Evolutionary Decision-Making and Planning for Autonomous Driving Based on Safe and Rational Exploration and Exploitation

Decision-making and motion planning are extremely important in autonomous driving to ensure safe driving in a real-world environment.This study proposes an online evolutionary decision-making and motion planning framework for autonomous driving based on a hybrid data-and model-driven method.First,a data-driven decision-making module based on deep reinforcement learning(DRL)is developed to pursue a rational driving performance as much as possible.Then,model predictive control(MPC)is employed to execute both longitudinal and lateral motion planning tasks.Multiple constraints are defined according to the vehicle’s physical limit to meet the driving task requirements.Finally,two principles of safety and rationality for the self-evolution of autonomous driving are proposed.A motion envelope is established and embedded into a rational exploration and exploitation scheme,which filters out unreasonable experiences by masking unsafe actions so as to collect high-quality training data for the DRL agent.Experiments with a high-fidelity vehicle model and MATLAB/Simulink co-simulation environment are conducted,and the results show that the proposed online-evolution framework is able to generate safer,more rational,and more efficient driving action in a real-world environment.

Influences of Mixed Traffic Flow and Time Pressure on Mistake-Prone Driving Behaviors among Bus Drivers

Bus safety is a matter of great importance in many developing countries, with driving behaviors among bus drivers identified as a primary factor contributing to accidents. This concern is particularly amplified in mixed traffic flow (MTF) environments with time pressure (TP). However, there is a lack of sufficient research exploring the relationships among these factors. This study consists of two papers that aim to investigate the impact of MTF environments with TP on the driving behaviors of bus drivers. While the first paper focuses on violated driving behaviors, this particular paper delves into mistake-prone driving behaviors (MDB). To collect data on MDB, as well as perceptions of MTF and TP, a questionnaire survey was implemented among bus drivers. Factor analyses were employed to create new measurements for validating MDB in MTF environments. The study utilized partial correlation and linear regression analyses with the Bayesian Model Averaging (BMA) method to explore the relationships between MDB and MTF/TP. The results revealed a modified scale for MDB. Two MTF factors and two TP factors were found to be significantly associated with MDB. A high presence of motorcycles and dangerous interactions among vehicles were not found to be associated with MDB among bus drivers. However, bus drivers who perceived motorcyclists as aggressive, considered road users’ traffic habits as unsafe, and perceived bus routes’ punctuality and organization as very strict were more likely to exhibit MDB. Moreover, the results from the three MDB predictive models demonstrated a positive impact of bus route organization on MDB among bus drivers. The study also examined various relationships between the socio-demographic characteristics of bus drivers and MDB. These findings are of practical significance in developing interventions aimed at reducing MDB among bus drivers operating in MTF environments with TP.

Driving pressure in mechanical ventilation:A review

Driving pressure(ΔP)is a core therapeutic component of mechanical ventilation(MV).Varying levels ofΔP have been employed during MV depending on the type of underlying pathology and severity of injury.However,ΔP levels have also been shown to closely impact hard endpoints such as mortality.Considering this,conducting an in-depth review ofΔP as a unique,outcome-impacting therapeutic modality is extremely important.There is a need to understand the subtleties involved in making sureΔP levels are optimized to enhance outcomes and minimize harm.We performed this narrative review to further explore the various uses ofΔP,the different parameters that can affect its use,and how outcomes vary in different patient populations at different pressure levels.To better utilizeΔP in MV-requiring patients,additional large-scale clinical studies are needed.

Self-assembly of perovskite nanocrystals:From driving forces to applications

Self-assembly of metal halide perovskite nanocrystals(NCs)into superlattices can exhibit unique collective properties,which have significant application values in the display,detector,and solar cell field.This review discusses the driving forces behind the self-assembly process of perovskite NCs,and the commonly used self-assembly methods and different self-assembly structures are detailed.Subsequently,we summarize the collective optoelectronic properties and application areas of perovskite superlattice structures.Finally,we conclude with an outlook on the potential issues and future challenges in developing perovskite NCs.

