交通运输工程学报

Research review on cooperative decision-making for vehicle swarms in vehicle-infrastructure cooperative environment

ZHANG Yi, PEI Hua-xin, YAO Dan-ya

Abstract: The research status of cooperative decision-making of vehicle swarms at home and abroad was analyzed from the aspects of mechanisms, methods, and typical application scenarios of cooperative decision-making for vehicle swarms in vehicle-infrastructure cooperative environments. Considering the different cooperative decision-making mechanisms of vehicle swarms, the research on two kinds of decision-making mechanisms, namely the centralized one and the distributed one, was systematically sorted out. Regarding the diversity of cooperative decision-making methods for vehicle swarms, the advantages and disadvantages of different decision-making methods were comparatively analyzed with the optimization-based and heuristics-based decision-making methods as the thread. As for the different application scenarios of cooperative decision-making for vehicle swarms, the theories and research on the cooperative decision-making for vehicle swarms were comprehensively analyzed in various application scenarios, such as ramps, intersections, road sections, and road networks, Concerning the progress of typical projects on the cooperative decision-making for vehicles at home and abroad, the tasks, construction, and implementation of representative projects on the cooperative decision-making for vehicle swarms in China, the United States, Japan, and Europe were sorted out, respectively. The future development trend of cooperative decision-making for vehicle swarms in vehicle-infrastructure cooperative environments was proposed from the three aspects of system structure, universal model, and demonstration scenarios. Research results show that the centralized cooperative decision-making mechanism for vehicle swarms can be employed to improve the vehicle traffic performance in local areas, whereas the distributed cooperative decision-making mechanism for vehicle swarms is conducive to promoting the global traffic operation. The optimization-based cooperative decision-making method for vehicle swarms can maximize the decision-making effect in specific scenarios, while feasible decision-making effects can be obtained by the heuristics-based cooperative decision-making method for vehicle swarms in most scenarios. Due to the different complexities of the cooperative decision-making problem for vehicle swarms in different scenarios, targeted modeling under a unified framework is required. The research results can provide a reference for the management and control of new hybrid traffic systems in vehicle-infrastructure cooperative environments.More>

2022, 22(3): 1-18. doi: 10.19818/j.cnki.1671-1637.2022.03.001

FullText HTML (630) PDF (6080KB) (690)

Research review on simulation and test of mixed traffic swarm in vehicle-infrastructure cooperative environment

SHANGGUAN Wei, LI Xin, CHAI Lin-guo, CAO Yue, CHEN Jing-jing, PANG Hao-jie, RUI Tao

Abstract: The developments of vehicle-infrastructure cooperation and corresponding simulation and test technologies were summarized, and the simulation requirements, classical methods, and technical bottlenecks in the rudiment, infancy, and developing stages were discussed with a focus on the typical simulation results. A new three-layer virtual-real interactive simulation and test framework was proposed based on the traffic subject modeling, swarm behavior simulation, and test result analysis. According to the simulation requirements of mixed traffic subjects, a model for the heterogeneous traffic subjects was constructed, and the operation mechanism of mixed traffic was analyzed to serve as the underlying model support for the simulation system. With the designed virtual-real interactive simulation and test framework, breakthroughs were accomplished in the scenario generation technology for the mixed traffic swarm intelligence, and a simulation method for the mixed traffic swarm intelligence was put forward. Then, simulation tests of decision-making and control methods for different swarm intelligences were carried out in the selected typical traffic scenarios, such as intersections and road sections, to verify the effectiveness of the proposed method. Finally, the future development directions of vehicle-infrastructure cooperation and corresponding suggestions were summarized. Research results show that show that compared with the traditional simulation and test method, the proposed virtual-real interactive simulation and test method reduces the system's simulation granularity from 500 ms to less than 100 ms, the simulation scale increases from 9 nodes and 500 traffic subjects to 150 nodes and 2 000 traffic subjects, and the number of simulated scenarios enhances from 36 to 98. The dynamic adjustment within a range of 0-100% penetration rate of heterogeneous traffic subjects is achieved, and the efficiency, scale, and coverage of the vehicle-infrastructure cooperative simulation and test of mixed traffic are effectively improved. The requirements of vehicle-infrastructure cooperative simulation and test in the new mixed traffic environment are rapidly evolving towards the larger swarm, higher intelligence, and larger scale. Carrying out research on the method and technology for the simulation and test on the vehicle-infrastructure cooperative swarm intelligence based on the virtual-real interaction and operating environment data simulation will effectively promote the development of the next generation of the intelligent traffic system.More>

2022, 22(3): 19-40. doi: 10.19818/j.cnki.1671-1637.2022.03.002

FullText HTML (664) PDF (19201KB) (411)

Review on impact of autonomous driving on travel behaviors

LI Rui-min, DAI Jing-chen

Abstract: The research status of travel behaviors under the influence of autonomous driving was analyzed from three aspects: vehicle travel demand, travel mode choice, and travel time utilzation. The data foundations and methodologies of existing studies on the impact of autonomous driving on travel behaviors were analyzed. The key factors affecting the choice of travel mode in the autonomous driving environment, existing challenges, and future development directions were discussed. Research results indicate that the studies on vehicle travel demand estimation primarily focus on the potential travel of underserved population and usually analyze the potential demand changes through assumptions, which have some deficiencies in terms of the reliability of the assumptions and the accuracy of the results. The studies on travel mode choice show that vehicle service and travel attributes, social demographic and family attributes, travel habits attributes, residential and environmental attributes, personal psychology and preference attributes are key influencing factors. Taking into account of different research objects, scenario design, and analysis methods, the influence of the factors such as gender, age, vehicle license holding, and family structure on travel behaviors remains to be examined. There is great uncertainty and heterogeneity in people's perceptions of the ways and benefits of the use of travel time in the era of autonomous driving, and there is a great need for theoretical models to further discuss the potential change of the use of travel time. Based on the limitations of existing studies on the impact of autonomous driving on travel behaviors, several future improvement directions are provided, including establishing standardized description of autonomous vehicles, enriching data collection methods, carrying out horizontal and longitudinal comparative studies, strengthening the consideration of the heterogeneity of influencing factors, and identifying the interaction mechanism among various travel behaviors.More>

