Tianjun Sun

For human-machine co-driving vehicles, the dynamic complexity of urban scenarios and the uncertainty of driving behavior impose stringent demands on the accuracy and generalization capability of ego vehicle trajectory prediction. Current research predominantly relies on multi-vehicle interaction information or single-scenario settings, overlooking the inherent dynamic correlation between driver and vehicle. This results in prediction models struggle to adapt to complex urban environments. To address this, this paper proposes a trajectory prediction framework based on driver-vehicle coupling feature encoding, enabling precise capture of vehicle trajectory evolution patterns under driver manipulation across diverse urban scenarios. The framework employs bidirectional LSTMs to perform temporal encoding on historical driver-vehicle coupling features and future road geometric features. Combined with an attention mechanism, it generates context vectors integrating temporal features and manipulation details, ultimately using LSTMs to recursively produce multi-step prediction results. Validation using diverse urban scenarios data from a dynamic driving simulator demonstrates that our prediction framework achieves precise trajectory prediction in typical scenarios such as lane changing, turning, and roundabout navigation, while exhibiting robust stability and generalization capabilities.

Quantifying uncertainty will significantly improve perceptual performance and provide more comprehensive environmental information for decision-making and planning modules of autonomous vehicles. Unfortunately, most perception methods exhibit excellent performance in accuracy but fall short in estimating associated uncertainty. To fill this gap, the variance inference ensemble network is proposed to enhance environmental perception and quantify uncertainty for 3-D object detection in point cloud. Specifically, the method is divided into three parts. Several variance inference neural networks that adopt multivariate Gaussian distribution for direct modeling are first constructed through a two-stage training strategy, extracting both the object details and variances from point cloud data in parallel. Following this, an uncertainty-aware fusion strategy is designed to integrate and filter the multiple results above based on the associated uncertainty and yield reliable and comprehensive results. Furthermore, a novel metric, uncertainty index, is coined to estimate the uncertainty of detected objects for the single deterministic network and ensemble network in a unified and quantitative manner. Finally, we validate our method on the KITTI dataset. The experiment demonstrates that our method outperforms the original baseline and recent uncertainty quantification methods across different scenarios.

Appropriate interaction with human-driven vehicles is crucial for advancing autonomous vehicles from demonstrations in controlled environments to applications in open-road scenarios. Accurately recognizing and quantifying driver interaction behaviors is essential for understanding their interaction intentions and subsequently formulating suitable responses. Previous studies have primarily focused on the generation of interaction behaviors. However, insufficient attention has been paid to the recognition and quantification of driver interaction behaviors. We establish a higher-order traffic flow network and present mathematical definitions for driver interaction behaviors and their strengths based on this framework. Furthermore, we find that humans identify interaction behaviors during driving by perceiving conflicts between different drivers’ intentions. Therefore, we propose a driver resistance field model (DRFM) to characterize driving intentions and describe these conflicts through operations within the resistance fields, enabling the recognition of interactions. Through computational social experiments and real-world experiments, the DRFM is demonstrated to accurately quantify the strength of driver-driver interactions. The analysis of extensive cases indicates that the quantification from the DRFM effectively capture the dynamics of interaction behaviors, aligning with human cognition of driver interactions, with remarkably higher correlation coefficients compared to baseline methods.

The purpose of this paper is to investigate the interactions between driver manipulation behaviors. Firstly, representative samples from driving behavior dataset were screened. Secondly, four characteristic indicators were constructed. Thirdly, the Apriori algorithm was established for correlation analysis on these four driving behaviors in order to obtain the influence relationship between different driving behaviors. Therefore, the correlation between different undesirable driving behaviors and reveals the intrinsic law of driver manipulation behavior were illustrated, which provided great significance for improving the warning effect and intervention efficiency of automotive active safety systems.

Big data already covers intelligent vehicles and is driving the autonomous driving industry’s transformation. However, the large amounts of driving data generated will result in complex issues and a huge workload for the test and verification processes of an autonomous driving system. Only effective and precise data extraction and recording aimed at the challenges of low efficiency, poor quality, and a long-time limit for traditional data acquisition can substantially reduce the algorithm development cycle. Based on the premise of driver-dominated vehicle movement, the virtual decision-making of autonomous driving systems under the accompanying state was considered as a reference. Based on a dynamic time warping algorithm and forming a data filtering approach under a dynamic time window, an automatic trigger recording control model for human-vehicle difference feature data was suggested. In this method, the data dimension was minimized, and the efficiency of the data mining was improved. The experimental findings showed that the suggested model decreased recorded invalid data by 75.35% on average and saved about 2.65 TB of data storage space per hour. Compared with industrial-grade methods, it saves an average of 307 GB of storage space per hour.

SDP (Stochastic Dynamic Programming) control strategy can mine the travel data of drivers, so that the energy-saving potential of vehicles could be improved. However, there are two problems need to be solved in the current SDP: firstly, the analysis and construction method of driver’s typical driving cycle is not clear; secondly, the adaptability of SDP algorithm to a typical driving cycle is insufficient. To solve the above problems, a driving cycle construction method for off-line SDP solution is proposed, which is based on “analysis, dimension reduction and clustering” process. In addition, a coupled control strategy (ECMS-SDP) based on driving conditions identification is developed. Because it is difficult to predict the driving conditions in real time, the working part of the coupled control strategy is calculated by the method of improved random forest. The simulation results show that ECMS-SDP control strategy can save 8%–15% fuel on average compared with CD-CS control strategy, and can save 4%–7% fuel on average compared with ECMS control strategy. The results prove that the ECMS-SDP coupled control strategy can respond well to the changing driving environment, and the fuel economy of vehicles is enhanced.