Effects of drive imbalance on the particle emission from a Bose-Einstein condensate in a one-dimensional lattice

Time-periodic driving has been an effective tool in the field of nonequilibrium quantum dynamics,which enables precise control of the particle interactions.We investigate the collective emission of particles from a Bose-Einstein condensate in a one-dimensional lattice with periodic drives that are separate in modulation amplitudes and relative phases.In addition to the enhancement of particle emission,we find that amplitude imbalances lead to energy shift and band broadening,while typical relative phases may give rise to similar gaps.These results offer insights into the specific manipulations of nonequilibrium quantum systems with tone-varying drives.

Research on Driver’s Fatigue Detection Based on Information Fusion

Driving fatigue is a physiological phenomenon that often occurs during driving.After the driver enters a fatigued state,the attentionis lax,the response is slow,and the ability todeal with emergencies is significantly reduced,which can easily cause traffic accidents.Therefore,studying driver fatigue detectionmethods is significant in ensuring safe driving.However,the fatigue state of actual drivers is easily interfered with by the external environment(glasses and light),which leads to many problems,such as weak reliability of fatigue driving detection.Moreover,fatigue is a slow process,first manifested in physiological signals and then reflected in human face images.To improve the accuracy and stability of fatigue detection,this paper proposed a driver fatigue detection method based on image information and physiological information,designed a fatigue driving detection device,built a simulation driving experiment platform,and collected facial as well as physiological information of drivers during driving.Finally,the effectiveness of the fatigue detection method was evaluated.Eye movement feature parameters and physiological signal features of drivers’fatigue levels were extracted.The driver fatigue detection model was trained to classify fatigue and non-fatigue states based on the extracted features.Accuracy rates of the image,electroencephalogram(EEG),and blood oxygen signals were 86%,82%,and 71%,separately.Information fusion theory was presented to facilitate the fatigue detection effect;the fatigue features were fused using multiple kernel learning and typical correlation analysis methods to increase the detection accuracy to 94%.It can be seen that the fatigue driving detectionmethod based onmulti-source feature fusion effectively detected driver fatigue state,and the accuracy rate was higher than that of a single information source.In summary,fatigue drivingmonitoring has broad development prospects and can be used in traffic accident prevention and wearable driver fatigue recognition.

Controlled thermally-driven mass transport in carbon nanotubes using carbon hoops

Controlling mass transportation using intrinsic mechanisms is a challenging topic in nanotechnology.Herein,we employ molecular dynamics simulations to investigate the mass transport inside carbon nanotubes(CNT)with temperature gradients,specifically the effects of adding a static carbon hoop to the outside of a CNT on the transport of a nanomotor inside the CNT.We reveal that the underlying mechanism is the uneven potential energy created by the hoops,i.e.,the hoop outside the CNT forms potential energy barriers or wells that affect mass transport inside the CNT.This fundamental control of directional mass transportation may lead to promising routes for nanoscale actuation and energy conversion.

Land use change and its driving factors in the ecological function area:A case study in the Hedong Region of the Gansu Province,China

Land use and cover change(LUCC)is important for the provision of ecosystem services.An increasing number of recent studies link LUCC processes to ecosystem services and human well-being at different scales recently.However,the dynamic of land use and its drivers receive insufficient attention within ecological function areas,particularly in quantifying the dynamic roles of climate change and human activities on land use based on a long time series.This study utilizes geospatial analysis and geographical detectors to examine the temporal dynamics of land use patterns and their underlying drivers in the Hedong Region of the Gansu Province from 1990 to 2020.Results indicated that grassland,cropland,and forestland collectively accounted for approximately 99% of the total land area.Cropland initially increased and then decreased after 2000,while grassland decreased with fluctuations.In contrast,forestland and construction land were continuously expanded,with net growth areas of 6235.2 and 455.9 km^(2),respectively.From 1990 to 2020,cropland was converted to grassland,and both of them were converted to forestland as a whole.The expansion of construction land primarily originated from cropland.From 2000 to 2005,land use experienced intensified temporal dynamics and a shift of relatively active zones from the central to the southeastern region.Grain yield,economic factors,and precipitation were the major factors accounting for most land use changes.Climatic impacts on land use changes were stronger before 1995,succeeded by the impact of animal husbandry during 1995-2000,followed by the impacts of grain production and gross domestic product(GDP)after 2000.Moreover,agricultural and pastoral activities,coupled with climate change,exhibited stronger enhancement effects after 2000 through their interaction with population and economic factors.These patterns closely correlated with ecological restoration projects in China since 1999.This study implies the importance of synergy between human activity and climate change for optimizing land use via ecological patterns in the ecological function area.