2022, 22(3): 41-54. doi: 10.19818/j.cnki.1671-1637.2022.03.003

FullText HTML (522) PDF (4588KB) (359)

Human-machine integration method for steering decision-making of intelligent vehicle based on reinforcement learning

WU Chao-zhong, LENG Yao, CHEN Zhi-jun, LUO Peng

Abstract: In terms of the continuous dynamic allocation problem of driving weights between human and autonomous driving systems in the human-machine integration (HMI) driving system of intelligent vehicles, especially the low adaptability problem of weight allocation methods caused by modeling errors, a HMI steering decision-making method based on the reinforcement learning was proposed. In view of drivers' steering characteristics, a driver model based on the two-point preview was built, and an autonomous steering control model of intelligent vehicles was established by adopting the predictive control theory. On this basis, a steering control framework of simultaneous human-machine in-loop for intelligent vehicles was constructed. According to the Actor-Critic reinforcement learning framework, a deep deterministic policy gradient (DDPG) agent for the human-machine driving weight allocation was designed, and a model-based gain function was proposed with the curvature adaptability, tracking accuracy, and ride comfort as targets. A reinforcement learning framework for the HMI driving weight allocation was constructed, which contains a driver model, an autonomous steering model, a driving weight allocation agent, and a gain function. To verify the effectiveness of the proposed method, eight drivers were recruited, and a total of 48 simulated driving experiments were carried out. Research results show that in the verification of curvature adaptability, the HMI-DDPG method is superior to the manned driving and HMI-Fuzzy methods. The trackability improves by an average of 70.69% and 39.67%, respectively, and the comfortability increases by an average of 18.34% and 7.55%, respectively. In the verification of speed adaptability, under the conditions of a vehicle speed of 40, 60, and 80 km·h^-1, the time proportion is 90.00%, 85.76%, and 60.74%, respectively, when the driver's weight is greater than 0.5. The phase trajectories of both the trackability and the comfort can effectively converge. Therefore, the proposed method can adapt to changes in curvature and vehicle speed and improve the trackability and comfort on the premise of ensuring safety.More>

2022, 22(3): 55-67. doi: 10.19818/j.cnki.1671-1637.2022.03.004

FullText HTML (373) PDF (13174KB) (271)

Guidance and cooperative operation method for group vehicles in vehicle-infrastructure cooperative environment

SHANGGUAN Wei, PANG Xiao-yu, LI Qiu-yan, CHAI Lin-guo

Abstract: To solve the traffic congestion problem caused by urban development, explore the potential of road traffic, and improve the driving efficiency of vehicles in the road network in vehicle-infrastructure cooperative environments, a guidance optimization method and a cooperative contral strategy for group vehicles were proposed. For the vehicle guidance allocation, the group vehicles allocation rules based on the road saturation, vehicle travel time, and delay were designed with the goals of optimal traffic efficiency and minimum vehicle emissions by the feasible path between the starting point and the destination. An optimization model for the group vehicles guidance allocation was built and solved by the multi-objective non-dominated sorting genetic algorithms-Ⅱ (NSGA-Ⅱ) and the multi-objective particle swarm optimization algorithm. Regarding the strategy for the vehicle cooperative operation control, a multi-vehicle cooperative operation model based on the idea of the gravitational field was created, and a multi-vehicle cooperative acceleration and deceleration strategy was proposed. The results of vehicle guidance optimization under different penetration rates of connected and automated vehicle (CAV) were compared through the simulation verification. The vehicle cooperative acceleration and deceleration strategy was simulated, and the guidance optimization method and the cooperative control strategy were co-simulated. Simulation results show that the multi-objective guidance allocation method can improve the vehicle speed and environmental benefits, and the average speed of the group vehicles is positively correlated with the CAV penetration rate. In the four-car group driving environment, the cooperative acceleration and deceleration strategy can increase the initial average acceleration of the vehicle by 15.0% and 8.2% respectively, during the acceleration and deceleration. The vehicle can quickly reach the target speed, and the safety of the vehicles can thereby be ensured. In the co-simulation environment, the accelerations of the group vehicles in the road network increase by 11.6% on average, their speeds increase by 1.6% on average, and their carbon-oxygen compound emissions reduce by about 4.9%. Therefore, the proposed method can be employed to improve the traffic efficiency of the road network, reduce the energy consumption of vehicles, and lower the adverse impact on the environment. 2 tabs, 10 figs, 31 refs.More>

2022, 22(3): 68-78. doi: 10.19818/j.cnki.1671-1637.2022.03.005

FullText HTML (316) PDF (7569KB) (186)

Prediction of traffic swarm movement situation based on generalized spatio-temporal graph convolution network

