In the context of banning gated communities, blocks returning to the human-oriented scale become the new normal, and pedestrian system design will be paid more attention in the urban planning field. Oct-Loft Creative ...In the context of banning gated communities, blocks returning to the human-oriented scale become the new normal, and pedestrian system design will be paid more attention in the urban planning field. Oct-Loft Creative Park is a template for open blocks in Shenzhen, with a convenient and humanized pedestrian system. This paper selects the creative park's pedestrian system as the research object, using the environment-behavior theory for analysis. Finally, optimization strategies of pedestrian system will be put forward.展开更多
The Multi-dimensional Pedestrian System( MPS) is an integral part of the new campus of University of Macao. It observes the principle of "pedestrian first " and features a pedestrian-vehicle dividing system....The Multi-dimensional Pedestrian System( MPS) is an integral part of the new campus of University of Macao. It observes the principle of "pedestrian first " and features a pedestrian-vehicle dividing system.Based on a comprehensive analysis of the location,climate,ecology and other factors of the project site,the conception of the idea of MPS and the related researches are illustrated. The transportation features of the MPS,as summarized,include multi-dimensions,short-distance and weather-resistance. Its features for the sake of livability include integration of nature, respect for the environment and sharing of landscape. Upon the completion of the project, the effects on its users were tested. Finally, some constructive rules for the construction of similar campus pedestrian systems were proposed.展开更多
Pedestrian positioning system(PPS)using wearable inertial sensors has wide applications towards various emerging fields such as smart healthcare,emergency rescue,soldier positioning,etc.The performance of traditional ...Pedestrian positioning system(PPS)using wearable inertial sensors has wide applications towards various emerging fields such as smart healthcare,emergency rescue,soldier positioning,etc.The performance of traditional PPS is limited by the cumulative error of inertial sensors,complex motion modes of pedestrians,and the low robustness of the multi-sensor collaboration structure.This paper presents a hybrid pedestrian positioning system using the combination of wearable inertial sensors and ultrasonic ranging(H-PPS).A robust two nodes integration structure is developed to adaptively combine the motion data acquired from the single waist-mounted and foot-mounted node,and enhanced by a novel ellipsoid constraint model.In addition,a deep-learning-based walking speed estimator is proposed by considering all the motion features provided by different nodes,which effectively reduces the cumulative error originating from inertial sensors.Finally,a comprehensive data and model dual-driven model is presented to effectively combine the motion data provided by different sensor nodes and walking speed estimator,and multi-level constraints are extracted to further improve the performance of the overall system.Experimental results indicate that the proposed H-PPS significantly improves the performance of the single PPS and outperforms existing algorithms in accuracy index under complex indoor scenarios.展开更多
Pedestrian self-organizing movement plays a significant role in evacuation studies and architectural design.Lane formation,a typical self-organizing phenomenon,helps pedestrian system to become more orderly,the majori...Pedestrian self-organizing movement plays a significant role in evacuation studies and architectural design.Lane formation,a typical self-organizing phenomenon,helps pedestrian system to become more orderly,the majority of following behavior model and overtaking behavior model are imprecise and unrealistic compared with pedestrian movement in the real world.In this study,a pedestrian dynamic model considering detailed modelling of the following behavior and overtaking behavior is constructed,and a method of measuring the lane formation and pedestrian system order based on information entropy is proposed.Simulation and analysis demonstrate that the following and avoidance behaviors are important factors of lane formation.A high tendency of following results in good lane formation.Both non-selective following behavior and aggressive overtaking behavior cause the system order to decrease.The most orderly following strategy for a pedestrian is to overtake the former pedestrian whose speed is lower than approximately 70%of his own.The influence of the obstacle layout on pedestrian lane and egress efficiency is also studied with this model.The presence of a small obstacle does not obstruct the walking of pedestrians;in contrast,it may help to improve the egress efficiency by guiding the pedestrian flow and mitigating the reduction of pedestrian system orderliness.展开更多
Due to multi-scale variations and occlusion problems,accurate traffic road pedestrian detection faces great challenges.This paper proposes an improved pedestrian detection method called Multi Scales Attention-YOLOv5x(...Due to multi-scale variations and occlusion problems,accurate traffic road pedestrian detection faces great challenges.This paper proposes an improved pedestrian detection method called Multi Scales Attention-YOLOv5x(MSA-YOLOv5x)based on the YOLOv5x framework.Firstly,by replacing the first convolutional operation of the backbone network with the Focus module,this method expands the number of image input channels to enhance feature expressiveness.Secondly,we construct C3_CBAM module instead of the original C3 module for better feature fusion.In this way,the learning process could achieve more multi-scale features and occluded pedestrian target features through channel attention and spatial attention.Additionally,a new feature pyramid detection layer and a new detection channel are embedded in the feature fusion part for enhancing multi-scale pedestrian detection accuracy.Compared with the baseline methods,experimental results on a public dataset demonstrate that the proposed method achieves optimal detection accuracy for traffic road pedestrian detection.展开更多
This study investigates pedestrian safety perception in Ho Chi Minh City under mixed traffic conditions by evaluating comfort,crash risk,and injury risk perceptions in two scenarios:walking along and crossing multilan...This study investigates pedestrian safety perception in Ho Chi Minh City under mixed traffic conditions by evaluating comfort,crash risk,and injury risk perceptions in two scenarios:walking along and crossing multilane roads.Using visual experiments with 510 participants,the study identifies how sidewalk quality,obstructions,crossing infrastructure,and traffic conditions shape pedestrian experiences.Statistical modeling reveals that protected sidewalks and comprehensive crossing features significantly enhance perceived safety and comfort.Findings emphasize the need for improved pedestrian infrastructure and traffic calming in dense urban settings to support safer,more inclusive mobility under mixed traffic conditions like Vietnam.展开更多
