Pedestrian Size Research Articles

A virtual test system representing the distribution of pedestrian impact configurations for future vehicle front-end optimization

ABSTRACTObjectives: The purpose of this study is to define a computationally efficient virtual test system (VTS) to assess the aggressivity of vehicle front-end designs to pedestrians considering the distribution of pedestrian impact configurations for future vehicle front-end optimization. The VTS should represent real-world impact configurations in terms of the distribution of vehicle impact speeds, pedestrian walking speeds, pedestrian gait, and pedestrian height. The distribution of injuries as a function of body region, vehicle impact speed, and pedestrian size produced using this VTS should match the distribution of injuries observed in the accident data. The VTS should have the predictive ability to distinguish the aggressivity of different vehicle front-end designs to pedestrians.Methods: The proposed VTS includes 2 parts: a simulation test sample (STS) and an injury weighting system (IWS). The STS was defined based on MADYMO multibody vehicle to pedestrian impact simulations accounting for the range of vehicle impact speeds, pedestrian heights, pedestrian gait, and walking speed to represent real world impact configurations using the Pedestrian Crash Data Study (PCDS) and anthropometric data. In total 1,300 impact configurations were accounted for in the STS. Three vehicle shapes were then tested using the STS. The IWS was developed to weight the predicted injuries in the STS using the estimated proportion of each impact configuration in the PCDS accident data. A weighted injury number (WIN) was defined as the resulting output of the VTS. The WIN is the weighted number of average Abbreviated Injury Scale (AIS) 2+ injuries recorded per impact simulation in the STS. Then the predictive capability of the VTS was evaluated by comparing the distributions of AIS 2+ injuries to different pedestrian body regions and heights, as well as vehicle types and impact speeds, with that from the PCDS database. Further, a parametric analysis was performed with the VTS to assess the sensitivity of the injury predictions to changes in vehicle shape (type) and stiffness to establish the potential for using the VTS for future vehicle front-end optimization.Results: An STS of 1,300 multibody simulations and an IWS based on the distribution of impact speed, pedestrian height, gait stance, and walking speed is broadly capable of predicting the distribution of pedestrian injuries observed in the PCDS database when the same vehicle type distribution as the accident data is employed. The sensitivity study shows significant variations in the WIN when either vehicle type or stiffness is altered.Conclusions: Injury predictions derived from the VTS give a good representation of the distribution of injuries observed in the PCDS and distinguishing ability on the aggressivity of vehicle front-end designs to pedestrians. The VTS can be considered as an effective approach for assessing pedestrian safety performance of vehicle front-end designs at the generalized level. However, the absolute injury number is substantially underpredicted by the VTS, and this needs further development.

Power laws in the cluster sizes of group-crossing pedestrians: empirical evidence and modelling of causes

.Pedestrian-vehicle mixed traffic at unsignalized crosswalks is a self-organizing system, because of the absence of external control. Consequently, pedestrian group-crossing behaviour may be expected to exhibit the power-law phenomenon, which is often an indicator of self-organized criticality. The assumption that the distribution of cluster sizes may obey power laws was theoretically supported by Xin et al. To test this claim empirically, we observed pedestrian collective motion under natural conditions, using video recordings. These real data confirmed the existence of the power-law phenomenon. A simulation study was then undertaken to uncover the mechanism that generated the observations. The observed patterns could be fully replicated if the behaviour of pedestrians and motorists were modelled, in particular the yielding behaviour of motorists, when approaching crossings and the adaptive behaviour of pedestrians as they increased in number. As well as providing empirical evidence, our findings also provide a deeper understanding of the patterns of the pedestrian collective motion. The results may have implications in the validation of mixed traffic simulation models.

Assessment of antenna characteristic effects on pedestrian and cyclists travel-time estimation based on Bluetooth and WiFi MAC addresses