QU Xu, GAN Rui, AN Bo-cheng, LI Lin-heng, CHEN Zhi-jun, RAN Bin

Abstract: To address the problem that traffic congestion on highways and urban expressways is becoming more and more serious and causes great difficulties for traffic management and control, a traffic speed prediction model was proposed based on the generalized spatio-temporal graph convolution network (GSTGCN). According to the complex spatio-temporal characteristics of traffic data, the generalized traffic data graph structure was defined, and the adjacency relationships of the generalized graph were constructed. By the basic theory of graph convolution network, the Chebyshev approximation and the first-order approximation were adopted to simplify the computational cost of the graph convolution operation, and a generalized graph convolution operator was established. With the generalized graph convolution module, standard convolution module, and linear fully-connected layer, a GSTGCN model was presented to extract the spatial and temporal characteristics of complex traffic data. The vehicle speed, flow, and occupancy datum were recorded by 38 detectors at 5-minute intervals for 21 weekdays on the expressway network in Milwaukee, Wisconsin, USA. The short-term traffic speed prediction accuracy and training efficiency of the GSTGCN model were evaluated on this data set. Analysis results show that compared with the results of the traditional auto regressive integrated moving average (ARIMA) model, the long short-term memory (LSTM) model, and the recent spatio-temporal graph convolution network (STGCN) model, the root mean square error, mean absolute error, and mean absolute percentage error of the GSTGCN model in the traffic speed prediction reduces by 22.79%, 22.97%, and 16.73%, respectively. Moreover, the training time of the GSTGCN model is 5.17% and 75.71% shorter than those of the STGCN model and LSTM model, respectively. Therefore, the GSTGCN model is able to effectively deal with the complex spatio-temporal traffic data structure, accurately predict the traffic speed, and provide information on the movement situation of traffic swarm for the traffic control and management. 4 tabs, 6 figs, 31 refs.More>

2022, 22(3): 79-88. doi: 10.19818/j.cnki.1671-1637.2022.03.006

FullText HTML (387) PDF (7111KB) (169)

Left-turn motion planning model of autonomous driving based on approximate grid risk assessment of vehicle-non-motor conflict

WANG Ping, QI Xu-dong, YANG Jing-wen, HAN Yu, LI Li, WANG Gui-ping, YAO Jun-feng, ZHAO Xiang-mo

Abstract: In order to improve the driving safety, reliability and efficiency of autonomous vehicles at complex vehicle-non-motor mixed intersections, a left-turn motion planning model of autonomous driving based on approximate grid risk assessment of vehicle-non-motor conflict was proposed and generalized. The division rules of static discrete sequence intersection grid area were set.According to the probability conversion relationship of traffic behaviors in multi traffic states, the motion state of non-motor vehicle in the subdivision grid was predicted, and the risk level of the conflict area between motor vehicles and non-motor vehicles was dynamically evaluated. On this basis, the model prediction method was used to design the lateral and longitudinal control algorithms of autonomous vehicle, the desired trajectory could be tracked by adaptively adjusting the heading and speed, and the grid conflict area could be avoided synchronously. Combined with the constraints of vehicle dynamics and external interaction environment, a traffic simulation platform of intersection four-phase signal control was developed. From the aspects of efficiency optimization, comfort optimization, and the offset between actual planned path and reference path, the model-in-the-loop test was used to verify the effectiveness of the left-turn motion planning in the vehicle-non-motor conflict area. Research results show that the proposed model can effectively extract and predict grid risk information dynamically, and improves the safe interaction, efficient traffic and driving comfort between autonomous vehicles and surrounding non-motor vehicles. Compared with similar algorithms, the offset of the planned path can reduce by 17.1%, the traffic efficiency can increase by 26.6% and the comfort can increase by 39.3% at most. Therefore, the proposed model has obvious advantages in the efficient passage through vehicle-non-motor conflict area and low space occupation by the planned path. 2 tabs, 14 figs, 32 refs.More>

2022, 22(3): 89-103. doi: 10.19818/j.cnki.1671-1637.2022.03.007

FullText HTML (376) PDF (11282KB) (149)

Vehicle trajectory prediction based on spatio-temporal information fusion in crowded driving scenario

LI Li, PING Zhen-dong, ZHU Jin-yu, XU Zhi-gang, WANG Gui-ping

Abstract: The spatio-temporal interaction information among vehicles was integrated into the convolutional social pooling network to formulate a human-driving vehicle trajectory prediction model in the crowded driving scenario. The long short term memory (LSTM) network was used to predict the speeds of the crowded vehicles. The prediction result was used to calculate the speed differences among the vehicles. The LSTM encoder was built to capture the time-series features of the crowded vehicle trajectories. The convolutional social pooling network was designed to captured the spatial dependence of the crowded vehicles. The emerging probabilities of all possible movements of the vehicles and corresponding trajectories were predicted by the LSTM decoder. The movement with the highest emerging probability and its trajectory were taken as final prediction result of trajectory. The real vehicle trajectory dataset was used in the parameter calibration and performance verification of the proposed model. Different methods of trajectory encoding/decoding and speed predicting were tested to figure out their influences on the model performance. The test results were used to identify the optimal model structure. Calculation results show that compared with historical speed, predicted speed used to calculate speed difference as model input can decrease by 19.45% in terms of root mean square error (RMSE). Compared with the gate recurrent unit, the LSTM network as speed predictor can decrease by 4.91% in terms of RMSE. Compared with the original convolutional social pooling network, the trajectory prediction errors of the proposed model respectively decrease by 20.32% and 21.04% in terms of RMSE and negative log-likelihood. The model performance is also significantly better than other variants of the original convolutional social pooling network. The computation time difference of the proposed model and original convolutional social pooling network is about 3 ms, which meets the request of real-time application. 8 tabs, 9 figs, 23 refs.More>

2022, 22(3): 104-114. doi: 10.19818/j.cnki.1671-1637.2022.03.008

FullText HTML (467) PDF (8168KB) (208)

SVM-LSTM-based car-following behavior recognition and information credibility discirmination

