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.

As one of the main methods in positioning and navigation, the Global Navigation Satellite System (GNSS) is widely applied in many areas, such as road engineering, automobile navigation and many other outdoor applications. In addition to an outdoor location-based service, there are many needs for indoor positioning. However, GNSS is hard to position indoors due to signal blockage, and outdoor–indoor transition areas where multipath effects are unavoidable and signal interference is severe. This article proposes a GNSS/INS/UWB tightly-coupled integration positioning methodology to address the problem of accurate positioning in GNSS-difficult areas and aims to provide seamless positioning in different scenarios. In indoor environments, the system uses the UWB/INS tightly-coupled integration positioning mode and periodically corrects the INS errors with range measurements obtained from the UWB transmitting unit to achieve centimeter-level accuracy. In outdoor GNSS-difficult areas like outdoor–indoor transition areas, a GNSS/INS/UWB tightly-coupled integration positioning system is proposed to ensure positioning accuracy continuity and further improve the reliability of the system. To evaluate the positioning performance of the proposed UWB enhanced multi-sensor tightly-coupled integration system, an experiment was conducted in a high-rise building. The experiment results show that the proposed integrated positioning system can provide seamless and accurate positioning results in both indoor and outdoor–indoor transition areas, with DRMS of 5.25 cm and RMSE of 10.18 cm.

With the increasing demand for rail transit, wireless communication technologies are playing a growing significant role in train control systems, which enables the railway systems to provide a higher capacity and more efficient services. However, due to the nature of radio frequency propagation, the quality of the train-to-ground wireless connections is highly dependent on a well-planned deployment of the wayside access points. To improve both the accuracy and the efficiency in railway network planning, in this paper, a deep learning technology is exploited to model the wireless propagation, which was very difficult to deterministically predict at a fast speed in our previous research due to the high computation demanding. In this proposed wireless propagation model, Kalman filter is utilized to update the neural network parameters online, which makes this model can meet the variation of the environment. The numeric evaluation result shows that the deep neural network based wireless channel model can precisely predict the outage probability with a very low computational cost.

The BeiDou Navigation Satellite System (BDS) has been widely applied in many areas, including communications, transportation, emergency rescue, public security, surveying, precise positioning, and many other industrial applications. The integration of BDS and an Inertial Navigation System (INS) has the ability to achieve a more consistent and accurate positioning solution. However, during long BDS outages, the performance of such an integrated system degrades because of the characteristics of the Inertial Measurement Unit (IMU) - the sensor errors accumulate rapidly with time when operated in a standalone mode. In this paper, a BDS/INS/odometer/map-matching (MM) positioning methodology for train navigation applications is proposed, to solve the problem of positioning during BDS outages when trains pass through signal obstructed areas such as under bridges, inside tunnels and through deep valleys. The seamless transition of the train operation in various scenarios can be therefore maintained. When the train operates in an open-sky environment the BDS signals are available to provide accurate positioning, and the integrated BDS/INS/odometer/MM system is used to correct the INS errors using BDS measurements and to calculate the velocity using the odometer in the navigation frame so as to improve the system reliability and accuracy. In addition, the integrated system also delivers accurate positioning measurements at a high update rate. When the BDS signals are blocked the integrated system switches to the INS/odometer/MM mode, and the integrated system corrects the INS errors by using odometer measurements. In order to evaluate the proposed system a real experiment was conducted on the Qinghai-Tibet railway in western China. The experimental results indicate that the proposed system can provide accurate and continuous positioning results in both open-sky and BDS signal-obstructed environments.

In this paper, novel adaptive fault-tolerant control (FTC) algorithms are put forward to address the position and velocity tracking control problems of high-speed trains (HSTs). The basic running resistances, additional resistances and interactive forces between the connected carriages are taken into account. On account of the system uncertainties, a novel multidimensional sliding mode surface with time-varying parameters estimated by the adaptive technique is proposed. And neural networks (NNS) are made use of approximating the additional resistances viewed as a bounded disturbance. Two cases for system parameters are taken into consideration: 1)Parameters with unknown boundary are unknown; 2)Parameters with known upper and lower boundary are unknown. As for the former, the unknown parameters are formulated via formal mathematical expression with indicator functions and the adaptive technique is introduced; As for the latter, a continuous functions related to the above multidimensional sliding mode surface is defined and the adaptive technique with Project functions is designed, which guarantees saturation characteristics of the presented controller. The closed-loop system stability can be proved by means of the Lyapunov theory and the feasibility and effectiveness of the presented control schemes can be revealed via simulation experiments.

GNSS/INS integrated navigation and positioning technology is widely used for train operations applications seeking a decrease in dependence on trackside equipment. However, during train operations, environmental factors such as bridges, tunnels, hillsides, and buildings will block the satellite signals. When INS works in a stand-alone mode, the system errors accumulate with time due to the inertial sensor characteristics. To overcome this problem, light detection and ranging (lidar) equipment can be introduced to help control INS divergence. The lidar point cloud data is first processed using the random sample consensus (RANSAC) algorithm to determine the track plane, then the moving average filtering (MAF) algorithm is used to detect and classify the rails. The frogs can be detected based on the changing rail number and the constraint condition for turnout parameters. With the assistance of a previously surveyed turnout database, the geographical positions of the detected frogs can be obtained and integrated with INS data to converge INS sensor errors when GNSS is not available. In this paper a lidar-aided GNSS/INS integrated navigation system is proposed that can ensure continuous positioning in both GNSS-available and GNSS-blocked scenarios. GNSS/INS is applied when the system operates in the GNSS-available scenario, and switches to the INS/lidar mode when the GNSS signals are blocked. To evaluate the performance of the proposed system, an actual train experiment was conducted along the Qinghai-Tibet railway. The results confirm that the proposed lidar-aided navigation system can achieve accurate and continuous positioning. When GNSS is not available, the proposed method can effectively suppress the error of INS with a DRMS of 1.66 m, which is a 70.4% improvement compared with a conventional GNSS/INS integrated system.