Landscape ecological risk assessment and its driving factors in the Weihe River basin,China

Weihe River basin is of great significance to analyze the changes of land use pattern and landscape ecological risk and to improve the ecological basis of regional development.Based on land use data of the Weihe River basin in 2000,2010,and 2020,with the support of Aeronautical Reconnaissance Coverage Geographic Information System(ArcGIS),GeoDa,and other technologies,this study analyzed the spatial-temporal characteristics and driving factors of land use pattern and landscape ecological risk.Results showed that land use structure of the Weihe River basin has changed significantly,with the decrease of cropland and the increase of forest land and construction land.In the past 20 a,cropland has decreased by 7347.70 km2,and cropland was mainly converted into forest land,grassland,and construction land.The fragmentation and dispersion of ecological landscape pattern in the Weihe River basin were improved,and land use pattern became more concentrated.Meanwhile,landscape ecological risk of the Weihe River basin has been improved.Severe landscape ecological risk area decreased by 19,177.87 km2,high landscape ecological risk area decreased by 3904.35 km2,and moderate and low landscape ecological risk areas continued to increase.It is worth noting that landscape ecological risks in the upper reaches of the Weihe River basin are still relatively serious,especially in the contiguous areas of high ecological risk,such as Tianshui,Pingliang,Dingxi areas and some areas of Ningxia Hui Autonomous Region.Landscape ecological risk showed obvious spatial dependence,and high ecological risk area was concentrated.Among the driving factors,population density,precipitation,normalized difference vegetation index(NDVI),and their interactions are the most important factors affecting the landscape ecological risk of the Weihe River basin.The findings significantly contribute to our understanding of the ecological dynamics in the Weihe River basin,providing crucial insights for sustainable management in the region.

A Multiscale Reliability-Based Design Optimization Method for Carbon-Fiber-Reinforced Composite Drive Shafts

Carbon fiber composites,characterized by their high specific strength and low weight,are becoming increasingly crucial in automotive lightweighting.However,current research primarily emphasizes layer count and orientation,often neglecting the potential of microstructural design,constraints in the layup process,and performance reliability.This study,therefore,introduces a multiscale reliability-based design optimization method for carbon fiber-reinforced plastic(CFRP)drive shafts.Initially,parametric modeling of the microscale cell was performed,and its elastic performance parameters were predicted using two homogenization methods,examining the impact of fluctuations in microscale cell parameters on composite material performance.A finite element model of the CFRP drive shaft was then constructed,achieving parameter transfer between microscale and macroscale through Python programming.This enabled an investigation into the influence of both micro and macro design parameters on the CFRP drive shaft’s performance.The Multi-Objective Particle Swarm Optimization(MOPSO)algorithm was enhanced for particle generation and updating strategies,facilitating the resolution of multi-objective reliability optimization problems,including composite material layup process constraints.Case studies demonstrated that this approach leads to over 30%weight reduction in CFRP drive shafts compared to metallic counterparts while satisfying reliability requirements and offering insights for the lightweight design of other vehicle components.

Optimization Control of Multi-Mode Coupling All-Wheel Drive System for Hybrid Vehicle

The all-wheel drive(AWD)hybrid system is a research focus on high-performance new energy vehicles that can meet the demands of dynamic performance and passing ability.Simultaneous optimization of the power and economy of hybrid vehicles becomes an issue.A unique multi-mode coupling(MMC)AWD hybrid system is presented to realize the distributed and centralized driving of the front and rear axles to achieve vectored distribution and full utilization of the system power between the axles of vehicles.Based on the parameters of the benchmarking model of a hybrid vehicle,the best model-predictive control-based energy management strategy is proposed.First,the drive system model was built after the analysis of the MMC-AWD’s drive modes.Next,three fundamental strategies were established to address power distribution adjustment and battery SOC maintenance when the SOC changed,which was followed by the design of a road driving force observer.Then,the energy consumption rate in the average time domain was processed before designing the minimum fuel consumption controller based on the equivalent fuel consumption coefficient.Finally,the advantage of the MMC-AWD was confirmed by comparison with the dynamic performance and economy of the BYD Song PLUS DMI-AWD.The findings indicate that,in comparison to the comparative hybrid system at road adhesion coefficients of 0.8 and 0.6,the MMC-AWD’s capacity to accelerate increases by 5.26%and 7.92%,respectively.When the road adhesion coefficient is 0.8,0.6,and 0.4,the maximum climbing ability increases by 14.22%,12.88%,and 4.55%,respectively.As a result,the dynamic performance is greatly enhanced,and the fuel savings rate per 100 km of mileage reaches 12.06%,which is also very economical.The proposed control strategies for the new hybrid AWD vehicle can optimize the power and economy simultaneously.