Monitoring pedestrian and cyclists movement is an important area of research in transport, crowd safety, urban design and human behaviour assessment areas. Media Access Control (MAC) address data has been recently used as potential information for extracting features from people’s movement. MAC addresses are unique identifiers of WiFi and Bluetooth wireless technologies in smart electronics devices such as mobile phones, laptops and tablets. The unique number of each WiFi and Bluetooth MAC address can be captured and stored by MAC address scanners. MAC addresses data in fact allows for unannounced, non-participatory, and tracking of people. The use of MAC data for tracking people has been focused recently for applying in mass events, shopping centres, airports, train stations, etc. In terms of travel time estimation, setting up a scanner with a big value of antenna’s gain is usually recommended for highways and main roads to track vehicle’s movements, whereas big gains can have some drawbacks in case of pedestrian and cyclists. Pedestrian and cyclists mainly move in built distinctions and city pathways where there is significant noises from other fixed WiFi and Bluetooth. Big antenna’s gains will cover wide areas that results in scanning more samples from pedestrians and cyclists’ MAC device. However, anomalies (such fixed devices) may be captured that increase the complexity and processing time of data analysis. On the other hand, small gain antennas will have lesser anomalies in the data but at the cost of lower overall sample size of pedestrian and cyclist’s data. This paper studies the effect of antenna characteristics on MAC address data in terms of travel-time estimation for pedestrians and cyclists. The results of the empirical case study compare the effects of small and big antenna gains in order to suggest optimal set up for increasing the accuracy of pedestrians and cyclists’ travel-time estimation.

Analysis of test tools for evaluation of contact-based sensor systems for pedestrian detection

Active systems of the passive vehicle safety, like deployable bonnets or external airbags, are installed in series vehicles since 2005 to fulfil pedestrian protection requirements. However, an appropriate test procedure for these systems was not available until 2010, when Euro NCAP published a revised version of the pedestrian test protocol (version 5.2). In this test protocol, a test method to evaluate contact-based sensor systems for pedestrian detection using so-called hardest-to-detect impactors is described. These impactors shall reproduce the lowest possible impact load induced in the vehicle front in case of a collision with a pedestrian, which is strongly dependent on the anthropometry of the pedestrian as well as the impact configuration. In this paper a holistic analysis of the accident kinematics of pedestrian–vehicle collisions and resulting loads induced in the vehicle front is conducted. The vehicle front is modelled using a generic vehicle buck, which is specifically designed for this analysis. As can be observed by the conducted simulations, the pedestrian size inducing the lowest possible impact load varies for different vehicle categories. Thus, it is not sufficient to consider only one pedestrian size. The comparison of these vehicle–pedestrian simulations with the loads being induced by three hardest-to-detect impactors shows that none of the three impactors is able to reproduce the ‘worst-case scenario’ for all analysed four vehicle categories. The results of this study, which was conducted on behalf of and funded by the German Federal Highway Research Institute (BASt), could be used for the development of a suitable, effective and objective test procedure for active systems of the passive vehicle safety with regard to pedestrian protection.

Joint global–local information pedestrian detection algorithm for outdoor video surveillance

The pedestrian size is usually small in practical outdoor surveillances. The small-scale pedestrian detection for outdoor surveillances is an important but difficult issue due to the limited information and the background interference. According to human cognition, the global information is important for the pedestrian detection. Therefore, a joint global–local information pedestrian detection algorithm is proposed to fully exploit and utilize the global information. The LBP feature is explicitly extracted from the low-frequency component of original images, which are utilized as the global information to suppress the background interference and enrich the description of pedestrian. Moreover, a structure-LBP is proposed to apply the inherent topology structure of human body to LBP. The structure-LBP feature extracted from original images can achieve a more discriminative description of pedestrians compared with the original LBP. The experimental results demonstrate that the proposed algorithm can improve the overall recognition performance for the small-scale pedestrians.

On Critical Densities and Velocities for Pedestrians Entering a Crowd

The problem of pedestrians trying to enter and/or advance against an incoming crowd is studied. The analysis shows that starting with a very limited set of assumptions (elliptical cross-section of pedestrians; constant ratio of forward to lateral separation) one is able to derive from purely kinematic considerations critical densities beyond which it is impossible for pedestrians to enter and/or advance into an incoming crowd. The results obtained indicate that for the common pedestrian size of a = 0.5 m, b = 0.3 m, the limit densities range from ρ = O(5.0 – 6.06) [p/m2], in good agreement with empirical observations.