Pedestrian detection has been a hot spot in computer vision over the past decades due to the wide spectrum of promising applications,and the major challenge is false positives that occur during pedestrian detection.Th...Pedestrian detection has been a hot spot in computer vision over the past decades due to the wide spectrum of promising applications,and the major challenge is false positives that occur during pedestrian detection.The emergence of various Convolutional Neural Network-based detection strategies substantially enhances pedestrian detection accuracy but still does not solve this problem well.This paper deeply analyzes the detection framework of the two-stage CNN detection methods and finds out false positives in detection results are due to its training strategy misclassifying some false proposals,thus weakening the classification capability of the following subnetwork and hardly suppressing false ones.To solve this problem,this paper proposes a pedestrian-sensitive training algorithm to help two-stage CNN detection methods effectively learn to distinguish the pedestrian and non-pedestrian samples and suppress the false positives in the final detection results.The core of the proposed algorithm is to redesign the training proposal generating scheme for the two-stage CNN detection methods,which can avoid a certain number of false ones that mislead its training process.With the help of the proposed algorithm,the detection accuracy of the MetroNext,a smaller and more accurate metro passenger detector,is further improved,which further decreases false ones in its metro passenger detection results.Based on various challenging benchmark datasets,experiment results have demonstrated that the feasibility of the proposed algorithm is effective in improving pedestrian detection accuracy by removing false positives.Compared with the existing state-of-the-art detection networks,PSTNet demonstrates better overall prediction performance in accuracy,total number of parameters,and inference time;thus,it can become a practical solution for hunting pedestrians on various hardware platforms,especially for mobile and edge devices.展开更多
This study presents an interpretable surrogate framework for predicting pedestrian-leg injury severity that integrates high-fidelity finite-element(FE)simulations with a TabNet-based deep-learning model.We generated a...This study presents an interpretable surrogate framework for predicting pedestrian-leg injury severity that integrates high-fidelity finite-element(FE)simulations with a TabNet-based deep-learning model.We generated a parametric dataset of 3000 impact scenarios-covering ten vehicle types and various legform impactors-using automated FE runs configured via Latin hypercube sampling.After preprocessing and one-hot encoding of categorical features,we trained TabNet alongside Support-Vector Regression,Random Forest,and Decision-Tree ensembles.All models underwent hyperparameter tuning via Optuna’s Bayesian optimization coupled with repeated four-fold crossvalidation(20 trials per model).TabNet achieved the best balance of explanatory power and predictive accuracy,with an average R^(2)=0.94±0.01 and RMSE=0.14±0.02.On an independent test set,85%,88%,and 90%of predictions for tibial acceleration,knee-flexion angle,and shear displacement,respectively,fell within±20%of true peaks.SHAPbased analyses confirm that collision-point location and bumper geometry dominate injury outcomes.These results demonstrate TabNet’s capacity to deliver rapid,robust,and explainable injury predictions,offering actionable design insights for vehicle front-end optimization and regulatory assessment in early development stages.展开更多
Visual indoor positioning methods have the potential for widespread application in complex large-scale indoor environments,such as shopping centers and hospitals.However,during the visual positioning process,passing p...Visual indoor positioning methods have the potential for widespread application in complex large-scale indoor environments,such as shopping centers and hospitals.However,during the visual positioning process,passing pedestrians may cause occlusion in the visual image,leading to large deviations in the visual positioning results.Aiming at the problem of feature occlusion in visual images caused by pedestrians,this paper proposes a visual indoor positioning system that combines semantic segmentation and image restoration.The paper proposes a method called the fast image segmentation repair(FISR),which segments and rapidly repairs the selected image to eliminate the influence of pedestrians on image feature extraction and improve positioning accuracy.In addition,the paper proposes a method called local feature based bag-of-visual-words combined with high-level semantic information(LFHS)for image retrieval.LFHS uses both local features and high-level semantic information to obtain more comprehensive and accurate representations of image features.This approach improves the accuracy and robustness of image retrieval by harnessing the combined power of local features and high-level semantic information.Experimental results show that the proposed positioning method reduces the average positioning error by 0.35 m compared with NetVLAD and 0.49 m compared with MixVPR,significantly improving the performance of visual positioning technology.展开更多
With the development of micro-electromechanical systems(MEMS), miniaturized, low-power and low-cost inertial measurement units(IMUs) have been widely integrated into mobile terminals and smart wearable devices. This p...With the development of micro-electromechanical systems(MEMS), miniaturized, low-power and low-cost inertial measurement units(IMUs) have been widely integrated into mobile terminals and smart wearable devices. This provides the prospect of a broad application for the inertial sensor-based pedestrian dead-reckoning(IPDR) systems. Especially for indoor navigation and indoor positioning, the IPDR systems have many unique advantages that other methods do not have. At present, a large number of technologies and methods for IPDR systems are proposed. In this paper, we have analyzed and outlined the IPDR systems based on about 80 documents in the field of IPDR in recent years. The article is structured in the form of an introduction-elucidation-conclusion framework. First, we proposed a general framework to explore the structure of an IPDR system. Then, according to this framework, the IPDR system was divided into six relatively independent sub-problems, which were discussed and summarized separately. Finally, we proposed a graph structure of IPDR systems, and a sub-directed graph, formed by selecting a combined path from the start node to the end node, skillfully constitutes a technical route of one specific IPDR system. At the end of the article, we summarized some key issues that need to be resolved before the IPDR systems are widely used.展开更多
Lower limb injures are frequently observed in passenger car traffic accidents.Previous studies of the injuries focus on long bone fractures by using either cadaver component tests or simulations of the long bone kinem...Lower limb injures are frequently observed in passenger car traffic accidents.Previous studies of the injuries focus on long bone fractures by using either cadaver component tests or simulations of the long bone kinematics,which lack in-depth study on the fractures in stress analysis.This paper aims to investigate lower limb impact biomechanics in real-world car to pedestrian accidents and to predict fractures of long bones in term of stress parameter for femur,tibia,and fibula.For the above purposes,a 3D finite element(FE) model of human body lower limb(HBM-LL) is developed based on human anatomy.The model consists of the pelvis,femur,tibia,fibula,patella,foot bones,primary tendons,knee joint capsule,meniscus,and ligaments.The FE model is validated by comparing the results from a lateral impact between simulations and tests with cadaver lower limb specimens.Two real-world accidents are selected from an in-depth accident database with detailed information about the accident scene,car impact speed,damage to the car,and pedestrian injuries.Multi-body system(MBS) models are used to reconstruct the kinematics of the pedestrians in the two accidents and the impact conditions are calculated for initial impact velocity and orientations of the car and pedestrian during the collision.The FE model is used to perform injury reconstructions and predict the fractures by using physical parameters,such as von Mises stress of long bones.The calculated failure level of the long bones is correlated with the injury outcomes observed from the two accident cases.The reconstruction result shows that the HBM-LL FE model has acceptable biofidelity and can be applied to predict the risk of long bone fractures.This study provides an efficient methodology to investigate the long bone fracture suffered from vehicle traffic collisions.展开更多
