Search Results (32)

Efficient pedestrian detection plays an important role in many practical daily life applications, such as autonomous cars, video surveillance, and intelligent driving assistance systems. The main goal of pedestrian detection systems, especially in vehicles, is to prevent accidents. By recognizing pedestrians in real time, these systems can alert drivers or even autonomously apply brakes, minimizing the possibility of collisions. However, occlusion is a major obstacle to pedestrian detection. Pedestrians are typically occluded by trees, street poles, cars, and other pedestrians. State-of-the-art detection methods are based on fully visible or little-occluded pedestrians; hence, their performance declines with increasing occlusion level. To meet this challenge, a pedestrian detector capable of handling occlusion is preferred. To increase the detection accuracy for occluded pedestrians, we propose a new method called the Discriminative Deformable Part Model (DDPM), which uses the concept of breaking human image into deformable parts via machine learning. In existing works, human image breaking into deformable parts has been performed by human intuition. In our novel approach, machine learning is used for deformable objects such as humans, combining the benefits and removing the drawbacks of the previous works. We also propose a new pedestrian dataset based on Eastern clothes to accommodate the detector’s evaluation under different intra-class variations of pedestrians. The proposed method achieves a higher detection accuracy on Pascal VOC and VisDrone Detection datasets when compared with other popular detection methods. Full article

In response to the sudden violation of pedestrians crossing the road, intelligent vehicles take into account factors such as the road conditions in the accident zone, traffic rules, and surrounding vehicles’ driving status to make emergency evasive decisions. Thus, the collision simulation models for pedestrians and three types of vehicles, i.e., sedans, Sport Utility Vehicles (SUVs), and Multi-Purpose Vehicle (MPVs), are built to investigate the impact of vehicle types, vehicle steering angles, collision speeds, collision positions, and pedestrian orientations on head injuries of pedestrians. The results indicate that the Head Injury Criterion (HIC) value of the head increases with the increase in collision speed. Regarding the steering angles, when a vehicle’s steering direction aligns with a pedestrian’s position, the pedestrian remains on top of the vehicle’s hood for a longer period and moves together with the vehicle after the collision. This effectively reduces head injuries to pedestrians. However, when the vehicle’s steering direction is opposite to the pedestrian’s position, the pedestrian directly collides with the ground, resulting in higher head injuries. Among them, MPVs cause the most severe injuries, followed by SUVs, and sedans have the least impact. Overall, intelligent vehicles have great potential to reduce head injuries of pedestrians in the event of sudden pedestrian-vehicle collisions by combining with Automatic Emergency Steering (AES) measures. In the future, efforts need to be made to establish an optimized steering strategy and optimize the handling of situations where steering is ineffective or even harmful. Full article

Nearly 3700 people are killed in broadside collisions in the U.S. every year. To reduce broadside collisions, we created and tested the CornerGuard, a prototype system that senses around a corner to alert a car driver of an impending collision with a pedestrian or automobile that is not in the line of sight (LOS). The CornerGuard leverages a microwave-transceiving radar sensor mounted on the car and a curved radio wave reflector installed at the corner to sense around the corner and detect a broadside collision threat. The car’s speed is constantly read by an onboard diagnostics (OBD) system to allow the sensor to differentiate between static objects and objects approaching around the corner. Field testing demonstrated that the CornerGuard can effectively and consistently detect threats at a consistent range without blind spots under broad weather conditions. Our proof of concept study shows that the CornerGuard can be enhanced to be readily integrated into automobile construction and street infrastructure. Full article

Autonomous outdoor moving objects like cars, motorcycles, bicycles, and pedestrians present different risks to the safety of Visually Impaired People (VIPs). Consequently, many camera-based VIP mobility assistive solutions have resulted. However, they fail to guarantee VIP safety in practice, i.e., they cannot effectively prevent collisions with more dangerous threats moving at higher speeds, namely, Critical Moving Objects (CMOs). This paper presents the first practical camera-based VIP mobility assistant scheme, ARAware, that effectively identifies CMOs in real-time to give the VIP more time to avoid danger through simultaneously addressing CMO identification, CMO risk level evaluation and classification, and prioritised CMO warning notification. Experimental results based on our real-world prototype demonstrate that ARAware accurately identifies CMOs (with 97.26% mAR and 88.20% mAP) in real-time (with a 32 fps processing speed for 30 fps incoming video). It precisely classifies CMOs according to their risk levels (with 100% mAR and 91.69% mAP), and warns in a timely manner about high-risk CMOs while effectively reducing false alarms by postponing the warning of low-risk CMOs. Compared to the closest state-of-the-art approach, DEEP-SEE, ARAware achieves significantly higher CMO identification accuracy (by 42.62% in mAR and 10.88% in mAP), with a 93% faster end-to-end processing speed. Full article