SHI Yu-chen, YAN Song, YAO Dan-ya, ZHANG Yi

Abstract: To effectively improve the traffic system security by using the real-time interaction information in intelligent vehicle-infrastructure cooperative systems (i-VICS), a credibility discrimination approach for traffic information based on the traffic business features was proposed. In particular, the model for the car-following behavior recognition and the information credibility discrimination was built based on the support vector machine (SVM) and long short-term memory (LSTM) neural network. It was composed of the SVM-based car-following behavior recognition model and the LSTM neural network-based car-following speed prediction model. The feature vector representing the vehicle driving states was set, and the vehicle driving states were divided into the following and non-following by the SVM-based car-following behavior recognition model. For following vehicles, their speeds were predicted by the LSTM neural network-based car-following speed prediction model according to the history data. With the SVM-LSTM-based information credibility discrimination model, the credibility of vehicle data was judged by checking whether the difference between the predicted speed and the actual speed of the following vehicles was within the reasonable range, and in this way, the information credibility discrimination was achieved. The public dataset was employed to train and test the proposed models, and several abnormal test datasets of various abnormity types and abnormity amplitude were built to verify the SVM-LSTM neural network-based model for the car-following behavior recognition and the information credibility discrimination. Research results show that the vehicle driving behavior recognition accuracy of the SVM-based car-following behavior recognition model is up to 99%, and the predicted car-following speed precision with an order of magnitude of cm·s^-1 can be achieved by the LSTM neural network-based car-following speed prediction model. The discrimination accuracy of the SVM-LSTM neural network-based model for the car-following behavior recognition and information credibility discrimination is up to 97% on the normal test datasets and multiple abnormal test datasets. Thus, the proposed approach can be applied for the real-time information credibility discriminations of road side units (RSUs) to on-board units (OBUs) and between OBUs. 8 tabs, 9 figs, 30 refs.More>

2022, 22(3): 115-125. doi: 10.19818/j.cnki.1671-1637.2022.03.009

FullText HTML (371) PDF (3450KB) (122)

Traffic capacity enhancement strategy for urban expressway diversion area under vehicle-infrastructure cooperative environment

LI Rui, RAN Bin, QU Xu

Abstract: According to the mixed traffic flow characteristics of vehicles including different types of automatic vehicles (AVs) and human-driven vehicles (HVs) in the urban expressway diversion area under a vehicle-infrastructure cooperative environment, the dynamic acceleration and variable lane-changing probability were introduced to improve the traffic flow rules of a cellular automata model. The lane-changing simulation experiments in the diversion area were designed by considering the coupling influence of factors such as the penetration rate of AVs on the main road, proportion of large vehicles, penetration rate of off-ramp AVs, rate of off-ramp vehicles, number of off-ramp lanes, and distance before lane-changing. The influences of indicators including the free lane-changing rate and average distance before lane-changing of off-ramp vehicles were compared and analyzed under multi-factor coupling actions, and change rules of road capacity of the urban expressway diversion area were studied. On the basis of the variable distance before lane-changing, a strategy for improving the road capacity of the diversion area with mixed traffic flows was proposed. Analysis results show that the road capacity improves as the free lane-changing rate of off-ramp vehicles in the diversion area increases. The penetration rate of AVs on the main road has the most significant impact on the road capacity, and the road capacity under the environment with fully AVs is twice that under the environment with fully HVs. The impact of the number of off-ramp lanes on the road capacity is not significant, and the road capacity of two off-ramp lanes improves by about 3%, compared with that of one off-ramp lane. The distance before lane-changing greatly affects the road capacity, and the road capacity of the diversion area enhances by 9.6%-10.6% when the distance before lane-changing increases from 100 m to 150 m. Therefore, mobile traffic signs can be utilized to guide vehicles to change lanes in advance, which can significantly enhance the traffic capacity of the diversion area. 1 tab, 20 figs, 31 refs.More>

2022, 22(3): 126-138. doi: 10.19818/j.cnki.1671-1637.2022.03.010

FullText HTML (307) PDF (16790KB) (138)

Vehicle-infrastructure cooperative control method of connected and signalized intersection in mixed traffic environment

WANG Run-min, ZHANG Xin-rui, ZHAO Xiang-mo, WU Xia, FAN Hai-jin

Abstract: In order to improve the adaption of intelligent vehicle-infrastructure cooperative control methods around connected and signalized intersection to real traffic environment, a novel intelligent vehicle-infrastructure cooperative optimization control method was proposed under the traffic scene of eight-phase connected and signalized intersection mixed intelligent and connected vehicle (ICV) with connected and human-driven vehicle (CHV). Based on modeling the kinematic characteristics and car-following behavior of ICV in the mixed traffic scene, a mixed platoon was formed. A rolling optimization-based cooperative control method of traffic signal and ICV trajectory was proposed based on the platoon model, safety constraints, and fuel consumption model. The cooperative control problem was divided into two layers based on the idea of asynchronous hierarchical optimization, the upper layer was traffic signal timing optimization, and the lower layer was ICV trajectory optimization. Taking the travel time delay and fuel consumption of the vehicle at the intersection as the optimization objectives, the genetic algorithm and three-stage trajectory optimization method were used to solve the traffic signal timing optimization and ICV trajectory optimization, respectively. The stability of the mixed vehicle platoon was verified under different steady-state speeds and penetration rates of ICV. The control effect of the proposed control method and the influence of key parameters on the control effect were analyzed. Analysis results indicate that the proposed control method can effectively improve the traffic efficiency and fuel economy of the intersection under various traffic flows and penetration rates of ICV. In the total ICV environment, the indexes respectively improve by 57.3% and 13.3% when the proposed control method is compared with the method without optimization. Compared with the condition without penetration, with the increase of the penetration rate of ICV, the control efficiency of the proposed control method constantly improves, and the indexes respectively increase by 42.0% and 14.2%. Even if the penetration rate of ICV is only 20%, the proposed control method can also achieve 20.4% improvement in the term of traffic efficiency. The longer traffic signal cycle and the shorter driver reaction time of CHV can provide a benefit for the cooperative control effect. 2 tabs, 13 figs, 40 refs.More>