Pedestrian detection and tracking are vital elements of today’s surveillance systems,which make daily life safe for humans.Thus,human detection and visualization have become essential inventions in the field of compu...Pedestrian detection and tracking are vital elements of today’s surveillance systems,which make daily life safe for humans.Thus,human detection and visualization have become essential inventions in the field of computer vision.Hence,developing a surveillance system with multiple object recognition and tracking,especially in low light and night-time,is still challenging.Therefore,we propose a novel system based on machine learning and image processing to provide an efficient surveillance system for pedestrian detection and tracking at night.In particular,we propose a system that tackles a two-fold problem by detecting multiple pedestrians in infrared(IR)images using machine learning and tracking them using particle filters.Moreover,a random forest classifier is adopted for image segmentation to identify pedestrians in an image.The result of detection is investigated by particle filter to solve pedestrian tracking.Through the extensive experiment,our system shows 93%segmentation accuracy using a random forest algorithm that demonstrates high accuracy for background and roof classes.Moreover,the system achieved a detection accuracy of 90%usingmultiple templatematching techniques and 81%accuracy for pedestrian tracking.Furthermore,our system can identify that the detected object is a human.Hence,our system provided the best results compared to the state-ofart systems,which proves the effectiveness of the techniques used for image segmentation,classification,and tracking.The presented method is applicable for human detection/tracking,crowd analysis,and monitoring pedestrians in IR video surveillance.展开更多
Pedestrian protection has played an important role for driver assistance systems.Our aim is to develop a video based driver assistance system for the detection of the potentially dangerous situation between the vehicl...Pedestrian protection has played an important role for driver assistance systems.Our aim is to develop a video based driver assistance system for the detection of the potentially dangerous situation between the vehicle and pedestrian,in order to warn the driver.In this paper,we address the problem of detecting pedestrian in real-world scenes and estimation of the walking direction with a single camera from a moving vehicle.Considering all the available cues for predicting the possibility of collision is very important.The direction in which the pedestrian is facing is one of the most important cues predicting where the pedestrian may move in the future.So we first address the problem of sin-gle-frame pedestrian orientation estimation in real-world scenes.Then again,we estimate the pedes-trian walking direction using multi-frame based on the result of single-frame orientation estimation.We propose a three-step method:pedestrian detection for single-frame step,orientation estimation for single-frame step and walking direction estimation for multi-frame step.To evaluate the proposed method in its robustness and accuracy,the experiments have been performed between numbers of images which is highly challenging uncontrolled conditions in real world.It shows a significant per-formance improvement in octant orientation estimation of about 64% accuracy in the orientation es-timation step and achieved surprisingly good accuracy in estimating the walking direction against 212 targeted objects.展开更多
A real-time pedestrian detection and tracking system using a single video camera was developed to monitor pedestrians. This system contained six modules: video flow capture, pre-processing, movement detection, shadow ...A real-time pedestrian detection and tracking system using a single video camera was developed to monitor pedestrians. This system contained six modules: video flow capture, pre-processing, movement detection, shadow removal, tracking, and object classification. The Gaussian mixture model was utilized to extract the moving object from an image sequence segmented by the mean-shift technique in the pre-processing module. Shadow removal was used to alleviate the negative impact of the shadow to the detected objects. A model-free method was adopted to identify pedestrians. The maximum and minimum integration methods were developed to integrate multiple cues into the mean-shift algorithm and the initial tracking iteration with the competent integrated probability distribution map for object tracking. A simple but effective algorithm was proposed to handle full occlusion cases. The system was tested using real traffic videos from different sites. The results of the test confirm that the system is reliable and has an overall accuracy of over 85%.展开更多
Road traffic safety can decrease when drivers drive in a low-visibility environment.The application of visual perception technology to detect vehicles and pedestrians in infrared images proves to be an effective means...Road traffic safety can decrease when drivers drive in a low-visibility environment.The application of visual perception technology to detect vehicles and pedestrians in infrared images proves to be an effective means of reducing the risk of accidents.To tackle the challenges posed by the low recognition accuracy and the substan-tial computational burden associated with current infrared pedestrian-vehicle detection methods,an infrared pedestrian-vehicle detection method A proposal is presented,based on an enhanced version of You Only Look Once version 5(YOLOv5).First,A head specifically designed for detecting small targets has been integrated into the model to make full use of shallow feature information to enhance the accuracy in detecting small targets.Second,the Focal Generalized Intersection over Union(GIoU)is employed as an alternative to the original loss function to address issues related to target overlap and category imbalance.Third,the distribution shift convolution optimization feature extraction operator is used to alleviate the computational burden of the model without significantly compromising detection accuracy.The test results of the improved algorithm show that its average accuracy(mAP)reaches 90.1%.Specifically,the Giga Floating Point Operations Per second(GFLOPs)of the improved algorithm is only 9.1.In contrast,the improved algorithms outperformed the other algorithms on similar GFLOPs,such as YOLOv6n(11.9),YOLOv8n(8.7),YOLOv7t(13.2)and YOLOv5s(16.0).The mAPs that are 4.4%,3%,3.5%,and 1.7%greater than those of these algorithms show that the improved algorithm achieves higher accuracy in target detection tasks under similar computational resource overhead.On the other hand,compared with other algorithms such as YOLOv8l(91.1%),YOLOv6l(89.5%),YOLOv7(90.8%),and YOLOv3(90.1%),the improved algorithm needs only 5.5%,2.3%,8.6%,and 2.3%,respectively,of the GFLOPs.The improved algorithm has shown significant advancements in balancing accuracy and computational efficiency,making it promising for practical use in resource-limited scenarios.展开更多