Rapid advances in technology and shifting tastes among motorists have reworked the contemporary automobile production sector. Driving is now much safer and more convenient than ever before thanks to a plethora of new technology and apps. Millions of people are hurt every year despite the fact that automobiles are networked and have several sensors and radars for collision avoidance. Each year, many of them are injured in car accidents and need emergency care, and sadly, the fatality rate is growing. Vehicle and pedestrian collisions are still a serious problem, making it imperative to advance methods that prevent them. This paper refines our previous efficient VANET-based pedestrian safety system based on two-way communication between smart cars and the cell phones of vulnerable road users. We implemented the scheme using C and NS3 to simulate different traffic scenarios. Our objective is to measure the additional overhead to protect vulnerable road users. We prove that our proposed scheme adds just a little amount of additional overhead and successfully satisfies the stringent criteria of safety applications. Full article

Nighttime object detection is challenging due to dim, uneven lighting. The IIHS research conducted in 2022 shows that pedestrian anti-collision systems are less effective at night. Common solutions utilize costly sensors, such as thermal imaging and LiDAR, aiming for highly accurate detection. Conversely, this study employs a low-cost 2D image approach to address the problem by drawing inspiration from biological dark adaptation mechanisms, simulating functions like pupils and photoreceptor cells. Instead of relying on extensive machine learning with day-to-night image conversions, it focuses on image fusion and gamma correction to train deep neural networks for dark adaptation. This research also involves creating a simulated environment ranging from 0 lux to high brightness, testing the limits of object detection, and offering a high dynamic range testing method. Results indicate that the dark adaptation model developed in this study improves the mean average precision (mAP) by 1.5−6% compared to traditional models. Our model is capable of functioning in both twilight and night, showcasing academic novelty. Future developments could include using virtual light in specific image areas or integrating with smart car lighting to enhance detection accuracy, thereby improving safety for pedestrians and drivers. Full article

This paper proposes a combined system (OPREVU-AES) that integrates optimized AEB and Automatic Emergency Steering (AES) to generate evasive maneuvers, and it provides an assessment of its effectiveness when compared to a commercial AEB system. The optimized AEB system regulates the braking response through a collision prediction model. OPREVU is a research project in which INSIA-UPM and CEDINT-UPM cooperate to improve driving assistance systems and to characterize pedestrians’ behavior through virtual reality (VR) techniques. The kinematic and dynamic analysis of OPREVU-AES is conducted using CarSim© software v2020.1. The avoidance trajectories are predefined for speeds above 40 km/h, which controls the speed and lateral stability during the overtaking and lane re-entry process. In addition, the decision algorithm integrates information from the lane and the blind spot detectors. The effectiveness evaluation is based on the reconstruction of a sample of vehicle-to-pedestrian crashes (INSIA-UPM database), using PCCrash© software v. 2013, and it considers the probability of head injury severity (ISP) as an indicator. The incorporation of AEB can avoid 53.8% of accidents, with an additional 2.5–3.5% avoided by incorporating automatic steering. By increasing the lateral activation range, the total avoidance rate is increased to 61.8–69.8%. The average ISP reduction is 65%, with significant reductions achieved in most cases where avoidance is not possible. Full article

In recent years, the application of artificial intelligence (AI) in the automotive industry has led to the development of intelligent systems focused on road safety, aiming to improve protection for drivers and pedestrians worldwide to reduce the number of accidents yearly. One of the most critical functions of these systems is pedestrian detection, as it is crucial for the safety of everyone involved in road traffic. However, pedestrian detection goes beyond the front of the vehicle; it is also essential to consider the vehicle’s rear since pedestrian collisions occur when the car is in reverse drive. To contribute to the solution of this problem, this research proposes a model based on convolutional neural networks (CNN) using a proposed one-dimensional architecture and the Inception V3 architecture to fuse the information from the backup camera and the distance measured by the ultrasonic sensors, to detect pedestrians when the vehicle is reversing. In addition, specific data collection was performed to build a database for the research. The proposed model showed outstanding results with 99.85% accuracy and 99.86% correct classification performance, demonstrating that it is possible to achieve the goal of pedestrian detection using CNN by fusing two types of data. Full article