2022, 22(3): 139-151. doi: 10.19818/j.cnki.1671-1637.2022.03.011

FullText HTML (525) PDF (5621KB) (256)

Multi-intersection coordinated control method based on group decision-making

MA Cheng-yuan, ZHU Ji-chen, LAI Jin-tao, ZHANG Zhen, YANG Xiao-guang

Abstract: On the basis of the group decision-making mechanism with competition and cooperation, the isolated signal optimization was modeled as the right-of-way competition process of all phases at intersections, and the coordination among many intersections was modeled as the cooperation process between upstream and downstream phases. A signal timing design method for road networks was proposed under considering both the multi-intersection synergy and the optimal control of isolated intersections. The perceptibility of the vehicle route information in road networks under the vehicle-road cooperative environment was used to quantitatively analyze the coupling relationship between upstream and downstream traffic in a dynamic and accurate manner. On this basis, a hierarchical dynamic decision-making framework was established to avoid the impact of the control decisions of upstream and downstream intersections on local decisions in single-layer decision-making, and the composite relationship between the traffic states of road networks and the signal control decision in the cooperative control model was decoupled. A distributed decision-making algorithm for signal timing was designed based on the competitiveness of each traffic flow at intersections, and the performances of the proposed group decision-making cooperative control method and the traditional cooperative control method was compared by a simulation test platform. Research results show that compared with the traditional cooperative control method, the group decision-making cooperative control method can dynamically adapt to the traffic demand of the road network, and has significant advantages in traffic efficiency and stability. Under the traffic demand levels with different saturation degrees, the average vehicle delay can reduces by more than 15%. In the case of high traffic saturation, the delay can reduce by 19.2%, so the control advantage is more obvious. As the upstream outflow of the vehicles can be reduced by the group decision-making cooperative control method when the vehicle queues at downstream intersections for inflow are long, the maximum queue length in road networks can be cut by over 40%. In this way, the overflow risk in road networks can be avoided. Through the distributed solution of the group decision-making cooperative control method, the calculation time of a single decision-making process is less than 0.01 s, so the method has the potential to be applied to the real-time signal timing decision in large-scale complex road network. 1 tab, 7 figs, 31 refs.More>

2022, 22(3): 152-161. doi: 10.19818/j.cnki.1671-1637.2022.03.012

FullText HTML (325) PDF (7974KB) (136)

Mixed traffic group throttling control strategy for traffic bottleneck of expressway

ZHAO Hang, ZHAO Min, SUN Di-hua, DU Cheng

Abstract: Considering the lane reduction bottleneck of expressways under mixed traffic condition composed of human-driven vehicles (HVs) and connected and automated vehicles (CAVs), a novel speed harmonization control strategy (throttling control strategy for short) was developed from the viewpoint of group control. A speed controller for the leading CAV was designed on the basis of the bottleneck traffic state and the Greenshields model. A nonlinear controller for the target changing was developed for the control during the CAV throttling group formation. A platoon-like controller for the CAV throttling group was built, and the group formation and group speed were thereby regulated dynamically according to the bottleneck traffic state. The speed control method for the leading CAV was combined to regulate the vehicles overtaking each throttling group periodically. A longitudinal safety controller for the CAV was presented to resolve the vehicle safety problem in the processes of group formation and group evolution. Simulation results show that, on the bottleneck road of the expressway, compared with the traditional traffic system, the proposed dynamic throttling control strategy is applied when the CAV penetration rate reaches 5% and vehicle flow is 2 000, 3 000, 5 000 and 6 000 veh·h^-1, respectively, the corresponding traffic efficiency improves about 5.87%, 16.97%, 11.07%, and 10.25%, respectively. On an expressway bottleneck road with a fixed traffic flow of 3 000 or 6 000 veh·h^-1, compared with the traditional traffic system, the traffic efficiency of the controlled traffic system can be enhanced by around 24% when the CAV penetration rate reaches 10%, 20%, and 30%, respectively. According to the analysis of space headways, the controlled CAVs can avoid collision during the entire throttling process and keep a safe distance of more than 9 m from their predecessors. Therefore, the throttling control strategy is effective in dealing with the bottleneck problem of expressway. 3 tabs, 15 figs, 30 refs.More>

2022, 22(3): 162-173. doi: 10.19818/j.cnki.1671-1637.2022.03.013

FullText HTML (198) PDF (17362KB) (120)

Distributed cooperative decision-making method for vehicle swarms in large-scale road networks

