Search

The rapid advance of autonomous vehicles (AVs) has motivated new perspectives and potential challenges for existing modes of transportation. Currently, driving assistance systems of Level 3 and below have been widely produced, and several applications of Level 4 systems to specific situations have also been gradually developed. By improving the automation level and vehicle intelligence, these systems can be further advanced towards fully autonomous driving. However, general development concepts for Level 5 AVs remain unclear, and the existing methods employed in the development processes of Levels 0–4 have been mainly based on task-driven function development related to specific scenarios. Therefore, it is difficult to identify the problems encountered by high-level AVs. The essential logical and physical mechanisms of vehicles have hindered further progression towards Level 5 systems. By exploring the physical mechanisms behind high-level autonomous driving systems and analyzing the essence of driving, we put forward a coordinated and balanced framework based on the brain–cerebellum–organ concept through reasoning and deduction. Based on a mixed mode relying on the crow inference and parrot imitation approach, we explore the research paradigm of autonomous learning and prior knowledge to realize the characteristics of self-learning, self-adaptation, and self-transcendence for AVs. From a systematic, unified, and balanced point of view and based on least action principles and unified safety field concepts, we aim to provide a novel research concept and develop an effective approach for the research and development of high-level AVs, specifically at Level 5.

Situation assessment is crucial for intelligent vehicles, enabling detection of potential risks to dynamic and complex traffic environments. In this paper, we propose a unified framework that tackles the coupling relationships between traffic participants and quantifies the possible range of vehicle trajectory generation and the expected losses caused by risk source attributes in the driving process. We first apply the state space trajectory planning scheme based on a sampling algorithm to generate the path candidates; each feasible path is designed through a parametric cubic spline. Then, to evaluate the risk range in the driving process, we quantify the interaction of traffic participants, and employ the principle of least action to calculate the cost of each feasible path when achieving the destination. The probability distribution map, namely the possible range of driving trajectories, can be obtained based on the path cost. Furthermore, the vehicle-to-vehicle interaction is calculated based on the equivalent force, which estimates the expected accident losses. Finally, the vehicle trajectory prediction and the expected loss are combined to output the probabilistic situation assessment of intelligent vehicles. The algorithm is implemented in different scenarios and applied to the trajectory planning process. Results demonstrate that, compared with the classical situation assessment metric, the developed method can determine and accurately identify the influence range of driving risk in real-time, predict a dangerous situation earlier, and ensure the vehicle avoids obstacles in advance.

Due to the limited sensing ability with the single-view camera and the real-time requirement for multi-view scenarios or deep learning-based methods in complex scenes, the output of lane detection is not applicable for the actual lane departure warning system. To tackle this challenge, the authors propose a fast and robust approach for lane detection based on well-designed multi-camera fusion, integrating vanishing point estimation, and specified feature fitting strategies. To meet real-time demand, several simple but effective image processing means are introduced and improved. Concretely, on account of statistical information, the authors’ method carries out an improved region of interest selection to speed up the detection. Afterwards, they used the B-spline fitting lane line on the strength of the RANdom SAmple consensus algorithm for the front view image detection and improved the Hough algorithm for the two rear-view images correspondingly. Using coordinate conversion and self-designed fusion strategy, they get the robust lane information based on symmetrical lane detection from the left/right sides of both front and side views. Experimental results in newly introduced multi-camera scenarios show that their multi-camera fusion framework contributes to significant improvement in accuracy and robustness in comparison with traditional methods.

Purpose For autonomous vehicles, trajectory prediction of surrounding vehicles is beneficial to improving the situational awareness of dynamic and stochastic traffic environments, which is a crucial and indispensable element to realize highly automated driving. Design/methodology/approach In this paper, the overall framework consists of two parts: first, a novel driver characteristic and intention estimation (DCIE) model is built to indicate the higher-level information of the vehicle using its low-level motion variables; then, according to the estimation results of the DCIE model, a classified Gaussian process model is established for probabilistic vehicle trajectory prediction under different motion patterns. Findings The whole method is later applied and analyzed in the highway lane-change scenarios with the parameters of models learned from the public naturalistic driving data set. Compared with other traditional methods, the performance of this proposed approach is proved superior, demonstrated by the higher accuracy in the long prediction horizon and a more reasonable description of uncertainty. Originality/value This hierarchical approach is proposed to make trajectory prediction accurately both in the short term and long term, which can also deal with the uncertainties caused by the perception system or indeterminate vehicle behaviors.

Intelligent‐driving technologies play crucial roles in reducing road‐traffic accidents and ensuring more convenience while driving. One of the significant challenges in developing an intel...