The need for continuous research to refine the models used in forensic accident reconstruction appears with the development of new car models that satisfy consumer complaints. This paper analyzed a sub-sequence of car and pedestrian accidents from the perspective of the distance traveled by them in the contact phase with the aim of improving the information regarding the reconstruction of road accidents. This research included the analysis of some real tests with pedestrian dummies, as well as simulations of the impact between different classes of vehicles and pedestrians in two different walking positions. Specialized software was used with complex multibody models of pedestrians, modifying the speed and deceleration parameters of the car at the time of the collision. For pedestrian characteristics, the friction coefficients of the ground, car and its mass were modified. The research results highlight the differences between the bilinear models used in accident reconstruction and the proposed study. They can also be used to determine the distance traveled by the vehicle in the first phase of a collision with pedestrians. Full article

The importance of an external interaction interface (eHMI) has grown in recent years. Most eHMI concepts focus on communicating autonomous vehicle (AV)’s yielding intention to pedestrians at a crossing. However, according to previous studies, pedestrians at a crossing rely mainly on the vehicle’s movement information (implicit communication) rather than information from eHMIs (explicit communication). This paper has the purpose of proposing a specific use case in which the eHMI of future AVs could play an indispensable role in the safety of other road users (ORUs). Often VRUs cannot see the traffic flow due to a series of parked or stopped vehicles, which is a frequent cause of fatal traffic collision accidents. Drivers may also not be able to see approaching pedestrians or other cars from the side for the same reason. In this paper, the impact of an eHMI is tested from the perspective of drivers with limited visibility when a jaywalker steps into the road. A combination of colors, shapes, and information levels is presented on an eHMI. We show that our proposed eHMI design, in the deadlock scenario of a jaywalker and a driver who both lack visibility, significantly reduced the reaction time compared to when there was no eHMI. In the experiment, the willingness to stop, varying from 0 to 5, was measured from the driver’s perspective. The results showed that most users felt uncertainty and did not move quickly when seeing the light band color alone. Textual information on the eHMI was significantly more effective in providing an urgent warning of this specific scenario than vertical and horizontal light bands with color without text. In addition, red color, blinking rapidly above 3 Hz, and egocentric messages were also necessary to reduce the PRT(perception response time). By using text-added eHMI (Vertical + Text eHMI), the mean time to achieve a score above 4 for willingness to stop was 2.113 s faster than when there was no eHMI. It was 2.571 s faster than the time until the slider of the participants reached the maximum level for willingness to stop. This is a meaningful amount of difference when considering a PRT of 2.5 s, which is the Korean road design standard. As eHMIs tend to be applied for smarter mobility, it is expected that they will be more effective in preventing accidents if the eHMI is standardized in autonomous driving level 2 to 3 vehicles driven by humans before fully autonomous driving becomes a reality. Full article

In this paper, multispectral pedestrian detection is mainly discussed, which can contribute to assigning human-aware properties to automated forklifts to prevent accidents, such as collisions, at an early stage. Since there was no multispectral pedestrian detection dataset in an intralogistics domain, we collected a dataset; the dataset employs a method that aligns image pairs with different domains, i.e. RGB and thermal, without the use of a cumbersome device such as a beam splitter, but rather by exploiting the disparity between RGB sensors and camera geometry. In addition, we propose a multispectral pedestrian detector called SSD 2.5D that can not only detect pedestrians but also estimate the distance between an automated forklift and workers. In extensive experiments, the performance of detection and centroid localization is validated with respect to evaluation metrics used in the driving car domain but with distinct categories, such as hazardous zone and warning zone, to make it more applicable to the intralogistics domain. Full article

