Dr Haoning Xi

Community energy management is critical for facilitating the transition towards sustainable and clean smart grids. Energy cooperation techniques with community shared energy storage should be developed to reduce the challenges of distributed energy resources' uncertain and variable nature to a reliable power system. The proposed coordinator-users model involves the coordinator for techno-economic-environment optimization to minimize the community energy cost and carbon dioxide emissions. The end-users, including consumers and prosumers, also can make self-driven decisions to reduce their energy costs and contribute to a low-carbon society. By complementing the model with demand response dissatisfaction cost to evaluate the willingness of the user to participate in the energy cooperation, the techno-economic-environment-satisfaction benefit can be optimized at the community and user level. The effectiveness and superior performance of the proposed model are evaluated with the real-world dataset. The improvements in economic and environmental benefits are significant.

Increasing traffic congestion and vehicular emissions have become a major public concern, which naturally leads to a bi-objective optimization problem, i.e., to minimize system total travel time and system total emissions. In this study, a road pricing scheme is proposed to manage the system's total travel time and total emissions simultaneously. Finally, the practical case study based on Chengdu in China is carried out to simulate a situation with the proposed road pricing. Simulation results evaluate the effectiveness of the road pricing scheme to manage the congestion and emissions.

The keep-right rule is widely implemented in traditional traffic systems, but it might not be the optimal one in intelligent vehicle-infrastructure cooperation systems (i-VICS). According to literature reviews, the free rule is more effective in i-VICS, due to its superior mutual information communication. This paper firstly analyzes the reason why keep-right rule is more popular in traditional traffic systems, and then compares the overtaking procedure under the keep-right rule in traditional traffic systems and free rule in i-VICS. Comparison results show that overtaking in i-VICS is more efficient. Furthermore, five improved alternatives considering speed constraints and vehicle types are proposed and simulated through VISSIM. After this, the simulation results are used to evaluate alternative rules in i-VICS through the TOPSIS method. Additional sensitivity analysis shows that rule 5 is more effective in i-VICS.

Governments all over the world have attached great importance to the management of long-distance transport drivers. According to the statistics, the most dangerous three behaviors of long-distance drivers are: fatigued driving, distracted driving, and unrestricted driving. In this paper, these three behaviors are defined as safety features of drivers, which are detected based on machine learning and image processing technology. For identity recognition of drivers, Eigenface, Fisherface, and LBPH algorithms are combined to achieve a recognition rate of 100%. Driving time of each driver is recorded automatically. For mobile phone detection of drivers, the HOG algorithm and SVM algorithm are combined to achieve the recognition rate of 80%. For unrestricted driving detection, Gray-level integral projection method is improved to raise the detection rate to 85%. Finally, the safety feature detection system for drivers is developed to ensure the safety of long-distance transport.

In one-way carsharing systems, striking a balance between vehicle supply and user demand across stations poses considerable operational challenges. While existing research on vehicle relocation, personnel dispatch, and trip pricing have shown effectiveness, they often struggle with the complexities of fluctuating and unpredictable demand and supply patterns in uncertain environments. This paper introduces a real-time relocation-dispatch-pricing (RDP) problem, within an evolving time-state-extended transportation network, to optimize vehicle relocation, personnel dispatch, and trip pricing in carsharing systems considering both demand and supply uncertainties. Furthermore, recognizing the critical role of future insights in real-time decision making and strategic adaptability, we propose a novel two-stage anticipatory-decision rolling horizon (ADRH) optimization framework where the first stage solves a real-time RDP problem to make actionable decisions with future supply and demand distributions, while also incorporating anticipatory guidance from the second stage. The proposed RDP problem under the ADRH framework is then formulated as a stochastic nonlinear programming (SNP) model. However, the state-of-the-art commercial solvers are inadequate for solving the proposed SNP model due to its solution complexity. Thus, we customize a hybrid parallel Lagrangian decomposition (HPLD) algorithm, which decomposes the RDP problem into manageable subproblems. Extensive numerical experiments using a real-world dataset demonstrate the computational efficiency of the HPLD algorithm and its ability to converge to a near-globally optimal solution. Sensitivity analyses are conducted focusing on parameters such as horizon length, fleet size, number of dispatchers, and demand elasticity. Numerical results show that the profits under the stochastic scenario are 18% higher than those under the deterministic scenario, indicating the significance of incorporating uncertain and future information into the operational decisions of carsharing systems.

