Explore the vast range of titles available.

This paper briefly introduces the blind spot detection and obstacle avoidance algorithm for autonomous electric vehicles that utilize laser radar and an onboard camera. Simulation experiments were conducted. During the experiments, the You Only Look Once version 5 (YOLOv5) model was tested for its ability to identify obstacles in road images captured by the onboard camera. Next, the effectiveness of the onboard camera combined with laser radar in accurately locating obstacles was assessed. Finally, the obstacle avoidance capability of the autonomous electric vehicle was tested. The results showed that the YOLOv5 model accurately located and identified obstacles in the image. The combination of the onboard camera and laser radar accurately determined the coordinates of obstacles. Moreover, the autonomous electric vehicle based on the onboard camera and laser radar successfully avoided obstacles and reached its destination without being affected by low‐light environments. The focus of this paper is to combine two road information acquisition methods, namely laser radar and onboard cameras, use the You Only Look Once version 5 (YOLOv5) model to identify road obstacles in the camera images, integrate the laser radar point cloud data to determine the real coordinates of the obstacles, and employ the artificial potential field method to plan the path. Subsequently, simulation experiments were carried out. The experimental results showed that the YOLOv5 model could accurately locate and identify obstacles in the images; the combination of onboard cameras and laser radar could accurately locate the real coordinates of obstacles; and the autonomous electric vehicles based on the camera and laser radar could effectively avoid obstacles and reach the destination.

Buses and heavy vehicles have more blind spots compared to cars and other road vehicles due to their large sizes. Therefore, accidents caused by these heavy vehicles are more fatal and result in severe injuries to other road users. These possible blind-spot collisions can be identified early using vision-based object detection approaches. Yet, the existing state-of-the-art vision-based object detection models rely heavily on a single feature descriptor for making decisions. In this research, the design of two convolutional neural networks (CNNs) based on high-level feature descriptors and their integration with faster R-CNN is proposed to detect blind-spot collisions for heavy vehicles. Moreover, a fusion approach is proposed to integrate two pre-trained networks (i.e., Resnet 50 and Resnet 101) for extracting high level features for blind-spot vehicle detection. The fusion of features significantly improves the performance of faster R-CNN and outperformed the existing state-of-the-art methods. Both approaches are validated on a self-recorded blind-spot vehicle detection dataset for buses and an online LISA dataset for vehicle detection. For both proposed approaches, a false detection rate (FDR) of 3.05% and 3.49% are obtained for the self recorded dataset, making these approaches suitable for real time applications.

When turning or passing through intersections, the vehicle’s A-pillar blind spot may obscure road conditions. Using cameras to capture and display external images on the A-pillar screen can eliminate the A-pillar blind spot. However, this solution still faces a mismatch between the displayed screen image and the surrounding environment image. To obtain more accurate blind spot images, accurate blind spot cropping areas are calculated through vehicle A-pillar detection and image registration technology. By introducing eye position and posture as features and considering the corresponding blind spot cropping areas as target variables, a decision tree model is trained to predict the blind spot cropping areas. This model can rapidly and accurately predict the blind spot areas for display on the A-pillar screen, with an average pixel error of 12.7, a coefficient of determination of 0.85, and an image processing speed within one millisecond. Experiments show that the A-pillar blind spot visualization system based on A-pillar detection and decision tree model prediction effectively eliminates the A-pillar blind spot, enhancing driving safety.

Providing blind-spot-free vehicle surround view to the driver is important for many driving maneuvers such as parking. Existing vehicle Surround View System (SVS) can only visualize front, left, rear and right side of the vehicle but leaves the under vehicle area unknown. However, perceiving the under vehicle area is critical for many tasks such as passing through speed bumps, avoiding potholes, driving on narrow roads with high curbs or the unpaved terrain. In this paper, we propose a novel Under Vehicle Reconstruction (UVR) algorithm which utilizes what the vehicle sees in the past and vehicle egomotion to “see” through the original invisible under vehicle area. First, front or back fisheye cameras, are utilized to build a local textured map for future usage. Second, vehicle’s precise location and orientation within the local map is estimated using the vehicle egomotion. Finally, correspondent under vehicle area texture is retrieved from the map using vehicle’s pose and stitched together with traditional Surround View System to provide a new blind-spot-free visualization. As far as we know, our work is the first solution that can provide full under vehicle area reconstruction which empowers many Advanced Driving Assistant System (ADAS) functionalities such as transparent hood or transparent vehicle. Experiments on both simulated and real data are presented to show the effectiveness and robustness of the proposed algorithm.

The ever-changing landscape of the automotive sector has been significantly influenced by technological advancements, particularly the incorporation of Radar technology, have played a vital role in shaping the safety features of modern vehicles. In the automotive industry, radar sensors play a important role in advanced driver assistance systems (ADAS), improving road safety and driver convenience and One of the ADAS latest application is the Blind Spot Monitoring (BSM) system, which addresses the great challenge of driver visibility in certain areas around the vehicle. The BSM system uses radar sensors to detect obscured vehicles or objects, thereby reducing accident risk and improving road safety and driver convenience. As traffic densities rise and vehicles become more interconnected, BSM systems have become essential for safer and more efficient driving. The continuous evolution of radar technology integration demonstrates the automotive industry’s commitment to driver safety and its dedication to using advanced technologies to enhance the driving experience.

