Object detection in low-light scenarios has been widely acknowledged as a significant research area in the field of computer vision, presenting a challenging task. Aiming at the low detection accuracy of mainstream single-stage object detection models in low-light scenarios, this paper proposes a detection model called DK_YOLOv5 based on YOLOv5, specifically designed for such scenarios. First, a low-light image enhancement algorithm with better results is selected to generate enhanced images that achieve relatively better visual effects and amplify target features. Second, the SPPF layer is improved to an R-SPPF module with faster inference speed and stronger feature expression ability. Next, we replace the C3 module with the C2f module and incorporate an attention mechanism to develop the C2f_SKA module, enabling richer gradient information flow and reducing the impact of noise features. Finally, the model detection head is replaced with a decoupled head suitable for the object detection task in this scenario to improve model performance. Additionally, we expand the Exdark dataset to include low-light data of underground mine scenario targets, named Mine_Exdark. Experimental results demonstrate that the proposed DK_YOLOv5 model achieves higher detection accuracy than other models in low-light scenarios, with an mAP0.5 of 71.9% on the Mine_Exdark dataset, which is 4.4% higher than that of YOLOv5.

Multi-sensor fusion has essential applications in the field of target detection. Considering the current actual demand for miniaturization of on-board computers for driverless vehicles, this paper uses the multimodal data YOLOv3 (MDY) algorithm for pedestrian detection on embedded devices. The MDY algorithm uses YOLOv3 as the basic framework to improve pedestrian detection accuracy by optimizing anchor frames and adding small target detection branches. Then the algorithm is accelerated by using TensorRT technology to improve the real-time performance in embedded devices. Finally, a hybrid fusion framework is used to fuse the LIDAR point cloud data with the improved YOLOv3 algorithm to compensate for the shortcomings of a single sensor and improve the detection accuracy while ensuring speed. The improved YOLOv3 improves AP by 6.4% and speed by 11.3 FPS over the original algorithm. The MDY algorithm achieves better performance on the KITTI dataset. To further verify the feasibility of the MDY algorithm, an actual test was conducted on an unmanned vehicle with Jetson TX2 embedded device as the on-board computer within the campus scenario, and the results showed that the MDY algorithm achieves 90.8% accuracy under real-time operation and can achieve adequate detection accuracy and real-time performance on the embedded device.

The decision-making control of autonomous driving in complex urban road environments is a difficult problem in the research of autonomous driving. In order to solve the problem of high dimensional state space and sparse reward in autonomous driving decision control in this environment, this paper proposed a Coordinated Convolution Multi-Reward Proximal Policy Optimization (CCMR-PPO). This method reduces the dimension of the bird’s-eye view data through the coordinated convolution network and then fuses the processed data with the vehicle state data as the input of the algorithm to optimize the state space. The control commands acc (acc represents throttle and brake) and steer of the vehicle are used as the output of the algorithm.. Comprehensively considering the lateral error, safety distance, speed, and other factors of the vehicle, a multi-objective reward mechanism was designed to alleviate the sparse reward. Experiments on the CARLA simulation platform show that the proposed method can effectively increase the performance: compared with the PPO algorithm, the line crossed times are reduced by 24 %, and the number of tasks completed is increased by 54 %.

Unsignalized roundabouts have a significant impact on traffic flow and vehicle safety. To address the challenge of autonomous vehicles passing through roundabouts with low penetration, improve their efficiency, and ensure safety and stability, we propose the proximal policy optimization (PPO) algorithm to enhance decision-making behavior. We develop an optimization-based behavioral choice model for autonomous vehicles that incorporates gap acceptance theory and deep reinforcement learning using the PPO algorithm. Additionally, we employ the CoordConv network to establish an aerial view for spatial perception information gathering. Furthermore, a dynamic multi-objective reward mechanism is introduced to maximize the PPO algorithm’s reward pool function while quantifying environmental factors. Through simulation experiments, we demonstrate that our optimized PPO algorithm significantly improves training efficiency by enhancing the reward value function by 2.85%, 7.17%, and 19.58% in scenarios with 20, 100, and 200 social vehicles, respectively, compared to the PPO+CCMR algorithm. The effectiveness of simulation training also increases by 11.1%, 13.8%, and 7.4%. Moreover, there is a reduction in crossing time by 2.37%, 2.62%, and 13.96%. Our optimized PPO algorithm enhances path selection during autonomous vehicle simulation training as they tend to drive in the inner ring over time; however, the influence of social vehicles on path selection diminishes as their quantity increases. The safety of autonomous vehicles remains largely unaffected by our optimized PPO algorithm.

This paper proposes a compensation technique for the global navigation satellite system (GNSS)/real-time kinematic (RTK) course angle data using an electronic compass for an unmanned system. Additionally, the proportion, integral, and derivative control based on a back-propagation neural network (BP-PID) is introduced to improve the steering safety and riding comfort. The course angle jitter was determined. Because the GNSS/RTK receiver cannot offer stable heading data under specific conditions, including but not limited to susceptibility to obstacles, complex electromagnetic environment, and fewer satellites. The compensation algorithm is based on the determination of the GNSS course angle variance ratio and the asynchronous characteristic between the GNSS and an electronic compass. The combined data provide accurate and robust navigation information for an outdoor unmanned system. To address the limitation of the in-system parameter adjustment, a back-propagation (BP) neural network is adhibited to a conventional proportion, integral, and derivative (PID) lateral control system. The BP-PID control module updates the incremental PID parameters through self-learning, and results in the smoother operation of the vehicle. The flowchart of the learning algorithm and method of calculating the parameters are presented. A typical measurement was conducted and the obtained results were compared with typical RTK navigation results. Thus, the effectiveness of the proposed compensation method was confirmed.

The rapid advancement of autonomous driving has heightened safety concerns, making it essential to adopt a comprehensive approach for secure navigation. Multi-modal methods for 3D object detection play a critical role in enhancing driving safety by integrating data from various sensor types. However, existing methods face challenges, such as feature misalignment and loss, which can lead to overfitting and undermine perception performance and reliability. Building on findings that suggest excluding direct camera branch features from the regression task can improve detection performance, this paper delves deeper into the detection pipeline and introduces a novel multi-modal 3D object detection approach. The proposed approach starts with the introduction of an attention-based module designed to align features across different modalities, thereby enhancing the fusion of features through channel and spatial attention mechanisms. Additionally, an image-guided feature candidate generation strategy is employed to identify candidate regions within the fused features. These original features are then divided into two distinct branches for regression and classification tasks, which are subsequently processed by the detection heads. This approach reduces the model's dependence on precise depth estimation from the image branch and minimizes the impact of sensor calibration errors. Experimental results validate that the proposed method delivers outstanding detection performance. Notably, our best model achieves a competitive performance of 71.0% mAP and 74.0% NDS, while demonstrating strong robustness in scenarios with missing camera data, underscoring its capability to manage complex real-world situations.