PEI Hua-xin, YANG Jing-xuan, HU Jian-ming, ZHANG Yi

Abstract: To resolve the cooperative decision-making problem for vehicle swarms in large-scale road networks under the vehicle-infrastructure cooperative environment, a distributed cooperative decision-making method for vehicle swarms was proposed. On the basis of the in-depth analysis on the traffic control characteristics, the road network decomposition model was built to decompose the large-scale cooperative decision-making problem into several homogeneous small-scale sub-problems, each covering three different types of traffic areas: the upstream road segment, intersection, and downstream road segment. By the virtual vehicle mapping technique, the cooperative decision-making model of vehicle swarms was constructed to transform the two-dimensional cooperative decision-making problem of vehicle swarms at intersections into a one-dimensional problem. Similar to the cooperative decision-making method for vehicle swarms in the road segment areas, the interaction and conflict resolution between vehicles at intersections were accomplished by controlling the equivalent time headway of vehicles in the virtual vehicle platoon, and then the unified cooperative decision-making parameters were used to solve the cooperative decision-making problem of vehicle swarms in different areas of each sub-problem. Upon the unification of the cooperative decision-making parameters of vehicle swarms in different areas, the cooperative mechanism between the upstream and downstream areas was designed to ensure that the appropriate driving decisions could be made by the upstream vehicles under the full consideration of the downstream traffic states. Simulation results show that under different traffic demand settings, smooth spatiotemporal trajectories are presented by all vehicles while passing through the conflict areas after the proposed method is adopted, and the violent fluctuations in vehicle spatiotemporal trajectories are avoided. Compared with the purely distributed method, the fuel consumption of vehicles reduces by up to 14% with the proposed method under the given simulation conditions. Therefore, the proposed distributed cooperative decision-making method for vehicle swarms is effective in reducing the impact of conflict areas on the traffic flow continuity after being implemented in large-scale road networks, and thus ensuring the safe, smooth, and environmentally friendly driving of vehicles. 7 figs, 30 refs.More>

2022, 22(3): 174-183. doi: 10.19818/j.cnki.1671-1637.2022.03.014

FullText HTML (281) PDF (6079KB) (136)

Stability of PID control system for vehicle platoon with input delay and communication delay

ZHU Xu, ZHANG Ze-hua, YAN Mao-de

Abstract: The internal stability and string stability of the PID control system were analyzed for vehicle platoon with input delay and communication delay, the sufficient and necessary conditions of the internal stability were emphatically studied, and the exhaustive and exact time delay margins were derived. In the internal stability analysis, considering that the PID control system for vehicle platoon is a neutral time delay system with input delay and communication delay, the sufficient and necessary strong stability conditions were proposed by analyzing the stability of the neutral operator via Rekasius substitution and Routh table. In order to facilitate selecting the PID parameters, a sufficient condition with a more concise form was derived. Then, the clustering method of characteristic roots was applied to obtain the exhaustive and exact time delay margins. Considering the vehicle platoon with an odd number of following vehicles, the upper bound of the input delay, which was independent of the scale of the vehicle platoon, was derived. In order to ensure that the interference and error propagated backward along the vehicle platoon without divergence, the error transfer function among the vehicles was analyzed, and the sufficient condition of string stability under the influence of two delays was given. Simulation results show that the internal stability and string stability of vehicle platoon can be guaranteed simultaneously by the distributed PID controllers under communication delay and input delay. The state errors quickly converge to zero within 15 s. When the velocities of the vehicles are constant, an desired safe distance maintains 50 m between the successive vehicles. When the leader vehicle accelerates at 0.5 m·s^-2 and decelerates at 0.8 m·s^-2, the velocities and accelerations of the following vehicles asymptotically change with those of the leader and are consistent with the leader when the leader's velocity is constant. Under the different driving conditions, the spacing errors caused by the acceleration and deceleration of the leader are less than 0.2 m, and propagate backward along vehicle platoon without divergence. 1 tab, 11 figs, 36 refs.More>

2022, 22(3): 184-198. doi: 10.19818/j.cnki.1671-1637.2022.03.015

FullText HTML (458) PDF (16384KB) (153)

Multi-scale filtering synchronization method for vehicle-infrastructure cooperative twin-simulation testing

QIU Wei-zhi, SHANGGUAN Wei, CHAI Lin-guo, CHU Duan-feng

Abstract: To enhance the synchronization performance of the vehicle-infrastructure cooperative twin-simulation testing system, the operation mechanism of twin objects was clarified. Then the interference factors affecting the synchronization performance of the system were analyzed to establish the synchronous mapping model for the twin state. In view of the asynchronous clock problem in twin state sampling, a clock error estimation strategy was designed to correct the measurement time deviation of the twin-simulation testing system. On this basis, a multi-scale filtering updating mechanism was introduced by combining the principle of the Kalman filtering. Furthermore, a measurement noise model considering the synchronization sampling errors was established, and the multi-scale filtering synchronization optimization method was proposed. Finally, the vehicle trajectories from the NGSIM dataset were selected to carry out experiments in a constructed prototype system of twin-simulation testing. Research results show that the synchronization performance can be well maintained by the proposed multi-scale filtering synchronization optimization method under different vehicle speeds. In terms of synchronizing the lateral coordinate, the mean absolute error (MAE) is less than 1 mm, and 99.5% of absolute error (AE) can be controlled to within 8 mm. In terms of synchronizing the longitudinal coordinate, the MAE is less than 9 mm, and 99.5% of AE can be controlled to within 38 mm. In terms of synchronizing the speed, the MAE is less than 2.8 cm·s^-1, and 99.5% of AE can be controlled to within 24 cm·s^-1. In terms of synchronizing the yaw angle, the MAE is less than 1.1×10^-3 rad, and 99.5% of AE can be controlled to within 1.1×10^-2 rad. Compared with the dead reckoning method, the proposed method can improve the synchronization accuracy by an average of 30.0% in terms of lateral coordinate, longitudinal coordinate, speed, and yaw angle, solve the asynchronous state problem for twin objects effectively, and guarantee the real-time synchronization and accurate operation of the vehicle-infrastructure cooperative twin-simulation testing system. 3 tabs, 10 figs, 31 refs.More>

2022, 22(3): 199-209. doi: 10.19818/j.cnki.1671-1637.2022.03.016

FullText HTML (198) PDF (7091KB) (66)

