Unmanned aerial vehicle (UAV)-assisted multi-access edge computing (MEC) networks significantly improve Internet of Things (IoT) communication in infrastructure-limited environments. However, in large-scale, complex, and dynamic 3D environments, uncertain sensor information and unexpected obstacles pose serious challenges to UAV flight safety, thereby necessitating robust and fast-responding trajectory optimization methods. This study presents a hierarchical trajectory generation and optimization scheme for UAV-assisted MEC in complex environments. Different from existing works, our scheme incorporates constraints from the UAV’s realistic dynamics model and a communication-computation energy consumption model. A global trajectory is first generated using an improved cost-function-based rapidly-exploring random tree star (RRT*) algorithm, followed by refinement via nonlinear model predictive control for high-precision smoothing. In addition, an attention-enhanced double deep Q-network (DDQN) module is employed as the high-level controller within the hierarchical trajectory optimization framework, providing optimized obstacle avoidance strategies in the presence of uncertain obstacles. Simulation results show that: 1) Compared with traditional approaches, the proposed scheme generates trajectories that demonstrate improved smoothness while rigorously satisfying physical and safety constraints, thereby enhancing the overall reliability of flight operations; and 2) in scenarios involving uncertain obstacles, the proposed framework effectively balances multiple objectives, including flight efficiency, obstacle avoidance safety, and energy consumption, leading to more robust and cost-effective aerial operation performance.

Connected autonomous driving and vehicle-toeverything (V2X) technology brings new opportunities for precise perception in occluded and sight-limited environments. The emergence of cooperative perception via V2X interaction has a positive impact on the safe and efficient driving of connected autonomous vehicles (CAVs). However, faced with the diversity and heterogeneity of V2X data, how to effectively process it to achieve better cooperative perception is a key and challenging task. This paper introduces a heterogeneous multiscale cooperative perception (HM-CoPept) framework with bird’s eye view features. It pays more attention to crucial cooperative data that affects driving to avoid blind areas and extend the sensing range. For the heterogeneity of V2X data, a bidirectional crossattention is innovatively proposed to fuse LiDAR and camera data complementary. Furthermore, the multiscale cooperation of V2X interaction data is proposed to break perception occlusion and limitation considering the benefit of multiscale features with different spatial importance. Cooperative perception is enhanced by learnable spatial confidence weight of safety distance constraint and foreground estimation. Test and validation are conducted on standard benchmarks, simulated (OPV2V) and real (DAIRV2X). HM-CoPept results show that performance increases by more than 16% compared to single-vehicle perception. Through extensive experiments and critical analysis, we demonstrate that our approach advances competitive methods and state-of-the-art in average precision. HM-CoPept enables precise and broad perception in complex driving environments and promotes the intelligent and autonomous development of CAVs.

The emergence of virtual coupling (VC) technology has the potential to substantially enhance the capacity of existing railway infrastructure. However, the complexity of line conditions introduces considerable disturbances to convoy operations, negatively affecting both energy consumption and operational efficiency. To address this issue, this study proposes a double loop trajectory optimization method for high-speed trains in a convoy, incorporating line condition disturbances and train dynamic characteristics. The proposed approach begins with an analysis of operating sequences under varying line conditions, leading to the development of a multi-resolution sequence optimization model for the leading train (LT). Subsequently, a train-following model is introduced to define the acceleration adjustment rules for the following train (FT) in different operating state. On this basis, a cooperative optimization framework is established to integrate the above models and define the detailed optimization procedure. Finally, a double loop seeker optimization algorithm is developed to obtain the optimal solution for the proposed model. Numerical experiments using field data from the Wuhan-Guangzhou high-speed railway line demonstrate the effectiveness of the proposed method. The experimental result proves that our method can generate trajectories with superior energy-saving and time-efficient performance while consistently maintaining safe and stable separation between virtually coupled trains.