적분영상 기반 특징 정보 예측을 통한 고속 보행자 검출

본 논문은 특징 정보 예측을 통한 빠른 보행자 검출 기법을 제안한다. 다양한 크기의 보행자를 검출하기 위해 보행자 모델의 크기나 입력영상의 크기를 변화시킨다. 보행자 모델의 크기를 변화시킬 경우 크기별 모델이 필요하며, 보행자 모델의 크기의 축소시키는 경우 모델 정보를 손상시킨다. 보행자 모델의 다양한 크기별 보행자의 특징을 추출해야 하므로 보행자 특징의 추출은 전체 수행시간 중 가장 많은 시간을 필요로 한다. 따라서 본 논문은 영상 크기에 따라 특징 추출을 반복하지 않고 입력영상에서 얻어진 특징 정보의 예측을 통해 보행자 검출의 특징추출을 수행한다. 제안하는 방법의 효율성을 검증하기 위해 다양한 채널을 가진 ChnFtrs 특징 및 Adaboost 알고리즘을 사용과 학습과 실험을 위한 영상으로 INRIA 보행자 DB를 사용하였다. This paper enhances the speed of a pedestrian detection using an estimation of feature information based on integral image. Pedestrian model or input image should be resized to the size of various pedestrians. In case that the size of pedestrian model would be changed, pedestrian models with respect to the size of pedestrians should be required. Reducing the size of pedestrian model, however, deteriorates the quality of the model information. Since various features according to the size of pedestrian models should be extracted, repetitive feature extractions spend the most time in overall process of pedestrian detection. In order to enhance the processing time of feature extraction, this paper proposes the fast extraction of pedestrian features based on the estimate of integral image. The efficiency of the proposed method is evaluated by comparative experiments with the Channel Feature and Adaboost training using INRIA person dataset.

구배 기반 가중 접근성을 이용한 역세권 보행 네트워크 분석에 관한 연구

최근보행은 교통의 가장 기본 이동수단으로 대중교통 중심개발, 지속가능한 개발, 녹색교통 활성화 등 새로운 패러다임의 부각으로 그 중요성과 관심도가 높아지고 있다. 기존의 보행 네트워크 분석 연구에서는 보행자의 목적지까지 이동거리와 공간구문론의 통합도를 대표적인 접근성 요소로 사용하고 있지만 단순한 평면상의 네트워크 이동거리는 지형적인 특성이 반영되지 않기 때문에 비슷한 네트워크 구조를 가진 분석 지역들의 경우 비슷한 결과 값을 나타낼 수 있다. 또한 분석 지역의 특성에 따라 보행자들의 이동규모와 이동거리가 다름에도 불구하고 이를 반영하지 못한다. 따라서 본 연구에서는 구배를 고려하고 보행자의 이동규모를 고려한 이동거리 분석을 통해 새로운 보행 네트워크 접근성 분석 방법론을 제시하는 것이 목적이다. 강남구에 위치한 논현역, 역삼역, 학동역을 연구지역으로 선정하고, 미국 스포츠의학협회에서 제시하는 대사량 공식을 통해 구배 가중거리 공식을 유도하고 보행자의 이동규모를 추정하기 위해 건물 면적 비율에 따라 역 이용객 수 데이터를 활용해 건물 이용자 수를 추정한 뒤 연구 지역의 1인당 평균 이동거리를 비교하고 분석하였다. 더 나아가 역세권 200m와 500m에 위치한 건물에 대한 보행자의 이동거리 분석을 실시하였으며 이는 보다 연구지역의 특성을 반영하는 것으로 나타났다. 따라서 본 연구는 단순 보행 네트워크 이동거리 분석에서 구배를 고려한 보행 네트워크 이동거리 분석, 더 나아가 보행자의 이동규모를 고려한 보행 네트워크 이동거리 분석을 실시하였고 보다 실질적인 보행자의 특성을 고려한 보행 네트워크 접근성 분석 방법론을 제시하였다. 이는 앞으로 보행 네트워크 분석에서 지형적 특성을 반영하고 보다 실질적인 보행자의 움직임을 고려한 보행 네트워크 분석 방법론으로 활용될 것으로 기대된다. Walking is the most basic personal mobility and its importance and concern is ever increasing with the highlighting of a new paradigm, such as transit oriented development, sustainable development and revitalization of green transport. The existing analytical research on pedestrian network is using a pedestrian's moving distance to a destination and integration in space syntax theory as its representative accessibility factors. However, the uniplanar network moving distance fails to reflect topographic characteristics, so the moving distance could show a similar result value in case of the regions for analysis that have a similar network structure to each other. Accordingly, the aim of this study is to suggest a new analytical methodology on pedestrian network accessibility in consideration of the grade in pedestrian sections and a pedestrian's size. this study, in its analysis of a uniplanar pedestrian network moving distance, analyzed the pedestrian network moving distance in consideration of the grade in pedestrian sections, and even the pedestrian network moving distance in consideration of a pedestrian's size, and suggested the methodology on pedestrian network accessibility analysis in consideration of a more substantive pedestrian's characteristics. It is hoped that the methodology used by this study will be used as the methodology on pedestrian network analysis which can reflect topographic characteristics in the pedestrian network analysis, and take a more substantive pedestrian's movement into account.