The COVID-19 virus is usually spread by small droplets when talking,coughing and sneezing,so maintaining physical distance between people is necessary to slow the spread of the virus.The World Health Organization(WHO)...The COVID-19 virus is usually spread by small droplets when talking,coughing and sneezing,so maintaining physical distance between people is necessary to slow the spread of the virus.The World Health Organization(WHO)recommends maintaining a social distance of at least six feet.In this paper,we developed a real-time pedestrian social distance risk alert system for COVID-19,whichmonitors the distance between people in real-time via video streaming and provides risk alerts to the person in charge,thus avoiding the problem of too close social distance between pedestrians in public places.We design a lightweight convolutional neural network architecture to detect the distance between people more accurately.In addition,due to the limitation of camera placement,the previous algorithm based on flat view is not applicable to the social distance calculation for cameras,so we designed and developed a perspective conversion module to reduce the image in the video to a bird’s eye view,which can avoid the error caused by the elevation view and thus provide accurate risk indication to the user.We selected images containing only person labels in theCOCO2017 dataset to train our networkmodel.The experimental results show that our network model achieves 82.3%detection accuracy and performs significantly better than other mainstream network architectures in the three metrics of Recall,Precision and mAP,proving the effectiveness of our system and the efficiency of our technology.展开更多
The braking behavior of drivers when a pedestrian comes out from the sidewalk to the road was analyzed using a driving simulator. Based on drivers＇ braking behavior, the braking control timing of the system for avoid...The braking behavior of drivers when a pedestrian comes out from the sidewalk to the road was analyzed using a driving simulator. Based on drivers＇ braking behavior, the braking control timing of the system for avoiding the collision with pedestrians was proposed. In this study, the subject drivers started braking at almost the same time in terms of TTC （Time to Collision）, regardless of the velocity of a subject vehicle and crossing velocity of pedestrians. This experimental result showed that brake timing of the system which can minimize the interference for braking between drivers and the system is 1.3 s of TTC. Next, the drivers＇ braking behavior was investigated when the system controlled braking to avoid collision at this timing. As a result, drivers did not show any change of braking behavior with no excessive interference between braking control by the system and braking operation by drivers for avoiding collisions with pedestrians which is equivalent to the excessive dependence on the system.展开更多
Urban intersections without traffic signals are prone to accidents involving motor vehicles and pedestrians.Utilizing computer vision technology to detect pedestrians crossing the street can effectively mitigate the o...Urban intersections without traffic signals are prone to accidents involving motor vehicles and pedestrians.Utilizing computer vision technology to detect pedestrians crossing the street can effectively mitigate the occurrence of such accidents.Faced with the complex issue of pedestrian occlusion at signal-free intersections,this paper proposes a target detection model called Head feature And ENMS fusion Residual connection For CNN(HAERC).Specifically,the model includes a head feature module that detects occluded pedestrians by integrating their head features with the overall target.Additionally,to address the misselection caused by overlapping candidate boxes in two-stage target detection models,an Extended Non-Maximum Suppression classifier(ENMS)with expanded IoU thresholds is proposed.Finally,leveraging the CityPersons dataset and categorizing it into four classes based on occlusion levels(heavy,reasonable,partial,bare),the HAERC model is experimented on these classes and compared with baseline models.Experimental results demonstrate that HAERC achieves superior False Positives Per Image(FPPI)values of 46.64%,9.59%,9.43%,and 6.78%respectively for the four classes,outperforming all baseline models.The study concludes that the HAERC model effectively identifies occluded pedestrians in the complex environment of urban intersections without traffic signals,thereby enhancing safety for long-range driving at such intersections.展开更多
With the development of positioning technology,loca-tion services are constantly in demand by people.As a primary location service pedestrian navigation has two main approaches based on radio and inertial navigation.T...With the development of positioning technology,loca-tion services are constantly in demand by people.As a primary location service pedestrian navigation has two main approaches based on radio and inertial navigation.The pedestrian naviga-tion based on radio is subject to environmental occlusion lead-ing to the degradation of positioning accuracy.The pedestrian navigation based on micro-electro-mechanical system inertial measurement unit(MIMU)is less susceptible to environmental interference,but its errors dissipate over time.In this paper,a chest card pedestrian navigation improvement method based on complementary correction is proposed in order to suppress the error divergence of inertial navigation methods.To suppress atti-tude errors,optimal feedback coefficients are established by pedestrian motion characteristics.To extend navigation time and improve positioning accuracy,the step length in subsequent movements is compensated by the first step length.The experi-mental results show that the positioning accuracy of the pro-posed method is improved by more than 47%and 44%com-pared with the pure inertia-based method combined with step compensation and the traditional complementary filtering com-bined method with step compensation.The proposed method can effectively suppress the error dispersion and improve the positioning accuracy.展开更多
Multispectral pedestrian detection technology leverages infrared images to provide reliable information for visible light images, demonstrating significant advantages in low-light conditions and background occlusion s...Multispectral pedestrian detection technology leverages infrared images to provide reliable information for visible light images, demonstrating significant advantages in low-light conditions and background occlusion scenarios. However, while continuously improving cross-modal feature extraction and fusion, ensuring the model’s detection speed is also a challenging issue. We have devised a deep learning network model for cross-modal pedestrian detection based on Resnet50, aiming to focus on more reliable features and enhance the model’s detection efficiency. This model employs a spatial attention mechanism to reweight the input visible light and infrared image data, enhancing the model’s focus on different spatial positions and sharing the weighted feature data across different modalities, thereby reducing the interference of multi-modal features. Subsequently, lightweight modules with depthwise separable convolution are incorporated to reduce the model’s parameter count and computational load through channel-wise and point-wise convolutions. The network model algorithm proposed in this paper was experimentally validated on the publicly available KAIST dataset and compared with other existing methods. The experimental results demonstrate that our approach achieves favorable performance in various complex environments, affirming the effectiveness of the multispectral pedestrian detection technology proposed in this paper.展开更多
文摘In the context of banning gated communities, blocks returning to the human-oriented scale become the new normal, and pedestrian system design will be paid more attention in the urban planning field. Oct-Loft Creative Park is a template for open blocks in Shenzhen, with a convenient and humanized pedestrian system. This paper selects the creative park's pedestrian system as the research object, using the environment-behavior theory for analysis. Finally, optimization strategies of pedestrian system will be put forward.