Perception serves as the vital cornerstone of autonomous driving system, influencing the decision-making and control performance of vehicles. The rich semantic color information of images, the low cost of cameras and the support of deep learning make visual perception play a pivotal role. However, there are occlusions and blind areas when capturing data using only the on-board camera. With the development of vehicle-to-everything (V2X), information interaction can be achieved based on vehicle-to-vehicle (V2V), cooperative perception of connected autonomous vehicles (CAVs) based on information interaction has become a new trend. This study delves into visual perception based on Transformer attention, and enhances the encoder-decoder through multi-scale feature extraction and queries initialization. Furthermore, a visual cooperative perception method driven by V2V interaction is proposed. Based on spatial registration, data association and multi-source cooperation, perception enhancement of far-sight and see-through is achieved. Experiments were conducted on the real-world dataset and the PreScan simulator, evaluating the proposed method under various traffic state and density scenarios. Experimental results demonstrate that visual cooperative perception can improve the perception effect of CAVs and adapt to more complex traffic environments.

The Virtual coupling is proposed as a blocking mode to address the increasing demand for railway transport capacity, which takes operation efficiency further improve by separating trains with a relative braking distance. Nevertheless, an insufficient protection for complete avoidance safety risks in time with limited and fluctuate spacing separation between consecutive trains is introduced. A time-space occupancy band model is established to hold a safety protection in time-space dimension and assess the transport capacity for train operation under virtual coupling. In addition, a train trajectory planning method aimed at improvement of transport capacity, is proposed as a two-step program consisting of train followed operation trajectory planning based on Markov Decision Process and a trajectory multi-objective optimization for train convoy. In order to meet the requirement of the train trajectory dynamic adjustment under disturbance, an approach based on trajectory strategy set is designed by two stages to consider objectives of safety and punctuality. Based on the field data from the Wuhan-Guangzhou high-speed railway line, numerical experiments are conducted to validate the applicability of the proposed model and method. A comparative analysis of the track resource occupancy for several application condition under virtual coupling, and signaling systems is provided. The results indicate that the effective performance of proposed method in terms of trains separation and track resource occupancy, and show that virtual coupling is a more satisfactory blocking mode that could provide a higher track resource utilization while operation conditions are taken into account.

Virtual coupling is proposed as an innovative solution to meet the growing transport demand and to further improve the service quality of railways. Nevertheless, obtaining the optimal driving strategy that enhances its transport capacity and energy efficiency remains a challenging task. In order to achieve these objectives, this paper presents a novel convoy optimization method that optimizes the recommended trajectories for virtually coupled trains. A resilience adjustment model is firstly proposed to evaluate the coupling process and generate candidate trajectories according to the adjustment rules. In addition, the convoy optimization problem is formulated as a two-stage programming model consisting of a multi-objective programming stage and a least-cost goal programming stage, which determine the optimal trajectories for trains. Taking into account the requirements of practical applications, the solution method is finally designed to solve the proposed model and achieve dynamic updates of the recommended trajectories throughout train operations. Based on the field data from the Wuhan-Guangzhou high-speed railway line, numerical experiments are conducted to validate the effectiveness of the proposed method. The experimental results indicate that the proposed method shows the best performance in terms of infrastructure utilization and energy consumption, and the spacing between virtually coupled trains is well maintained regardless of ideal or disturbing conditions.

Taking advantage of both vehicle-to-everything (V2X) communication and automated driving technology, connected and automated vehicles are quickly becoming one of the transformative solutions to many transportation problems. However, in a mixed traffic environment at signalized intersections, it is still a challenging task to improve overall throughput and energy efficiency considering the complexity and uncertainty in the traffic system. In this study, we proposed a hybrid reinforcement learning (HRL) framework which combines the rule-based strategy and the deep reinforcement learning (deep RL) to support connected eco-driving at signalized intersections in mixed traffic. Vision-perceptive methods are integrated with vehicle-to-infrastructure (V2I) communications to achieve higher mobility and energy efficiency in mixed connected traffic. The HRL framework has three components: a rule-based driving manager that operates the collaboration between the rule-based policies and the RL policy; a multi-stream neural network that extracts the hidden features of vision and V2I information; and a deep RL-based policy network that generate both longitudinal and lateral eco-driving actions. In order to evaluate our approach, we developed a Unity-based simulator and designed a mixed-traffic intersection scenario. Moreover, several baselines were implemented to compare with our new design, and numerical experiments were conducted to test the performance of the HRL model. The experiments show that our HRL method can reduce energy consumption by 12.70% and save 11.75% travel time when compared with a state-of-the-art model-based Eco-Driving approach.