While Product-Service Systems (PSS) have a potential sustainability impact by increasing a product’s life and reducing resource consumption, the lack of ownership might lead to less responsible user behavior. Smart PSS can overcome this obstacle and guarantee correct and safe PSS use. In this context, intelligent connected vehicles (ICVs) with PSS can effectively reduce traffic accidents and ensure the safety of vehicles and pedestrians by guaranteeing optimal and safe vehicle operation. A core subsystem to support that is the collision-warning system (CWS). Existing CWSs are, however, limited to in-car warning; users have less access to the warning information, so the result of CWS for collision avoidance is insufficient. Therefore, CWS needs to be extended to include more elements and stakeholders in the collision scenario. This paper aims to provide a novel understanding of extended CWS (ECWS), outline the conceptual framework of ECWS, and contribute a conceptual modeling approach of ECWS from the smart PSS perspective at the functional level. It defines an integrated solution of intelligent products and warning services. The function is modeled based on the Theory of Inventive Problem Solving (TRIZ). Functions of an ECWS from the perspective of smart PSS can be comprehensively expressed to form an overall solution of integrated intelligent products, electronic services, and stakeholders. Based on the case illustration, the proposed method can effectively help function modeling and development of the ECWS at a conceptual level. This can effectively avoid delays due to traffic accidents and ensure the safety of vehicles and pedestrians. Full article

During major disasters, such as a subduction earthquake and the associated tsunami, combinations of uncommon conditions such as non-functioning traffic signals, a large number of pedestrians on traffic lanes, and debris scattered on roads can be widespread. It is vital to take these uncommon conditions into account since they can significantly influence the evacuation progress. Agent-Based Models (ABMs) with capabilities to reproduce evacuees’ behaviors as emergent phenomena is promising method to simulate combinations of such rare conditions. This paper presents a new model to cover the current research gap in accurately modeling car–car and car–pedestrian interactions at non-signalized junctions. Specifically, the details of accurately approximating car trajectories at junctions and automated construction, approximating free-flow speed of cars along curved trajectories, and accurately calculating the points of collision and time to collision are presented. As a demonstrative application, we simulated a hypothetical evacuation scenario with non-functioning traffic signals in which different numbers of slow evacuees are allowed to use cars. While the ABM is yet to be thoroughly validated, the presented demonstrative scenarios indicates that a considerable number of the needy can be allowed to use cars for evacuation if their routes and evacuation start time window are well planned. Full article

Traffic collisions pose a serious problem for cities due to the annually increasing number of vehicles. Information about incidents that occur on roads is important for the corresponding monitoring bodies, authorities, and emergency services. To ensure traffic safety, the data have to be visible, clean, and transparently displayed. This research was, therefore, aimed at developing a methodology for monitoring motor vehicle collision data and applying visualization techniques to evidence from New York City. The method showed that the largest number of motor vehicle traffic crashes occurred in Lower Manhattan due to its high population and traffic density. With these data, the road agencies of the city can put potentially dangerous road sections under control and make them safer for both drivers and pedestrians. Further development of the system may be associated with data analytics and visualization, resulting in new layers of heatmaps that not only provide details on car collision hotspots, which serve as the main target indicator for traffic safety authorities, but also break them down into social facilities, such as schools. This feature will enable assessment of how safe it is around a school and the evaluation of the impact of an underpass or a traffic enforcement camera on the number of collisions. The motor vehicle traffic crash (MVTC) monitoring system will help in comparing city districts and regions in terms of safety, seeing trends, realizing what exactly is happening at interchanges, and understanding the reasons behind. The methodology, in addition, can be supplemented with an analysis of risk factors for MVTCs, the efficiency of adopted measures and road renovations that are carried out, and many other functions. Full article

V2X is used for communication between the surrounding pedestrians, vehicles, and roadside units. In the Forward Collision Warning (FCW) of Phase One scenarios in V2X, multimodal modalities and multiple warning stages are the two main warning strategies of FCW. In this study, three warning modalities were introduced, namely auditory warning, visual warning, and haptic warning. Moreover, a multimodal warning and a novel multi-staged HUD warning were established. Then, the above warning strategies were evaluated in objective utility, driving performance, visual workload, and subjective evaluation. As for the driving simulator of the experiment, SCANeR was adopted to develop the driving scenario and an open-cab simulator was built based on Fanatec hardware. Kinematic parameters, location-related data and eye-tracking data were then collected. The results of the Analysis of Variance (ANOVA) indicate that the multimodal warning is significantly better than that of every single modality in utility and longitudinal car-following performance, and there is no significant difference in visual workload between multimodal warning and the baseline. The utility and longitudinal driving performance of multi-staged warning are also better than those of single-stage warning. Finally, the results provide a reference for the warning strategy design of the FCW in Intelligent Connected Vehicles. Full article