基金Sponsored by the State Key Laboratory of Subtropical Building Science(Grant No.2011ZA01)
文摘The Multi-dimensional Pedestrian System( MPS) is an integral part of the new campus of University of Macao. It observes the principle of "pedestrian first " and features a pedestrian-vehicle dividing system.Based on a comprehensive analysis of the location,climate,ecology and other factors of the project site,the conception of the idea of MPS and the related researches are illustrated. The transportation features of the MPS,as summarized,include multi-dimensions,short-distance and weather-resistance. Its features for the sake of livability include integration of nature, respect for the environment and sharing of landscape. Upon the completion of the project, the effects on its users were tested. Finally, some constructive rules for the construction of similar campus pedestrian systems were proposed.
基金supported by the National Natural Science Foundation of China under(Grant No.52175531)in part by the Science and Technology Research Program of Chongqing Municipal Education Commission under Grant(Grant Nos.KJQN202000605 and KJZD-M202000602)。
文摘Pedestrian positioning system(PPS)using wearable inertial sensors has wide applications towards various emerging fields such as smart healthcare,emergency rescue,soldier positioning,etc.The performance of traditional PPS is limited by the cumulative error of inertial sensors,complex motion modes of pedestrians,and the low robustness of the multi-sensor collaboration structure.This paper presents a hybrid pedestrian positioning system using the combination of wearable inertial sensors and ultrasonic ranging(H-PPS).A robust two nodes integration structure is developed to adaptively combine the motion data acquired from the single waist-mounted and foot-mounted node,and enhanced by a novel ellipsoid constraint model.In addition,a deep-learning-based walking speed estimator is proposed by considering all the motion features provided by different nodes,which effectively reduces the cumulative error originating from inertial sensors.Finally,a comprehensive data and model dual-driven model is presented to effectively combine the motion data provided by different sensor nodes and walking speed estimator,and multi-level constraints are extracted to further improve the performance of the overall system.Experimental results indicate that the proposed H-PPS significantly improves the performance of the single PPS and outperforms existing algorithms in accuracy index under complex indoor scenarios.
基金Project supported by the National Natural Science Foundation of China(Grant No.71603146).
文摘Pedestrian self-organizing movement plays a significant role in evacuation studies and architectural design.Lane formation,a typical self-organizing phenomenon,helps pedestrian system to become more orderly,the majority of following behavior model and overtaking behavior model are imprecise and unrealistic compared with pedestrian movement in the real world.In this study,a pedestrian dynamic model considering detailed modelling of the following behavior and overtaking behavior is constructed,and a method of measuring the lane formation and pedestrian system order based on information entropy is proposed.Simulation and analysis demonstrate that the following and avoidance behaviors are important factors of lane formation.A high tendency of following results in good lane formation.Both non-selective following behavior and aggressive overtaking behavior cause the system order to decrease.The most orderly following strategy for a pedestrian is to overtake the former pedestrian whose speed is lower than approximately 70%of his own.The influence of the obstacle layout on pedestrian lane and egress efficiency is also studied with this model.The presence of a small obstacle does not obstruct the walking of pedestrians;in contrast,it may help to improve the egress efficiency by guiding the pedestrian flow and mitigating the reduction of pedestrian system orderliness.
文摘Due to multi-scale variations and occlusion problems,accurate traffic road pedestrian detection faces great challenges.This paper proposes an improved pedestrian detection method called Multi Scales Attention-YOLOv5x(MSA-YOLOv5x)based on the YOLOv5x framework.Firstly,by replacing the first convolutional operation of the backbone network with the Focus module,this method expands the number of image input channels to enhance feature expressiveness.Secondly,we construct C3_CBAM module instead of the original C3 module for better feature fusion.In this way,the learning process could achieve more multi-scale features and occluded pedestrian target features through channel attention and spatial attention.Additionally,a new feature pyramid detection layer and a new detection channel are embedded in the feature fusion part for enhancing multi-scale pedestrian detection accuracy.Compared with the baseline methods,experimental results on a public dataset demonstrate that the proposed method achieves optimal detection accuracy for traffic road pedestrian detection.
文摘This study investigates pedestrian safety perception in Ho Chi Minh City under mixed traffic conditions by evaluating comfort,crash risk,and injury risk perceptions in two scenarios:walking along and crossing multilane roads.Using visual experiments with 510 participants,the study identifies how sidewalk quality,obstructions,crossing infrastructure,and traffic conditions shape pedestrian experiences.Statistical modeling reveals that protected sidewalks and comprehensive crossing features significantly enhance perceived safety and comfort.Findings emphasize the need for improved pedestrian infrastructure and traffic calming in dense urban settings to support safer,more inclusive mobility under mixed traffic conditions like Vietnam.
文摘Pedestrian detection has been a hot spot in computer vision over the past decades due to the wide spectrum of promising applications,and the major challenge is false positives that occur during pedestrian detection.The emergence of various Convolutional Neural Network-based detection strategies substantially enhances pedestrian detection accuracy but still does not solve this problem well.This paper deeply analyzes the detection framework of the two-stage CNN detection methods and finds out false positives in detection results are due to its training strategy misclassifying some false proposals,thus weakening the classification capability of the following subnetwork and hardly suppressing false ones.To solve this problem,this paper proposes a pedestrian-sensitive training algorithm to help two-stage CNN detection methods effectively learn to distinguish the pedestrian and non-pedestrian samples and suppress the false positives in the final detection results.The core of the proposed algorithm is to redesign the training proposal generating scheme for the two-stage CNN detection methods,which can avoid a certain number of false ones that mislead its training process.With the help of the proposed algorithm,the detection accuracy of the MetroNext,a smaller and more accurate metro passenger detector,is further improved,which further decreases false ones in its metro passenger detection results.Based on various challenging benchmark datasets,experiment results have demonstrated that the feasibility of the proposed algorithm is effective in improving pedestrian detection accuracy by removing false positives.Compared with the existing state-of-the-art detection networks,PSTNet demonstrates better overall prediction performance in accuracy,total number of parameters,and inference time;thus,it can become a practical solution for hunting pedestrians on various hardware platforms,especially for mobile and edge devices.