Resilience adjustment refers to the generation of a control strategy by evaluating the interaction between related factors. Tracking intervals of the high-speed train platoon change dynamically, which directly influences the operation safety and efficiency, and constrains the train operation trajectories. In China, the tracking interval is getting shorter. To ensure safety and improve efficiency, we research a dynamic interval resilience adjustment strategy based on the moving block system. Firstly, the optimal offline operation strategy is obtained by solving the multi-objective optimization model with the improved gravitational search algorithm (I-GSA). The resilience adjustment mechanism is developed to evaluate the tracking interval and choose the appropriate driving strategy to adjust operation states based on the resilience tracking interval model. Then, we study the relation between operation strategy and departure interval, and a seeker optimization algorithm (SOA) is used to obtain the optimal departure intervals and driving strategies. Simulations are conducted based on the sections between Chibi North station and Changsha South station in Wuhan-Guangzhou high-speed railway. The results indicate that the total operation time decreased by 191s and the operation safety can be ensured at any time.

Trajectory planning plays a crucial role in train operation by providing with the authorized speed at each position. The traditional static train trajectory planning methods are always designed offline according to a preplanned timetable, and they ignored the uncertainties of parameters, resulted by line condition, resistance coefficient, and delay. These uncertain disturbances have not been considered adequately in previous studies. This paper deals with the dynamic optimal train trajectory planning problem with uncertainties. First, in order to identify uncertain resistance coefficients and calculate the dynamic limited speed, we present the optimization framework using onboard equipment such as a global navigation satellite system (GNSS) terminal, a power supply system, and a communication device to sample the real-time traffic information. Then, by taking the energy consumption and punctuality as objectives, we propose a moving horizon train trajectory planning optimization model with an adaptive weight allocation mechanism based on trip time error. The innovation of this paper lies not only in the establishment of a novel dynamic optimization model for train trajectory planning but also the strategy that combines real-time traffic information with the trajectory planning procedure. By contrast with most existing solutions, the proposed approach fully takes advantage of the real-time information and thus avoids the difficulties for modeling the uncertain coefficients for train trajectory planning. The efficiency of the proposed approach is illustrated by showing some numerical results of simulations with the infrastructure data from Beijing-Shanghai High-speed Railway of China.

A speed trajectory profile indicating the authorized train speed at each position can be used to guide the driver or the automatic train operation (ATO) system to operate the train more efficiently, which is the most important part of the Chinese Train Control System (CTCS) and will decide the safety and efficiency of train operation. The efforts produced by the train to follow the speed trajectory will directly affect the evaluation of train operation. This paper studies the optimization approach for the speed trajectory of high-speed train in a single section. First, we take the energy consumption as the measure of satisfaction of the railway company, and the trip time is being regarded as the passenger satisfaction criterion; then, we present optimal speed trajectory searching strategies under different track characteristics by dividing the section into some subsections according to different speed limitations. After that, we develop a multiobjective optimization model for the speed trajectory, which is subject to the constraints such as safety requirement, track profiles, passenger comfort, and the dynamic performance. For obtaining the Pareto frontier of train speed trajectory, which has equal satisfaction degree on all the objects, a hybrid evolutionary algorithm is designed and applied to solve the model based on the differential evolution and simulating annealing algorithms. By showing some numerical results of simulations, the efficiency of the proposed model and solution methodology is illustrated.

Train operation energy consumption occupies a large proportion of whole rail transport resource consumption. Aiming at improving energy utilization, current research is mainly about saving energy while security and time are limiting conditions. However, actual train operation is a complex process which has strict requirements on safety, energy consumption, precise parking and some other factors. This paper summarizes train control optimization research, analyzes train running characteristics and influential factors, and introduces train traction computing method. Multi-objective Particle Swarm Optimization with inertia weight algorithm is applied to study train operation optimization. With energy-saving, punctuality and precise parking being optimization goals, train operation optimization models and fitness evaluation functions are established. Through optimization, optimal operating condition sequence and corresponding condition conversion points are obtained. Effectiveness of the proposed algorithm is verified by offline simulations, and the results present good optimization performance.