Analysis of the Spatiotemporal Variation Characteristics and Driving Factors of Land Vegetation GPP in a Certain Region of Asia

Gross primary productivity (GPP) of vegetation is a critical indicator of ecosystem growth and carbon sequestration. The spatiotemporal variation characteristics of land vegetation GPP trends in a specific region of Asia from 2001 to 2020 were analyzed by Sen and MK trend analysis methods in this study .Moreover , a GPP change attribution model was established to explore the driving influences of factors such as Leaf Area Index (LAI), Land Surface Temperature (LST), Vapor Pressure Deficit (VPD), Soil Moisture, Solar Radiation and Wind Speed on GPP. The results indicate that summer GPP values are significantly higher than those in other months, accounting for 60.8% of the annual total GPP;spring and autumn contribute 18.91% and 13.04%, respectively. In winter, due to vegetation being nearly dormant, the contribution is minimal at 7.19%. Spatially, GPP shows a decreasing trend from southeast to northwest. LAI primarily drives the spatial and seasonal variations of regional GPP, while VPD, surface temperature, solar radiation, and soil moisture have varying impacts on GPP across different dimensions. Additionally, wind speed exhibits a minor contribution to GPP across different dimensions.

Research on Anthropomorphic Obstacle Avoidance Trajectory Planning for Adaptive Driving Scenarios Based on Inverse Reinforcement Learning Theory

The forward design of trajectory planning strategies requires preset trajectory optimization functions,resulting in poor adaptability of the strategy and an inability to accurately generate obstacle avoidance trajectories that conform to real driver behavior habits.In addition,owing to the strong time-varying dynamic characteristics of obstacle avoidance scenarios,it is necessary to design numerous trajectory optimization functions and adjust the corresponding parameters.Therefore,an anthropomorphic obstacle-avoidance trajectory planning strategy for adaptive driving scenarios is proposed.First,numerous expert-demonstrated trajectories are extracted from the HighD natural driving dataset.Subsequently,a trajectory expectation feature-matching algorithm is proposed that uses maximum entropy inverse reinforcement learning theory to learn the extracted expert-demonstrated trajectories and achieve automatic acquisition of the optimization function of the expert-demonstrated trajectory.Furthermore,a mapping model is constructed by combining the key driving scenario information that affects vehicle obstacle avoidance with the weight of the optimization function,and an anthropomorphic obstacle avoidance trajectory planning strategy for adaptive driving scenarios is proposed.Finally,the proposed strategy is verified based on real driving scenarios.The results show that the strategy can adjust the weight distribution of the trajectory optimization function in real time according to the“emergency degree”of obstacle avoidance and the state of the vehicle.Moreover,this strategy can generate anthropomorphic trajectories that are similar to expert-demonstrated trajectories,effectively improving the adaptability and acceptability of trajectories in driving scenarios.

GDMNet: A Unified Multi-Task Network for Panoptic Driving Perception

To enhance the efficiency and accuracy of environmental perception for autonomous vehicles,we propose GDMNet,a unified multi-task perception network for autonomous driving,capable of performing drivable area segmentation,lane detection,and traffic object detection.Firstly,in the encoding stage,features are extracted,and Generalized Efficient Layer Aggregation Network(GELAN)is utilized to enhance feature extraction and gradient flow.Secondly,in the decoding stage,specialized detection heads are designed;the drivable area segmentation head employs DySample to expand feature maps,the lane detection head merges early-stage features and processes the output through the Focal Modulation Network(FMN).Lastly,the Minimum Point Distance IoU(MPDIoU)loss function is employed to compute the matching degree between traffic object detection boxes and predicted boxes,facilitating model training adjustments.Experimental results on the BDD100K dataset demonstrate that the proposed network achieves a drivable area segmentation mean intersection over union(mIoU)of 92.2%,lane detection accuracy and intersection over union(IoU)of 75.3%and 26.4%,respectively,and traffic object detection recall and mAP of 89.7%and 78.2%,respectively.The detection performance surpasses that of other single-task or multi-task algorithm models.