Traffic sign detection algorithm based on pyramid multi-scale fusion

GAO Tao, XING Ke, LIU Zhan-wen, CHEN Ting, YANG Zhao-chen, LI Yong-hui

Abstract: In order to address the problems of misdetection and missing detection for small target traffic signs in traditional traffic sign detection algorithms, a traffic sign detection algorithm based on pyramidal multi-scale fusion was proposed. To improve the feature extraction capability of the algorithm for traffic signs, the residual structure of ResNet was adopted to build the backbone network of the algorithm, and, the number of shallow convolutional layers of the backbone network was increased to extract more accurate semantic information of smaller scale traffic signs. Based on the idea of feature pyramid network, four different prediction scales were introduced in the detection network to enhance the fusion between deep and shallow features. To further improve the detection accuracy of the algorithm, the GIoU loss function was introduced to localize the anchor boxes of traffic signs. Meanwhile, the k-means algorithm was introduced to cluster the traffic sign label information and generate more accurate prior bounding boxes. In order to verify the generalization of the algorithm and solve the problem of inter-class imbalance of TT100K data set used in the experiment, the data set was enhanced and expanded. Experimental results show that the accuracy, recall and average accuracy of the proposed algorithm are 86.7%, 89.4% and 87.9%, respectively, significantly improving compared with traditional target detection algorithms. The adoption of multi-scale fusion detection mechanism, GIoU loss function and k-means improves the detection performance of the algorithm to different degrees, and its precision improves by 4.7%, 1.8% and 1.2%, respectively. The algorithm has better performance in the detection of traffic signs under different scales, and its recall rate is 90%, 93% and 88% under the scales of (0, 32], (32, 96] and (96, 400] in TT100K dataset, respectively. Comparing with YOLOv3, the proposed algorithm can correctly locate and classify traffic signs under the interference of different weather, noise and geometric transformation, which proves that the proposed algorithm has good robustness and generalization, and is suitable for road traffic sign detection. 7 tabs, 18 figs, 30 refs.More>

2022, 22(3): 210-224. doi: 10.19818/j.cnki.1671-1637.2022.03.017

FullText HTML (302) PDF (28402KB) (87)

A fused network based on PReNet and YOLOv4 for traffic object detection in rainy environment

CHEN Ting, YAO Da-chun, GAO Tao, QIU Hui-hui, GUO Chang-xin, LIU Zhan-wen, LI Yong-hui, BIAN Hao-yi

Abstract: In order to improve the detection accuracy of vehicle target in severe rainy day under traffic environment, a deep learning network DTOD-PReYOLOv4 (derain and traffic object detection-PReNet and YOLOv4) was proposed based on the fusion of PReNet and YOLOv4, which integrated the improved image restoration subnet D-PReNet and the improved target detection subnet TOD-YOLOv4. D-PReNet could extract rain streak features more effectively, since it introduced the multi-scale expansion convolution fusion module (MSECFM) and the attentional mechanism residual module (AMRM) with SEBlock into PReNet. TOD-YOLOv4 improved not only the detection accuracy of small traffic target, but also the detection efficiency, since it replaced the backbone module CSPDarknet53 of YOLOv4 with the lightweight CSPDarknet26 of YOLOv4, added CRB into PANet of YOLOv4 neck, and utilized k-means++ instead of the original network clustering algorithm. DTOD-PReYOLOv4 was verified based on the constructed vehicle target database VOD-RTE in rainy day traffic scenario. Research results show that compared with the current series of YOLO networks, the proposed DTOD-PReYOLOv4 can better extract the features with lower resolutions by superimposing RB over ResBlock_body1 in the shallow layer. It can effectively reduce the convolutional layer redundancy and improve the memory utilization, since ResBlock_body3, ResBlock_body4 and ResBlock_body5 in deep layer can be properly cropped to ResBlock_body3×2, ResBlock_body4×2 and ResBlock_body5×2, respectively. It also can alleviate the degradation of small target detection effect caused by the deepening of network layers by adding jump connection to Concat+Conv×5 in PANet to form CRB. In the process of multi-scale detection, k-means++ algorithm is adopted to allocate smaller prior boxes that are more suitable for the larger feature images, but larger prior boxes that are more suitable for smaller feature images, which further improves the accuracy of target detection. The harmonic mean value of precision and recall rate, average precision and detection speed of DTOD-PReYOLOv4 respectively increase by 5.02%, 6.70% and 15.63 frames per second compared with MYOLOv4, by 3.51%, 4.31% and 2.17 frames per second compared with TOD-YOLOv4, by 46.07%, 48.05% and 18.97 frames per second compared with YOLOv3, and by 31.06%, 29.74% and 16.26 frames per second compared with YOLOv4. 4 tabs, 12 figs, 44 refs.More>

2022, 22(3): 225-237. doi: 10.19818/j.cnki.1671-1637.2022.03.018

FullText HTML (477) PDF (26171KB) (193)

Perception of moving objects in traffic scenes based on heterogeneous graph learning

YANG Biao, YAN Guo-cheng, LIU Zhan-wen, LIU Xiao-feng

Abstract: In order to improve the operation efficiency and transportation safety of unmanned vehicles in traffic scenes, the perception of moving objects in traffic scenes was investigated based on the heterogeneous graph learning.In view of the influence of complex interaction relations between moving objects on their motions in actual traffic scenes, an integrated perception framework of multi-object detection-tracking-prediction was proposed based on the heterogeneous graph learning. YOLOv5 and DeepSORT were combined to detect and track the moving objects, and the trajectories of the objects were obtained. The long short-term memory (LSTM) network was used to learn the objects' motion information from their historical trajectories, and a heterogeneous graph was introduced to learn the interaction information between the objects and improve the prediction accuracies of the trajectories of moving objects. The LSTM network was also utilized to decode the objects' motion and interaction information to obtain their future trajectories, and the method was evaluated on the public transportation datasets Argoverse, Apollo, and NuScenes to verify its effectiveness.Analysis results show that the combination of YOLOv5 and DeepSORT can realize the detection and tracking of moving objects and achieve a detection accuracy rate of 75.4% and a continuous tracking rate of 61.4% for moving objects in traffic scenes. The heterogeneous graph can effectively capture the complex interaction relations between moving objects, and the captured interaction relations can improve the accuracy of trajectory prediction. The error of the predicted average displacement of moving objects reduces by 63.0% after the interaction relations captured by the heterogeneous graph are added. As a result, it is effective to consider the interaction relations between moving objects in traffic scenes. The historical and future motion information of moving objects can be perceived by introducing the heterogeneous graph to capture the interaction relations between moving objects, so as to facilitate unmanned vehicles to better understand complex traffic scenes. 4 tabs, 9 figs, 36 refs.More>