Effects of Vehicle Impact Velocity, Vehicle Front-End Shapes on Pedestrian Injury Risk

Objective: This study aimed at investigating the effects of vehicle impact velocity, vehicle front-end shape, and pedestrian size on injury risk to pedestrians in collisions with passenger vehicles with various frontal shapes. Method: A series of parametric studies was carried out using 2 total human model for safety (THUMS) pedestrian models (177 and 165 cm) and 4 vehicle finite element (FE) models with different front-end shapes (medium-size sedan, minicar, one-box vehicle, and sport utility vehicle [SUV]). The effects of the impact velocity on pedestrian injury risk were analyzed at velocities of 20, 30, 40, and 50 km/h. The dynamic response of the pedestrian was investigated, and the injury risk to the head, chest, pelvis, and lower extremities was compared in terms of the injury parameters head injury criteria (HIC), chest deflection, and von Mises stress distribution of the rib cage, pelvis force, and bending moment diagram of the lower extremities. Result: Vehicle impact velocity has the most significant influence on injury severity for adult pedestrians. All injury parameters can be reduced in severity by decreasing vehicle impact velocities. The head and lower extremities are at greater risk of injury in medium-size sedan and SUV collisions. The chest injury risk was particularly high in one-box vehicle impacts. The fracture risk of the pelvis was also high in one-box vehicle and SUV collisions. In minicar collisions, the injury risk was the smallest if the head did not make contact with the A-pillar. Conclusion: The vehicle impact velocity and vehicle front-end shape are 2 dominant factors that influence the pedestrian kinematics and injury severity. A significant reduction of all injuries can be achieved for all vehicle types when the vehicle impact velocity is less than 30 km/h. Vehicle designs consisting of a short front-end and a wide windshield area can protect pedestrians from fatalities. The results also could be valuable in the design of a pedestrian-friendly vehicle front-end shape. [Supplementary materials are available for this article. Go to the publisher's online edition of Traffic Injury Prevention for the following free supplemental resource: Head impact conditions and injury parameters in four-type vehicle collisions and validation result of the finite element model of one-box vehicle and minicar. ]

Pedestrian Detection: An Evaluation of the State of the Art

Pedestrian detection is a key problem in computer vision, with several applications that have the potential to positively impact quality of life. In recent years, the number of approaches to detecting pedestrians in monocular images has grown steadily. However, multiple data sets and widely varying evaluation protocols are used, making direct comparisons difficult. To address these shortcomings, we perform an extensive evaluation of the state of the art in a unified framework. We make three primary contributions: 1) We put together a large, well-annotated, and realistic monocular pedestrian detection data set and study the statistics of the size, position, and occlusion patterns of pedestrians in urban scenes, 2) we propose a refined per-frame evaluation methodology that allows us to carry out probing and informative comparisons, including measuring performance in relation to scale and occlusion, and 3) we evaluate the performance of sixteen pretrained state-of-the-art detectors across six data sets. Our study allows us to assess the state of the art and provides a framework for gauging future efforts. Our experiments show that despite significant progress, performance still has much room for improvement. In particular, detection is disappointing at low resolutions and for partially occluded pedestrians.

Evaluation of the kinematics and injury potential to different sizes of pedestrians impacted by a utility vehicle with a frontal guard

A review of accident data reveals that in most pedestrian accidents, the head and lower extremity injuries are the predominant areas of injury to the pedestrian. The vehicle front geometry profile and stiffness, as well as impact speed are important factors governing pedestrian kinematics. Accident data show that the fatality rate for pedestrian/utility vehicle impact is greater than that for pedestrian/passenger car impact. The addition of a front guard on light trucks and sports utility vehicles to mitigate damage during off-road activity, or to provide mounting points for extra lights, makes the pedestrian more vulnerable to the impact. In this study, a computational technique is utilised to quantify the influence of the added front guard on the impacted pedestrian. A computer-aided design model of a typical commercial frontal guard is developed, and the finite element analysis, along with impact test, is conducted to obtain the stiffness properties of the guard. Different sizes of pedestrian models in the MADYMO code are utilised, and the validated facet-surface model of a pickup truck is used to generate a vehicle front surface. The entire model is validated by comparing the pedestrian kinematics and injury parameters with the published data. This study demonstrates that for all sizes of pedestrians, the mid-body region is more vulnerable when a guard is added to the vehicle. The results from this study can be utilised in the designing of front guards, the frontal crash zone of utility vehicles and installation of these aftermarket guards in order to protect vulnerable road users.