基金sponsored by the National Natural Science Foundation of China(No.U21A20165,No.52072057).
文摘This study presents an interpretable surrogate framework for predicting pedestrian-leg injury severity that integrates high-fidelity finite-element(FE)simulations with a TabNet-based deep-learning model.We generated a parametric dataset of 3000 impact scenarios-covering ten vehicle types and various legform impactors-using automated FE runs configured via Latin hypercube sampling.After preprocessing and one-hot encoding of categorical features,we trained TabNet alongside Support-Vector Regression,Random Forest,and Decision-Tree ensembles.All models underwent hyperparameter tuning via Optuna’s Bayesian optimization coupled with repeated four-fold crossvalidation(20 trials per model).TabNet achieved the best balance of explanatory power and predictive accuracy,with an average R^(2)=0.94±0.01 and RMSE=0.14±0.02.On an independent test set,85%,88%,and 90%of predictions for tibial acceleration,knee-flexion angle,and shear displacement,respectively,fell within±20%of true peaks.SHAPbased analyses confirm that collision-point location and bumper geometry dominate injury outcomes.These results demonstrate TabNet’s capacity to deliver rapid,robust,and explainable injury predictions,offering actionable design insights for vehicle front-end optimization and regulatory assessment in early development stages.
基金Supported by the National Natural Science Foundation of China(No.61971162,61771186)the Natural Science Foundation of Heilongjiang Province(No.PL2024F025)+2 种基金the Open Research Fund of National Mobile Communications Research Laboratory Southeast University(No.2023D07)the Outstanding Youth Program of Natural Science Foundation of Heilongjiang Province(No.YQ2020F012)the Fundamental Scientific Research Funds of Heilongjiang Province(No.2022-KYYWF-1050).
文摘Visual indoor positioning methods have the potential for widespread application in complex large-scale indoor environments,such as shopping centers and hospitals.However,during the visual positioning process,passing pedestrians may cause occlusion in the visual image,leading to large deviations in the visual positioning results.Aiming at the problem of feature occlusion in visual images caused by pedestrians,this paper proposes a visual indoor positioning system that combines semantic segmentation and image restoration.The paper proposes a method called the fast image segmentation repair(FISR),which segments and rapidly repairs the selected image to eliminate the influence of pedestrians on image feature extraction and improve positioning accuracy.In addition,the paper proposes a method called local feature based bag-of-visual-words combined with high-level semantic information(LFHS)for image retrieval.LFHS uses both local features and high-level semantic information to obtain more comprehensive and accurate representations of image features.This approach improves the accuracy and robustness of image retrieval by harnessing the combined power of local features and high-level semantic information.Experimental results show that the proposed positioning method reduces the average positioning error by 0.35 m compared with NetVLAD and 0.49 m compared with MixVPR,significantly improving the performance of visual positioning technology.
基金supported by National Key Research and Development of China (No. 2017YFB1002800)
文摘With the development of micro-electromechanical systems(MEMS), miniaturized, low-power and low-cost inertial measurement units(IMUs) have been widely integrated into mobile terminals and smart wearable devices. This provides the prospect of a broad application for the inertial sensor-based pedestrian dead-reckoning(IPDR) systems. Especially for indoor navigation and indoor positioning, the IPDR systems have many unique advantages that other methods do not have. At present, a large number of technologies and methods for IPDR systems are proposed. In this paper, we have analyzed and outlined the IPDR systems based on about 80 documents in the field of IPDR in recent years. The article is structured in the form of an introduction-elucidation-conclusion framework. First, we proposed a general framework to explore the structure of an IPDR system. Then, according to this framework, the IPDR system was divided into six relatively independent sub-problems, which were discussed and summarized separately. Finally, we proposed a graph structure of IPDR systems, and a sub-directed graph, formed by selecting a combined path from the start node to the end node, skillfully constitutes a technical route of one specific IPDR system. At the end of the article, we summarized some key issues that need to be resolved before the IPDR systems are widely used.
基金supported by National Hi-tech Research and Development Program of China (863 Program,Grant No. 2006AA110101)"111 Program" of Ministry of Education and State Administration of Foreign Experts Affairs of China (Grant No. 111-2-11)+1 种基金General Motors Research and Development Center (Grant No. RD-209)Project of State Key Laboratory of Advanced Design and Manufacturing for Vehicle Body,Hunan University,China (Grant No. 60870004)
文摘Lower limb injures are frequently observed in passenger car traffic accidents.Previous studies of the injuries focus on long bone fractures by using either cadaver component tests or simulations of the long bone kinematics,which lack in-depth study on the fractures in stress analysis.This paper aims to investigate lower limb impact biomechanics in real-world car to pedestrian accidents and to predict fractures of long bones in term of stress parameter for femur,tibia,and fibula.For the above purposes,a 3D finite element(FE) model of human body lower limb(HBM-LL) is developed based on human anatomy.The model consists of the pelvis,femur,tibia,fibula,patella,foot bones,primary tendons,knee joint capsule,meniscus,and ligaments.The FE model is validated by comparing the results from a lateral impact between simulations and tests with cadaver lower limb specimens.Two real-world accidents are selected from an in-depth accident database with detailed information about the accident scene,car impact speed,damage to the car,and pedestrian injuries.Multi-body system(MBS) models are used to reconstruct the kinematics of the pedestrians in the two accidents and the impact conditions are calculated for initial impact velocity and orientations of the car and pedestrian during the collision.The FE model is used to perform injury reconstructions and predict the fractures by using physical parameters,such as von Mises stress of long bones.The calculated failure level of the long bones is correlated with the injury outcomes observed from the two accident cases.The reconstruction result shows that the HBM-LL FE model has acceptable biofidelity and can be applied to predict the risk of long bone fractures.This study provides an efficient methodology to investigate the long bone fracture suffered from vehicle traffic collisions.