Making long-term trajectory prediction accurately for surrounding vehicles is the crucial prerequisite for intelligent vehicles to accomplish superb decision making and motion planning. In this paper, to achieve high-quality prediction accuracy both in the short and long term, we propose an integrated probabilistic framework with the combination of driver awareness model and Gaussian process model. The former model can obtain high-level semantic information using low-level two-dimensional motion elements. And the latter incorporates the vehicle physical model to reach good prediction performance with strengthened historical input sequence. Furthermore, experiments on the public naturalistic driving dataset in lane-changing scenarios are conducted to verify our novel approach. Compared with another advanced method, the superiorities of our proposed approach are demonstrated with higher estimation and prediction accuracy, as well as more reasonable uncertainty description in terms of the whole prediction process.

Traffic accident statistics have shown the necessity of risk assessment when driving in the dynamic traffic environment. If the risk associated with different traffic elements (i.e., road, environment and vehicles) could be evaluated accurately, potential accidents could be significantly avoided or mitigated. This paper proposes a driving risk assessment model that can quantitatively evaluate the driving risk associated with intelligent vehicles via the coupled analysis of different traffic elements. First, we present a concept of the internal field and external field for establishing the driving risk coupling model, through employing the internal field to define the risk range of driver's perspective and the external field to calculate the risk coefficients of those traffic elements. Then, the relative risk coefficients are computed by incorporating both naturalistic driving study (NDS) and driver attitude questionnaire (DAQ) using a multinomial logit model. Specifically, we perform a large-scale naturalistic driving study to investigate the objective driving risks. Typical driver behavior parameters, such as velocity, time headway, and acceleration, are analyzed. Besides, a self-reported survey of 364 drivers is conducted to subjectively evaluate the potential risks that drivers may face in various situations. Finally, validation of the model is conducted by comparing the accuracy with the typical risk assessment index, i.e., TTC and THW. Results demonstrate that the proposed approach is effective in evaluating the comprehensive driving risks by quantifying the influence factors of driving risks in dynamic environments.

Probabilistic trajectory prediction for other vehicles can be an effective way to improve the understanding of dynamic and stochastic traffic environment for automated vehicles. One challenge is how to predict the vehicle trajectory accurately both in the short-term and long-term horizon. In this paper, we propose an integrated approach combining the driver characteristic and intention estimation (DCIE) model with the Gaussian process (GP) model. Our proposed method makes use of both vehicle low-level and high-level information and inquires parameters by learning from public naturalistic driving dataset. Our method is applied and analyzed in the highway lane change scenarios. Compared with other traditional methods, the advantages of this proposed method are demonstrated by more accurate prediction and more reasonable uncertainty description during the whole prediction horizon.

This paper proposes a method to quantify the driving risks of different traffic elements. We incorporate the objective analysis using naturalistic driving study (NDS) and the subjective analysis with driver attitude questionnaire (DAQ). The objective driving risks are investigated by a large-scale NDS experiment with multiple sources. Meanwhile, typical driver behavior parameters, such as velocity, time headway, and acceleration, are selected and analyzed. A self-reported survey of 364 drivers is conducted to subjectively evaluate the potential risks which drivers may suffer in various situations. NDS and DAQ are then combined together using the multinomial logit model to obtain the relative risks. Results demonstrate that the proposed method can provide an effective measure to quantify the influence factors of driving risks in dynamic environment. It is interesting to note that the risk value from subjective evaluation tends to be higher, implying that the subjective evaluation might be emotional and multi-factor coupling.

Increasing automation calls for evaluating the effectiveness and intelligence of automated vehicles. This paper proposes a framework for quantitatively evaluating the intelligence of automated vehicles. Firstly, we establish the evaluation environment for automated vehicles including test field, test task, and evaluation index. The test tasks include the single vehicle decision-making (turning, lane-changing, overtaking, etc.) and the maneuver execution of multi-vehicle interaction (obstacle avoidance, trajectory optimization, etc.). The intelligence evaluation index is the action amount of driving process considering the safety, efficiency, rationality and comfort. Then, we calculate the actual action amount of the automated vehicle in different scenarios in the test field. Finally, the least action calculated theoretically corresponds to the highest intelligence degree of the automated vehicle, and is employed as a standard to quantify the performance of other tested automated vehicles. The effectiveness of this framework is verified with two naturalistic driving datasets that contain the normal driving scenarios and high-risk scenarios. Specifically, the naturalistic lane-changing data filters 40,416 frames and 179 similar lane-changing trajectories. Compared with the lane-changing behavior of a large number of drivers, experimental results verify that the proposed algorithm can achieve the intelligence degree of drivers in the lane change scenario. Meanwhile, in 253 reconstructed high-risk scenarios, the intelligent risk avoidance ability of the proposed intelligence degree evaluation algorithm can be verified by comparing with the driver behavior and TTC algorithm. These experimental results show that the proposed framework can effectively quantify intelligence and evaluate the performance of automated vehicles under various scenarios.