Review of Three-phase Soft Switching Inverters and Challenges for Motor Drives

For electric vehicles (EVs),it is necessary to improve endurance mileage by improving the efficiency.There exists a trend towards increasing the system voltage and switching frequency,contributing to improve charging speed and power density.However,this trend poses significant challenges for high-voltage and high-frequency motor controllers,which are plagued by increased switching losses and pronounced switching oscillations as consequences of hard switching.The deployment of soft switching technology presents a viable solution to mitigate these issues.This paper reviews the applications of soft switching technologies for three-phase inverters and classifies them based on distinct characteristics.For each type of inverter,the advantages and disadvantages are evaluated.Then,the paper introduces the research progress and control methods of soft switching inverters (SSIs).Moreover,it presents a comparative analysis among the conventional hard switching inverters (HSIs),an active clamping resonant DC link inverter (ACRDCLI) and an auxiliary resonant commuted pole inverter (ARCPI).Finally,the problems and prospects of soft switching technology applied to motor controllers for EVs are put forward.

Research on the quasi-isentropic driving model of aluminized explosives in the detonation wave propagation direction

Taking CL-20(Hexanitrohexaazaisowurtzitane)-based aluminized explosives with high gurney energy as the research object, this research experimentally investigates the work capability of different aluminized explosive formulations when driving metal flyer plates in the denotation wave propagation direction.The research results showed that the formulations with 43 μm aluminum(Al) powder particles(The particle sizes of Al powder were in the range of 2~43 μm) exhibited the optimal performance in driving flyer plates along the denotation wave propagation direction. Compared to the formulations with Al powder 13 μm, the formulations with Al powder 2 μm delivered better performance in accelerating metal flyer plates in the early stage, which, however, turned to be poor in the later stage. The CL-20-based explosives containing 25% Al far under-performed those containing 15% Al. Based on the proposed quasi-isentropic hypothesis, relevant isentropy theories, and the functional relationship between detonation parameters and entropy as well as Al reaction degree, the characteristic lines of aluminized explosives in accelerating flyer plates were theoretically studied, a quasi-isentropic theoretical model for the aluminized explosive driving the flyer plate was built and the calculation methods for the variations of flyer plate velocity, Al reaction degree, and detonation product parameters with time and axial positions were developed. The theoretical model built is verified by the experimental results of the CL-20-based aluminized explosive driving flyer plate. It was found that the model built could accurately calculate the variations of flyer plate velocity and Al reaction degree over time. In addition, how physical parameters including detonation product pressure and temperature varied with time and axial positions was identified. The action time of the positive pressure after the detonation of aluminized explosives was found prolonged and the downtrend of the temperature was slowed down and even reversed to a slight rise due to the aftereffect reaction between the Al powder and the detonation products.

Optimal Design of the Modular Joint Drive Train for Enhancing Cobot Load Capacity and Dynamic Performance

Automation advancements prompts the extensive integration of collaborative robot(cobot)across a range of industries.Compared to the commonly used design approach of increasing the payload-to-weight ratio of cobot to enhance load capacity,equal attention should be paid to the dynamic response characteristics of cobot during the design process to make the cobot more flexible.In this paper,a new method for designing the drive train parameters of cobot is proposed.Firstly,based on the analysis of factors influencing the load capacity and dynamic response characteristics,design criteria for both aspects are established for cobot with all optimization design criteria normalized within the design domain.Secondly,with the cobot in the horizontal pose,the motor design scheme is discretized and it takes the joint motor diameter and gearbox speed ratio as optimization design variables.Finally,all the discrete values of the optimization objectives are obtained through the enumeration method and the Pareto front is used to select the optimal solution through multi-objective optimization.Base on the cobot design method proposed in this paper,a six-axis cobot is designed and compared with the commercial cobot.The result shows that the load capacity of the designed cobot in this paper reaches 8.4 kg,surpassing the 5 kg load capacity commercial cobot which is used as a benchmark.The minimum resonance frequency of the joints is 42.70 Hz.