基金supported by the MSIT(Ministry of Science and ICT),Korea,under the ITRC(Information Technology Research Center)support program(IITP-2023-2018-0-01426)supervised by the IITP(Institute for Information&Communications Technology Planning&Evaluation)+2 种基金Also,this work was partially supported by the Taif University Researchers Supporting Project Number(TURSP-2020/115)Taif University,Taif,Saudi Arabia.This work was also supported by Princess Nourah bint Abdulrahman University Researchers Supporting Project number(PNURSP2023R239)PrincessNourah bint Abdulrahman University,Riyadh,Saudi Arabia.
文摘Pedestrian detection and tracking are vital elements of today’s surveillance systems,which make daily life safe for humans.Thus,human detection and visualization have become essential inventions in the field of computer vision.Hence,developing a surveillance system with multiple object recognition and tracking,especially in low light and night-time,is still challenging.Therefore,we propose a novel system based on machine learning and image processing to provide an efficient surveillance system for pedestrian detection and tracking at night.In particular,we propose a system that tackles a two-fold problem by detecting multiple pedestrians in infrared(IR)images using machine learning and tracking them using particle filters.Moreover,a random forest classifier is adopted for image segmentation to identify pedestrians in an image.The result of detection is investigated by particle filter to solve pedestrian tracking.Through the extensive experiment,our system shows 93%segmentation accuracy using a random forest algorithm that demonstrates high accuracy for background and roof classes.Moreover,the system achieved a detection accuracy of 90%usingmultiple templatematching techniques and 81%accuracy for pedestrian tracking.Furthermore,our system can identify that the detected object is a human.Hence,our system provided the best results compared to the state-ofart systems,which proves the effectiveness of the techniques used for image segmentation,classification,and tracking.The presented method is applicable for human detection/tracking,crowd analysis,and monitoring pedestrians in IR video surveillance.
文摘Pedestrian protection has played an important role for driver assistance systems.Our aim is to develop a video based driver assistance system for the detection of the potentially dangerous situation between the vehicle and pedestrian,in order to warn the driver.In this paper,we address the problem of detecting pedestrian in real-world scenes and estimation of the walking direction with a single camera from a moving vehicle.Considering all the available cues for predicting the possibility of collision is very important.The direction in which the pedestrian is facing is one of the most important cues predicting where the pedestrian may move in the future.So we first address the problem of sin-gle-frame pedestrian orientation estimation in real-world scenes.Then again,we estimate the pedes-trian walking direction using multi-frame based on the result of single-frame orientation estimation.We propose a three-step method:pedestrian detection for single-frame step,orientation estimation for single-frame step and walking direction estimation for multi-frame step.To evaluate the proposed method in its robustness and accuracy,the experiments have been performed between numbers of images which is highly challenging uncontrolled conditions in real world.It shows a significant per-formance improvement in octant orientation estimation of about 64% accuracy in the orientation es-timation step and achieved surprisingly good accuracy in estimating the walking direction against 212 targeted objects.
基金Project(50778015)supported by the National Natural Science Foundation of ChinaProject(2012CB725403)supported by the Major State Basic Research Development Program of China
文摘A real-time pedestrian detection and tracking system using a single video camera was developed to monitor pedestrians. This system contained six modules: video flow capture, pre-processing, movement detection, shadow removal, tracking, and object classification. The Gaussian mixture model was utilized to extract the moving object from an image sequence segmented by the mean-shift technique in the pre-processing module. Shadow removal was used to alleviate the negative impact of the shadow to the detected objects. A model-free method was adopted to identify pedestrians. The maximum and minimum integration methods were developed to integrate multiple cues into the mean-shift algorithm and the initial tracking iteration with the competent integrated probability distribution map for object tracking. A simple but effective algorithm was proposed to handle full occlusion cases. The system was tested using real traffic videos from different sites. The results of the test confirm that the system is reliable and has an overall accuracy of over 85%.
文摘Road traffic safety can decrease when drivers drive in a low-visibility environment.The application of visual perception technology to detect vehicles and pedestrians in infrared images proves to be an effective means of reducing the risk of accidents.To tackle the challenges posed by the low recognition accuracy and the substan-tial computational burden associated with current infrared pedestrian-vehicle detection methods,an infrared pedestrian-vehicle detection method A proposal is presented,based on an enhanced version of You Only Look Once version 5(YOLOv5).First,A head specifically designed for detecting small targets has been integrated into the model to make full use of shallow feature information to enhance the accuracy in detecting small targets.Second,the Focal Generalized Intersection over Union(GIoU)is employed as an alternative to the original loss function to address issues related to target overlap and category imbalance.Third,the distribution shift convolution optimization feature extraction operator is used to alleviate the computational burden of the model without significantly compromising detection accuracy.The test results of the improved algorithm show that its average accuracy(mAP)reaches 90.1%.Specifically,the Giga Floating Point Operations Per second(GFLOPs)of the improved algorithm is only 9.1.In contrast,the improved algorithms outperformed the other algorithms on similar GFLOPs,such as YOLOv6n(11.9),YOLOv8n(8.7),YOLOv7t(13.2)and YOLOv5s(16.0).The mAPs that are 4.4%,3%,3.5%,and 1.7%greater than those of these algorithms show that the improved algorithm achieves higher accuracy in target detection tasks under similar computational resource overhead.On the other hand,compared with other algorithms such as YOLOv8l(91.1%),YOLOv6l(89.5%),YOLOv7(90.8%),and YOLOv3(90.1%),the improved algorithm needs only 5.5%,2.3%,8.6%,and 2.3%,respectively,of the GFLOPs.The improved algorithm has shown significant advancements in balancing accuracy and computational efficiency,making it promising for practical use in resource-limited scenarios.