Analysis of loading of lower extremities based on bending moment in car-to-pedestrian collisions

In this study, simulations of car-to-pedestrian collisions were conducted using multibody pedestrian models and a finite-element (FE) car model in order to examine the loadings on the lower extremity of the pedestrian. Bending moment diagrams were plotted versus time for the tibia, knee and femur; and the loadings on the lower extremities were evaluated using time-series comparisons. An analysis using a human FE pedestrian model also was conducted. The bending moment of the lower extremity was found to be comparable between the multibody and FE models. Car-to-pedestrian collisions were also simulated for various sizes of pedestrians. The bending moments were larger as the pedestrian size was larger, whereas the curve shape of the bending moment diagrams was similar for all sizes of pedestrians. It was shown that the bending moment diagrams using the multibody pedestrian models were useful for investigating the loading on the lower extremities for various sizes of pedestrians.

Experimental study on microscopic moving characteristics of pedestrians in built corridor based on digital image processing

Experiments of moving characteristics of Chinese pedestrians in a built corridor were conducted. Pedestrians’ trajectories were first extracted with digital image processing method and then mapped into real space coordinates by adopting a direct linear transformation approach. Moving characteristics of single pedestrian, interaction between pedestrian and the built corridor, and also interaction between pair pedestrians were analyzed and compared with those of the experiments with French pedestrians. It is found that the average relaxation time of the Chinese pedestrians is 0.714 s, and the mean free speed is 1.51 m/s, which are both larger than those of the French pedestrians. Meanwhile, the Chinese pedestrians also kept a distance of about 0.4 m to the wall. When walking too close to the wall, the pedestrian has a tendency to walk far away, which is quantified as exponential decay force. Given a same distance, the force from the wall on the French pedestrian is larger than on the Chinese pedestrian due to the different pedestrian size and personal space. When one pedestrian evade another standing still pedestrian, the interaction between them shows a non-isotropic feature. We found that the right preference in Chinese pedestrians is more obvious than in French pedestrians. We further quantified the interaction among pedestrians, and found the force from those who locate on the right-forward direction doesn’t change much while from those who locate on the left-forward direction varies with the increase of distance. The basic data presented in this study can be used to validate and refine pedestrian traffic models.

Using Image-Based Metrics to Model Pedestrian Detection Performance With Night-Vision Systems

The primary purpose of night-vision systems in civilian vehicles is to help drivers detect pedestrians. Pedestrian detection distance with night-vision systems has been modeled based on image metrics. However, the probability of pedestrian detection, in particular considering the factor of distance, has not been modeled based on image metrics. In this paper, we first describe a model of the probability of pedestrian detection, which compares several combinations of image-based clutter, contrast, and pedestrian size metrics using a simple mathematical equation. Next, we describe a model of the probability of pedestrian detection as a function of distance and image-based metrics by combining the model of pedestrian-detection probability and a model that represents the relationship between the distance to a pedestrian and an image-based pedestrian size metric. In the final model, image-based metrics are used to predict pedestrian-detection performance and can also be used to evaluate and support the development of night-vision systems in vehicles.

Exploring Trampling and Crushing in a Crowd

Using published field data on accidents in crowds involving either the trampling or crushing of pedestrians, a theory has been developed to model similar dangerous situations. Whether a particular situation is fatal is statistical, but general rules may be established. In cases involving trampling, which occurs when pedestrians are moving, it was found that the density of pedestrians is an important isotropic quantity that appears to determine the probability of a fatal accident. Previous studies have also identified the density of a crowd as determining the probability of crushing within a stationary crowd, although different densities are involved in the initiation of these two types of accidents. In a moving crowd, trampling will occur before crushing, but trampling does not occur in a static crowd. The cross-sectional size of pedestrians within the crowd is important in determining whether a particular density is dangerous in both types of accidents. Consequently, comparisons of accidents involving different groups of pedestrians, possibly in different countries, need to be standardized for the characteristics of the group.