Dynamic Development Characteristics and Driving Factors of High Quality Development Level in China’s Five Major Urban Agglomerations

High-quality development is the primary task of comprehensively building a socialist,modern country,as well as the primary task of building urban agglomerations in China.Based on the five development concepts,this paper used the entropy method to measure the High Quality Development Index(HQDI)of the five major urban agglomerations.The results showed that the HQDI of the five major urban agglomerations shows a fluctuating upward trend.First,using the Dagum Gini coefficient to explore the sources of HQDI development differences in urban agglomerations,we found that the main source of HQDI differences in urban agglomerations was inter-regional differences,while intra-regional differences were not important.Second,kernel density estimation was used to test the dynamic evolution trend of HQDI within urban agglomerations.There was a polarisation phenomenon in the HQDI of urban agglomerations,such as the Pearl River Delta urban agglomeration and the Chengdu-Chongqing urban agglomeration.But overall,the degree of imbalance had decreased.Third,using geographic detectors to examine the driving factors of HQDI in urban agglomerations,we found that the main driving forces for improving HQDI in urban agglomerations were economic growth,artificial intelligence technology and fiscal decentralisation.All the interaction factors had greater explanatory power for the spatial differentiation of HQDI,which can be divided into two types:two-factor improvement and non-linear improvement.This study is conducive to improving and enriching the theoretical system for evaluating the high quality development of urban agglomerations,and provides policy references for promoting the high quality development of urban agglomerations.

Perception Enhanced Deep Deterministic Policy Gradient for Autonomous Driving in Complex Scenarios

Autonomous driving has witnessed rapid advancement;however,ensuring safe and efficient driving in intricate scenarios remains a critical challenge.In particular,traffic roundabouts bring a set of challenges to autonomous driving due to the unpredictable entry and exit of vehicles,susceptibility to traffic flow bottlenecks,and imperfect data in perceiving environmental information,rendering them a vital issue in the practical application of autonomous driving.To address the traffic challenges,this work focused on complex roundabouts with multi-lane and proposed a Perception EnhancedDeepDeterministic Policy Gradient(PE-DDPG)for AutonomousDriving in the Roundabouts.Specifically,themodel incorporates an enhanced variational autoencoder featuring an integrated spatial attention mechanism alongside the Deep Deterministic Policy Gradient framework,enhancing the vehicle’s capability to comprehend complex roundabout environments and make decisions.Furthermore,the PE-DDPG model combines a dynamic path optimization strategy for roundabout scenarios,effectively mitigating traffic bottlenecks and augmenting throughput efficiency.Extensive experiments were conducted with the collaborative simulation platform of CARLA and SUMO,and the experimental results show that the proposed PE-DDPG outperforms the baseline methods in terms of the convergence capacity of the training process,the smoothness of driving and the traffic efficiency with diverse traffic flow patterns and penetration rates of autonomous vehicles(AVs).Generally,the proposed PE-DDPGmodel could be employed for autonomous driving in complex scenarios with imperfect data.

Prediction and driving factors of forest fire occurrence in Jilin Province,China

Forest fires are natural disasters that can occur suddenly and can be very damaging,burning thousands of square kilometers.Prevention is better than suppression and prediction models of forest fire occurrence have developed from the logistic regression model,the geographical weighted logistic regression model,the Lasso regression model,the random forest model,and the support vector machine model based on historical forest fire data from 2000 to 2019 in Jilin Province.The models,along with a distribution map are presented in this paper to provide a theoretical basis for forest fire management in this area.Existing studies show that the prediction accuracies of the two machine learning models are higher than those of the three generalized linear regression models.The accuracies of the random forest model,the support vector machine model,geographical weighted logistic regression model,the Lasso regression model,and logistic model were 88.7%,87.7%,86.0%,85.0%and 84.6%,respectively.Weather is the main factor affecting forest fires,while the impacts of topography factors,human and social-economic factors on fire occurrence were similar.