The COVID-19 pandemic has severely disrupted travel behavior across diverse socio-economic areas, with a significant impact on transportation systems, public health, and the economy. As countries both recover and plan for future virus-driven stresses, it is crucial to identify the drivers of building travel behavior resilience, such as vaccination. Using an integrated dataset with over 150 million US county-level mobile device data from 01/01/2020 to 20/04/2021, we employ Bayesian structural time series (BSTS) models to infer the relative impact of the vaccination intervention on five types of travel behavior across Metropolitan, Micropolitan and Rural areas. Further, we develop partial least squares regression (PLSR) models to accurately estimate how COVID-19 vaccination rates, epidemiological indicators (i.e., COVID-19 incidence rates, death rates, and testing rates) and weather conditions (i.e., temperature, rain, and snow) would impact various travel behaviors across the diverse areas during the recovery period of the pandemic. The model results shed light on the positive role of vaccinations in fostering the recovery of travel behaviors and reveal the disparities in travel behavior resilience in response to vaccination rates, epidemiological indicators, and weather conditions across diverse areas. Our findings can offer evidential insights for policymakers, transport planners, and public health officials, guiding the development of equitable, sustainable, and resilient transportation systems prepared to adapt to future pandemics.

The COVID-19 pandemic has resulted in substantial negative impacts on social equity. To investigate transport inequities in communities with varying medical resources and COVID controlling measures during the COVID pandemic and to develop transport-related policies for the post-COVID-19 world, it is necessary to evaluate how the pandemic has affected travel behavior patterns in different socio-economic segments (SES). We first analyze the travel behavior change percentage due to COVID, e.g., increased working from home (WFH), decreased in-person shopping trips, decreased public transit trips, and canceled overnight trips of individuals with varying age, gender, education levels, and household income, based on the most recent US Household Pulse Survey census data during Aug 2020 ~ Dec 2021. We then quantify the impact of COVID-19 on travel behavior of different socio-economic segments, using integrated mobile device location data in the USA over the period 1 Jan 2020¿20 Apr 2021. Fixed-effect panel regression models are proposed to statistically estimate the impact of COVID monitoring measures and medical resources on travel behavior such as nonwork/work trips, travel miles, out-of-state trips, and the incidence of WFH for low SES and high SES. We find that as exposure to COVID increases, the number of trips, traveling miles, and overnight trips started to bounce back to pre-COVID levels, while the incidence of WFH remained relatively stable and did not tend to return to pre-COVID level. We find that the increase in new COVID cases has a significant impact on the number of work trips in the low SES but has little impact on the number of work trips in the high SES. We find that the fewer medical resources there are, the fewer mobility behavior changes that individuals in the low SES will undertake. The findings have implications for understanding the heterogeneous mobility response of individuals in different SES to various COVID waves and thus provide insights into the equitable transport governance and resiliency of the transport system in the "post-COVID" era.

Researches showed that it was difficult to achieve objectives of reducing congestion and emission simultaneously. Road congestion pricing can manage the traffic demand efficiently, and thus reduce congestion, but it may not decrease vehicular emissions. The objectives of this study are multi-class users: users with different value of time (multi-VOT users) and users with different vehicle types (multi-vehicle users). By establishing the biobjective optimization model considering congestion and emissions simultaneously, it is proved that the Paretoefficient link flow can be decentralized as multi-class user equilibrium flow pattern by an effective road pricing scheme, whatever time cost criterion or monetary cost is used to choose the route, and thus the congestion and emissions can be reduced simultaneously.