基金This research was funded by the Fundamental Research Funds for the Central Universities,3072022TS0605the China University Industry-University-Research Innovation Fund,2021LDA10004.
文摘The COVID-19 virus is usually spread by small droplets when talking,coughing and sneezing,so maintaining physical distance between people is necessary to slow the spread of the virus.The World Health Organization(WHO)recommends maintaining a social distance of at least six feet.In this paper,we developed a real-time pedestrian social distance risk alert system for COVID-19,whichmonitors the distance between people in real-time via video streaming and provides risk alerts to the person in charge,thus avoiding the problem of too close social distance between pedestrians in public places.We design a lightweight convolutional neural network architecture to detect the distance between people more accurately.In addition,due to the limitation of camera placement,the previous algorithm based on flat view is not applicable to the social distance calculation for cameras,so we designed and developed a perspective conversion module to reduce the image in the video to a bird’s eye view,which can avoid the error caused by the elevation view and thus provide accurate risk indication to the user.We selected images containing only person labels in theCOCO2017 dataset to train our networkmodel.The experimental results show that our network model achieves 82.3%detection accuracy and performs significantly better than other mainstream network architectures in the three metrics of Recall,Precision and mAP,proving the effectiveness of our system and the efficiency of our technology.
文摘The braking behavior of drivers when a pedestrian comes out from the sidewalk to the road was analyzed using a driving simulator. Based on drivers＇ braking behavior, the braking control timing of the system for avoiding the collision with pedestrians was proposed. In this study, the subject drivers started braking at almost the same time in terms of TTC （Time to Collision）, regardless of the velocity of a subject vehicle and crossing velocity of pedestrians. This experimental result showed that brake timing of the system which can minimize the interference for braking between drivers and the system is 1.3 s of TTC. Next, the drivers＇ braking behavior was investigated when the system controlled braking to avoid collision at this timing. As a result, drivers did not show any change of braking behavior with no excessive interference between braking control by the system and braking operation by drivers for avoiding collisions with pedestrians which is equivalent to the excessive dependence on the system.
基金Beijing Natural Science Foundation(9234025)National Social Science Fund Project of China(21FGLB014)Humanity and Social Science Youth Foundation of Ministry of Education of China(21YJC630094).
文摘Urban intersections without traffic signals are prone to accidents involving motor vehicles and pedestrians.Utilizing computer vision technology to detect pedestrians crossing the street can effectively mitigate the occurrence of such accidents.Faced with the complex issue of pedestrian occlusion at signal-free intersections,this paper proposes a target detection model called Head feature And ENMS fusion Residual connection For CNN(HAERC).Specifically,the model includes a head feature module that detects occluded pedestrians by integrating their head features with the overall target.Additionally,to address the misselection caused by overlapping candidate boxes in two-stage target detection models,an Extended Non-Maximum Suppression classifier(ENMS)with expanded IoU thresholds is proposed.Finally,leveraging the CityPersons dataset and categorizing it into four classes based on occlusion levels(heavy,reasonable,partial,bare),the HAERC model is experimented on these classes and compared with baseline models.Experimental results demonstrate that HAERC achieves superior False Positives Per Image(FPPI)values of 46.64%,9.59%,9.43%,and 6.78%respectively for the four classes,outperforming all baseline models.The study concludes that the HAERC model effectively identifies occluded pedestrians in the complex environment of urban intersections without traffic signals,thereby enhancing safety for long-range driving at such intersections.
文摘With the development of positioning technology,loca-tion services are constantly in demand by people.As a primary location service pedestrian navigation has two main approaches based on radio and inertial navigation.The pedestrian naviga-tion based on radio is subject to environmental occlusion lead-ing to the degradation of positioning accuracy.The pedestrian navigation based on micro-electro-mechanical system inertial measurement unit(MIMU)is less susceptible to environmental interference,but its errors dissipate over time.In this paper,a chest card pedestrian navigation improvement method based on complementary correction is proposed in order to suppress the error divergence of inertial navigation methods.To suppress atti-tude errors,optimal feedback coefficients are established by pedestrian motion characteristics.To extend navigation time and improve positioning accuracy,the step length in subsequent movements is compensated by the first step length.The experi-mental results show that the positioning accuracy of the pro-posed method is improved by more than 47%and 44%com-pared with the pure inertia-based method combined with step compensation and the traditional complementary filtering com-bined method with step compensation.The proposed method can effectively suppress the error dispersion and improve the positioning accuracy.
基金supported by the Henan Provincial Science and Technology Research Project under Grants 232102211006,232102210044,232102211017,232102210055 and 222102210214the Science and Technology Innovation Project of Zhengzhou University of Light Industry under Grant 23XNKJTD0205+1 种基金the Undergraduate Universities Smart Teaching Special Research Project of Henan Province under Grant Jiao Gao[2021]No.489-29the Doctor Natural Science Foundation of Zhengzhou University of Light Industry under Grants 2021BSJJ025 and 2022BSJJZK13.
文摘Multispectral pedestrian detection technology leverages infrared images to provide reliable information for visible light images, demonstrating significant advantages in low-light conditions and background occlusion scenarios. However, while continuously improving cross-modal feature extraction and fusion, ensuring the model’s detection speed is also a challenging issue. We have devised a deep learning network model for cross-modal pedestrian detection based on Resnet50, aiming to focus on more reliable features and enhance the model’s detection efficiency. This model employs a spatial attention mechanism to reweight the input visible light and infrared image data, enhancing the model’s focus on different spatial positions and sharing the weighted feature data across different modalities, thereby reducing the interference of multi-modal features. Subsequently, lightweight modules with depthwise separable convolution are incorporated to reduce the model’s parameter count and computational load through channel-wise and point-wise convolutions. The network model algorithm proposed in this paper was experimentally validated on the publicly available KAIST dataset and compared with other existing methods. The experimental results demonstrate that our approach achieves favorable performance in various complex environments, affirming the effectiveness of the multispectral pedestrian detection technology proposed in this paper.