2022, 22(3): 238-250. doi: 10.19818/j.cnki.1671-1637.2022.03.019

FullText HTML (257) PDF (19432KB) (79)

Driverless vehicle positioning algorithm based on simultaneous positioning and mapping in low-visibility environment

GAO Yang, CAO Wang-xin, XIA Hong-yao, ZHAO Yi-hui

Abstract: In order to achieve high-precision positioning for driverless vehicles in a large-scale and low-light environment, a fused positioning algorithm LVG_SLAM was proposed based on the system framework of the VINS-Mono algorithm. In LVG_SLAM, a RFAST low-light image enhancement module and a VG fusion positioning module were proposed and then added. The RFAST low-light image enhancement module applied a wavelet transform to separate the detailed information from the brightness information. In the RFAST module, the unified threshold and mean filter were applied to filter the detailed noisy information from the oringinal image while the bilateral texture filter algorithm was applied to enhance the detail information. After that, the multi-scale retinex algorithm was proposed to further enhance the contrast of the image to improve the success rate of corner extraction in a low-light environment, benefit from which, both the stability of image tracking and the robustness of the positioning algorithm were improved. Using an unscented Kalman filter (UKF) algorithm, the VG fusion positioning module loosely fused the positioning information from both the global navigation satellite system (GNSS) and the inertial navigation equipment. The fused positioning result was introduced as a constraint into the back end of the LVG_SLAM algorithm, benefit from which, the influence of cumulative error on the positioning accuracy of the algorithm was reduced by a joint nonlinear optimization. Analysis results show that compared with the VINS-Mono algorithm, the LVG_SLAM algorithm performs better on the EuRoC and Kitti public datasets, and the root mean square error reduces by 38.76% and 58.39%, respectively, so that the motion trajectory estimated by the LVG_SLAM algorithm is closer to the real trajectory. In an experiment of night road scene, the LVG_SLAM algorithm successfully constrains the positioning error into a certain range, and detects the closed loop, which greatly improves the positioning performance. The root mean square error, average error, maximum error, and median error reduce by 79.61%, 82.50%, 71.31%, and 83.77%, respectively. Compared with the VINS-Mono algorithm, the proposed LVG_SLAM algorithm has obvious advantages in both positioning accuracy and robustness. 4 tabs, 12 figs, 26 refs.More>

2022, 22(3): 251-262. doi: 10.19818/j.cnki.1671-1637.2022.03.020

FullText HTML (192) PDF (17931KB) (69)

Scenario factor analysis and test case generation for vehicle-infrastructure cooperative mixed traffic

ZHAO Tong, SHANGGUAN Wei, CHAI Lin-guo, GUO Peng

Abstract: Considering the test requirements of multi-vehicle efficiency and single-vehicle safety scenarios in a mixed traffic environment, a test case generation method for the vehicle-infrastructure cooperation was developed based on the scenario factor analysis of mixed traffic. For higher diversity and coverage of test cases, the interaction mechanism of mixed traffic characteristic factors was analyzed, the hierarchical model of mixed traffic scenario factors was constructed, and the consistency description index of the importance of scenario factors was proposed. On this basis, a complexity model of test cases was built. For the simulation and test of multi-vehicle efficiency scenarios, a complexity-inspired generation combination test case method was proposed, and a combination strategy with strong coupling of scenario factors was designed. For the simulation and test of single-vehicle safety scenarios, a Monte Carlo test case generation method based on the complexity clustering was put forward, and a sampling mechanism of characteristic parameters of risk scenarios was designed. Typical scenarios of vehicle-infrastructure cooperative mixed traffic were selected for simulation experiments, to verify the effectiveness of the proposed test case generation method. Research results show that for the ramp-merging scenario test of expressways in mixed traffic of multi-vehicle efficiency, compared with the traditional pairwise test method, the proposed method improves the maximum complexity of scenarios, proportion of high-complexity scenarios, and coverage of test cases by 11.93%, 60.02%, and 12.08%, respectively. For the vehicle-infrastructure cooperative lane-changing warning scenario test of single-vehicle safety, compared with the traditional Monte Carlo test method, the proposed method raises the number of dangerous scenarios by 195%, with reducing the parameter estimation error by 5.95%, and increases the number of high-risk scenarios by 119%, with reducing the parameter estimation error by 4.78%. Therefore, the proposed method can improve the diversity and coverage of test cases, contribute to carry out the functional test of the vehicle-infrastructure cooperative system under complex environments and risk conditions, and can effectively meet the test requirements of multi-vehicle efficiency and single-vehicle safety scenarios. 6 tabs, 11 figs, 34 refs.More>

2022, 22(3): 263-276. doi: 10.19818/j.cnki.1671-1637.2022.03.021

FullText HTML (667) PDF (7793KB) (180)

User Center

Journal Info

Collection of Source JournalsMore>

2022 Vol. 22, No. 3

Search Type

Search Journal

LinksMore>