Pedestrian accidents involving heavy vehicles such as trucks and buses remain a critical safety issue, primarily due to structural blind spots. While existing systems like radar-based FCW and BSD have been adopted, they are not fully optimized for pedestrian detection, particularly under adverse weather conditions. This study focused on the empirical validation of a 360-degree pedestrian collision avoidance system using multichannel ultrasonic sensors specifically designed for heavy-duty vehicles. Eight sensors were strategically positioned to ensure full spatial coverage, and scenario-based field experiments were conducted under controlled rain (50 mm/h) and fog (visibility <30 m) conditions. Pedestrian detection performance was evaluated across six distance intervals (50–300 cm) using indicators such as mean absolute error (MAE), coefficient of variation (CV), and false-negative rate (FNR). The results demonstrated that the system maintained average accuracy of 97.5% even under adverse weather. Although rain affected near-range detection (FNR up to 17.5% at 100 cm), performance remained robust at mid-to-long ranges. Fog conditions led to lower variance and fewer detection failures. These empirical findings demonstrate the system’s effectiveness and robustness in real-world conditions and emphasize the importance of evaluating both distance accuracy and detection reliability in pedestrian safety applications.

This study evaluates the effectiveness of Stopping for Right-Turning Large Vehicles Policy in Nanjing, designed to mitigate accidents attributed to blind spots and delayed braking of large trucks at intersections. Using high-resolution conflict data from four signalized intersections in Jiangning District, collected via unmanned aerial vehicles (UAVs) and roadside video, the research employs K-means clustering for conflict severity classification and binomial Logit regression to identify critical determinants. Results reveal the policy exhibited limited statistical significance in reducing severe conflicts (p > 0.05). Regression analysis quantified four critical determinants: absence of motorized/non-motorized segregation (OR=1.82, + 81.6% severity odds), elevated stop-line speeds (OR=1.32, + 31.9%), failure to yield (OR=2.45, + 145%), and crossing the street within the zebra crossing (OR=0.19, −81.0%). The analysis demonstrates that infrastructural deficiencies and behavioral non-compliance outweigh the policy’s standalone impact. Based on these findings, the study proposes a holistic optimization framework integrating physical separation measures, enhanced signage, dynamic traffic signal adjustments, and data-driven enforcement strategies. Methodologically, this study innovatively combines unsupervised learning for conflict categorization, providing a scalable framework for evaluating urban traffic policies. This research underscores the necessity of multi-dimensional interventions—spanning infrastructure, enforcement, and public education—to achieve sustainable improvements in intersection safety. The findings offer actionable insights for policymakers to refine regulatory measures and enhance road safety in rapidly urbanizing environments.

Accidents involving cyclists and trucks are among the most severe road accidents. In 2021, 199 cyclists were killed in accidents involving a truck in the EU. The main accident situation is a truck turning right and a cyclist going straight ahead. A large proportion of these accidents are caused by the inadequate visibility in an HGV (Heavy Goods Vehicle). The blind spot, in particular, is a significant contributor to these accidents. A BSD (Blind Spot Detection) system is expected to significantly reduce these accidents. There are only a few studies that estimate the potential of assistance systems, and these studies include a combined assessment of cyclists and pedestrians. In the present study, accident simulations are used to assess a warning and an autonomously intervening assistance system that could prevent truck to cyclist accidents. The main challenges are local sight obstructions such as fences, hedges, etc., rule violations by cyclists, and the complexity of correctly predicting the cyclist’s intentions, i.e., detecting the trajectory. Taking these accident circumstances into consideration, a BSD system could prevent between 26.3% and 65.8% of accidents involving HGVs and cyclists.

Vehicles and roads cooperate to perceive traffic targets, which can reduce the perception blind spots of vehicles and improve driving safety. In this paper, we proposes a vehicle re-identification method oriented to vehicle-road coordination. This method first designs a lightweight vehicle re-identification network based on ShufflenetV2 to solve the computational efficiency problem of vehicle-road coordination scenarios, which can efficiently complete vehicle feature extraction; then, due to the real-time requirements of scenario communication, an adaptive feature conversion mechanism is designed in combination with the LSH algorithm, which can make the re-identification module to dynamically perform binary bit feature conversion and adjust the dimension according to the communication channel state; finally, a loss function for the conversion of vehicle re-identification features is designed, which can greatly reduce the accuracy loss rate of converting floating-point features to bit features. Experiments show that our method can efficiently complete the information extraction and comparison of vehicle re-identification features in the vehicle-road coordination scenario, and can improve the perception efficiency of vehicle-road coordination while taking into account performance and bandwidth.

Pedestrians are more likely to be seriously injured in vehicle collisions. In fact, multiple collisions between vehicles and pedestrians occur on residential roads that lack street-to-sidewalk dividers and have numerous blind spots. Traditional traffic safety features and equipment, such as speed bumps and traffic signs, are not always sufficient to prevent pedestrian accidents on such residential roads. Therefore, we suggest a collision risk warning service for residential roads as a solution to this issue. We use CCTVs with computer vision techniques and radar to accurately detect objects in real-time and to trace their trajectories. In addition, we employ a time-to-collision-based method to identify dangerous situations. The service warns drivers and pedestrians about hazardous situations using a light-emitting diode sign board. We applied our service to three different roads on a university campus in Seoul, Korea, and then conducted a user survey to evaluate the service. In summary, more than 90% of respondents stated that the service was necessary for these specific locations, and 76.9% noted that the service significantly contributed to traffic safety on the campus. This implies that the proposed service improved traffic safety and can be applied to various locations on residential roads.