions: road-segment and vehicle electrification. Simulation methods are employed to evaluate the effectiveness of different strategies, as well as road network carbon emissions, under four vehicle electrification structures: Pyramid, Inverted Pyramid, Olive, and Dumbbell. Results indicate that carbon emission intensity rises significantly due to traffic congestion during peak hours. Under the LCA framework, Battery Electric Vehicles (BEVs) and Plug-in Hybrid Electric Vehicles (PHEVs) show significantly lower emissions than traditional Internal Combustion Engine Vehicles (ICEVs). Under the specified scenario assumptions, the ranking-based optimization scheme is estimated to yield carbon reductions approximately 2 times (segment control) and 3 times (electrification) those of the stochastic selection scheme, respectively. The study concludes that integrating EV promotion policies with precise carbon reduction control strategies can effectively mitigate urban road network carbon emissions.