Deep Reinforcement Learning (DRL) has demonstrated its great potential for Adaptive Traffic Signal Control (ATSC) tasks at single‐intersection. In the transportation network multi‐agent environment, cooperative learning among multi‐agents has become a hot research topic. Based on the distributed control model, this paper presents a hybrid reward function model for the dynamic density method of intersections, which emphasizes the priority of emergency vehicles (EMV) while maximizing the traffic efficiency of social vehicles, and solves the problem of sparse reward due to the ambiguous guidance relationship between the multi‐agent Deep Reinforcement Learning (MDRL) state and reward function of the urban road network scenario. On the other hand, based on multi‐agent A2C (MA2C) algorithm, this paper presents Shared Experience MA2C (SEMA2C) between agents. In the transportation network, each intersection represented by an agent has similar task objectives. SMEA2C algorithm takes the current agent as the main body of self‐learning, and utilizes the principle of importance sampling to learn from the experience data of the agents located at adjacent intersections. The experimental results show that the proposed SEMA2C performs well in multi‐agent traffic signal control tasks, and has greater advantages than the similar algorithms.

In order to further improve the accuracy of short‐term traffic flow prediction on designated sections of highways, a combined prediction model is designed in this paper to predict the traffic flow on designated sections of highways. Firstly, for the shortcomings of artificial rabbits optimization (ARO) algorithm, sine cosine ARO (SARO) is proposed by incorporating sine cosine algorithm (SCA) idea into ARO, and introducing the non‐linear sinusoidal learning factor. Secondly, three mobile inverted bottleneck convolution (MBConv) modules are utilized to form the MB3 module, and with BiGRU are utilized to form the MB3‐BiGRU combined prediction model. Finally, the MB3‐BiGRU model is optimized by SARO to achieve short‐term prediction of traffic flow. The analysis results show that using the United Kingdom highway dataset as the data source, the SARO‐MB3‐BiGRU presented in this paper reduces the root mean squared error (RMSE) by 32.58%, the mean absolute error (MAE) by 30.25%, and the decision coefficient ( R ² ) reaches 0.96729, as compared to BiGRU. Compared with other common models and algorithms, the SARO has good solving capabilities and versatility, and the SARO‐MB3‐BiGRU model has been greatly improved in terms of prediction accuracy and generalization ability, which has better prediction ability and engineering reference value.

This paper presents a lateral control scheme based on event‐triggered model predictive control for trajectory tracking of autonomous vehicles. Firstly, the augmentation system is constructed based on the known road curvature information, and the model predictive controller is utilized to obtain the optimal control sequence. Then, an event‐triggered mechanism is introduced to improve the real‐time performance of the control system. The strategy targets to reduce the computational complexity and solving frequency of the optimization problem. In addition, a contraction constraint is structured using the backstepping control strategy to ensure the stability of the control system. Finally, experiments are conducted through the CarSim/Simulink joint simulation platform, and compared with the traditional model predictive control, the method proposed in this paper has better tracking accuracy and improves the real‐time performance of the control system.

This paper is concerned with the set-membership filtering problem for the two-dimensional shift-varying systems subject to the stochastic communication protocol and uniform quantization effects. To prevent the data transmission from collisions, only one sensor can transmit the measured information to the filter at each sampling shift instant, and the selected sensor is determined by the scheduling strategy of the stochastic communication protocol. On the contrary, the uniform quantization mechanism has been employed to mitigate the influence of quantization error on filtering performance. Incorporating the stochastic communication protocol and uniform quantization mechanism, this paper proposes a set-membership filter design framework for the two-dimensional shift-varying systems with unknown-but-bounded noises. Sufficient conditions are derived for the existence of desired set-membership filter by utilizing double mathematical induction, such that the estimation error resides within the ellipsoidal set. Moreover, the optimal filtering algorithm is given by minimizing the ellipsoidal constraints. Finally, several examples are provided to illustrate the effectiveness of the proposed filter design algorithm.

The asynchronous filtering problem for two-dimensional (2D) Markovian jump systems (MJSs) under hybrid cyber attacks and stochastic communication protocol (SCP) is the focus of this investigation. To avoid data transmission collisions, only a single sensor can send measurement data through the communication network at each sample shift moment, and the sensor chosen is decided by the SCP mechanism. Besides, the impact of stochastic hybrid cyber attacks on filtering performance is also considered. In the background of hybrid cyber attacks and SCP, an augmented model of 2D MJSs is constructed. Then, a novel asynchronous filter structure is proposed that can simultaneously characterize the MJSs and filter pattern mismatch as well as the mismatch between the actual SCP pattern information and the available SCP pattern information. Then, sufficient criteria conditions are established to ensure that dynamic filtering error systems meet robust H_∞ mean-square stability with a certain level of disturbance attenuation. Moreover, the conservativeness of the stability criterion is further reduced by the 2D summation inequality lemma. Finally, an industrial heating exchange process is employed to illustrate the effectiveness of the developed asynchronous filtering algorithm.