Pedestrian Visibility Under Automobile Low-Beam Headlight Illumination: With and Without Headlight Covers

Pedestrians and bicyclists are the most vulnerable of all participants in traffic. Their safety at night should be a priority for headlamp designers, rule-making agencies, and consumers alike. A disturbing trend—the use of very dark aftermarket headlamp covers on vehicle headlamps—is emerging. This trend is observed primarily in sports cars owned by young drivers. These dark headlamp covers, often referred to as blackouts, serve no real purpose other than that of supposedly enhancing vehicle appearance. For a driver using such headlamp covers at night, it may appear as though the overall visibility is not much affected. However, this perception is misleading and wrong, as will be clearly demonstrated. The objective was to quantify the detrimental effects dark headlamp covers have on pedestrian visibility. To do so, detection distances were obtained in the field using 15 observers who were approaching stationary pedestrian mockups. The independent variables were pedestrian size (child, adult), diffuse clothing reflectance (dark, light), and illumination (with and without headlamp covers). Clothing reflectance turned out to have the strongest effect on the detection distance, closely followed by the factor of illumination. The conclusions that can be drawn from the work presented are simple. Nighttime pedestrians should not wear dark clothing, instead they should wear white or reflective clothing, and motorists should not drive with dark headlamp covers at night.

Stereo- and neural network-based pedestrian detection

Pedestrian detection is essential to avoid dangerous traffic situations. We present a fast and robust algorithm for detecting pedestrians in a cluttered scene from a pair of moving cameras. This is achieved through stereo-based segmentation and neural network-based recognition. The algorithm includes three steps. First, we segment the image into sub-image object candidates using disparities discontinuity. Second, we merge and split the sub-image object candidates into sub-images that satisfy pedestrian size and shape constraints. Third, we use intensity gradients of the candidate sub-images as input to a trained neural network for pedestrian recognition. The experiments on a large number of urban street scenes demonstrate that the proposed algorithm: (1) can detect pedestrians in various poses, shapes, sizes, clothing, and occlusion status; (2) runs in real-time; and (3) is robust to illumination and background changes.

Visual Target Detection Models for Civil Twilight and Night Driving Conditions

A luminance contrast-based computer visibility model is discussed and compared with the civil twilight method, which has recently been introduced. The civil twilight method attempts to predict the visibility of ordinary objects (reflectance 3 to 79 percent, average size) by using only the headlamp illuminance at the target. It is suggested that the one-factor approach used by the civil twilight method is insufficient to satisfactorily address target visibility in the field. Developers of more advanced visibility models generally attempt to design models based on the current state of the visibility research and with enough capability to obtain a reasonable degree of realism. The level of the benchmark illuminance (3.2 lx) used in the civil twilight method is considered by the authors to be too high, leading to very short detection distances for pedestrians under automobile headlamp illumination at night. The developers of the civil twilight method claim that the 3.2-lx visibility benchmark is based on systematic visual observations made by astronomers over a century ago. The use of the civil twilight method for pedestrian detection under automobile headlamp illumination at night is strongly discouraged by the authors of this paper, because the method may be misused by forensic experts if there is a need to produce arbitrarily short pedestrian detection distances, regardless of factors such as clothing reflectance, contrast, pedestrian size, windshield transmittance, and atmospheric transmissivity.

Effect on Vehicle-Pedestrian Conflicts of “Turning Traffic Must Yield to Pedestrians” Sign

Left- and right-turn movements at signalized intersections have been found to be three to six times more hazardous to pedestrians than through movements mainly because drivers fail to observe or yield the right of way to pedestrians. The objective was to evaluate the Turning Traffic Must Yield to Pedestrians sign, which was aimed at reminding turning motorists of their legal obligation to yield the right of way to pedestrians and, consequently, reducing vehicle-pedestrian conflicts. The sign was installed at 12 marked crosswalks in two cities, and vehicle-pedestrian conflict data were collected before and after its installation. The sign was effective in reducing left-turn conflicts 20 to 65 percent and right-turn conflicts 15 to 30 percent; both reductions were statistically significant at the 0.05 level. The sign was significantly more effective in reducing left-turn conflicts than it was for right-turn conflicts. Percentage conflicts between pedestrians and turning traffic decreased as the pedestrian group size increased. In other words, larger pedestrian group sizes encountered fewer conflicts with turning traffic. This result is true for both left- and right-turn movements, regardless of the presence of the sign. Based on the results of this study, it was recommended that the “Turning Traffic Must Yield to Pedestrians” sign be considered for inclusion in the Manual on Uniform Traffic Control Devices.