A pedestrian red light violation warning method and system based on intelligent transportation internet of things

Abstract

This invention relates to the technical field of intelligent transportation IoT, and provides a method and system for pedestrian red-light violation warning based on intelligent transportation IoT. The method includes: acquiring multi-source local sensing data at a traffic intersection; determining the current state of traffic lights based on the multi-source local sensing data to obtain an autonomous judgment result of the traffic light state; acquiring and monitoring external traffic light state information, and when the update of external traffic light state information is delayed or inconsistent with the autonomous judgment result of the traffic light state, prioritizing the autonomous judgment result to obtain a prioritized traffic light state; judging the pedestrian's red-light violation behavior based on the prioritized traffic light state and combined with the received pedestrian movement trajectory prediction of the crosswalk area; and triggering a warning based on the red-light violation behavior. This invention improves the timeliness and accuracy of pedestrian red-light violation warnings.

Description

Technical Field

[0001] This invention relates to the technical field of intelligent transportation Internet of Things (IoT), specifically to a method and system for warning pedestrians who run red lights based on intelligent transportation IoT. Background Technology

[0002] In urban traffic management, intelligent transportation IoT pedestrian red-light violation warning systems typically rely on real-time traffic signal status information transmitted from the city's traffic control center to determine pedestrian red-light violations and trigger warnings. This single and authoritative source of information is the foundation for the normal operation of the warning system.

[0003] To support advanced functions, new controllers need to transmit more data, which is typically transmitted via public wireless network channels. During peak urban traffic hours, these public wireless network channels at critical intersections become extremely congested due to numerous devices vying for resources, leading to data packet delays, loss, and even multiple retries. This communication instability caused by network congestion results in unpredictable and inconsistent delays in the IoT platform's reception of actual traffic light status information. For example, a signal status update that should arrive within tens of milliseconds may be delayed by hundreds of milliseconds or even several seconds.

[0004] This irregular delay in traffic light status information directly affects the internal logic of the warning system in determining "running a red light." Even when the actual traffic light has changed from green to red, due to the information transmission lag, the warning system may still incorrectly perceive it as a "green" or "yellow" light for several seconds afterward. This disconnect between the system's internal understanding of traffic rules and the actual physical world means that the warning system's judgment is outdated when intervention is most needed.

[0005] At busy urban intersections during rush hour, some pedestrians habitually rush across crosswalks the moment the red light turns on. This prevalent behavioral pattern overlaps with the delay in the warning system's assessment of the red light status, creating a critical window of opportunity. By the time the warning system finally receives the delayed red light information and triggers an alert, the rushing pedestrian has already reached the middle of the crosswalk. While this delayed warning still alerts pedestrians, it misses the optimal opportunity to effectively dissuade them before they enter the danger zone, significantly diminishing its effectiveness.

[0006] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0007] This application discloses a pedestrian red-light violation warning method and system based on intelligent transportation Internet of Things, aiming to solve the technical problem that existing intelligent transportation Internet of Things pedestrian red-light violation warning systems cannot timely and accurately judge pedestrian red-light violation behavior and trigger warnings when traffic signal status information transmission is delayed or inconsistent, resulting in a significant reduction in warning effectiveness.

[0008] The technical solution of this application is as follows: In a first aspect, this application discloses a method for warning pedestrians who run red lights based on the Internet of Things for intelligent transportation, comprising the following steps: Acquire multi-source local sensing data at traffic intersections; Based on multi-source local sensing data, the current state of traffic lights is determined, and the autonomous judgment result of traffic light state is obtained. Acquire and monitor external traffic light status information. When the external traffic light status information is delayed or inconsistent with the autonomous judgment result of the traffic light status, the autonomous judgment result of the traffic light status is adopted first, and the traffic light status is adopted first. Based on the priority traffic light status and combined with the pedestrian movement trajectory prediction received from the crosswalk area, the pedestrian's red light violation behavior is determined. A warning is triggered based on the act of running a red light.

[0009] This technical solution effectively addresses the problem of delayed or inconsistent transmission of external traffic light status information in existing technologies, which leads to lagging judgment in the warning system. By prioritizing the adoption of local autonomous judgment results, the real-time and accurate judgment of traffic light status is ensured, thereby triggering warnings in a timely manner before or at the initial stage of pedestrian red-light running, significantly improving the effectiveness and safety of the warnings.

[0010] Furthermore, acquiring and monitoring external traffic light status information, and when the update of external traffic light status information is delayed or inconsistent with the autonomous judgment result of traffic light status, prioritizing the autonomous judgment result of traffic light status, the steps to obtain the prioritized traffic light status include: Obtain visual status information of vehicle signal lights; Obtain information on the actual traffic status of vehicles in pedestrian crossing areas; Based on visual status information and actual vehicle traffic status information, determine the safety status of the area traversed by the traveler; When the update of external traffic light status information is delayed or the external traffic light status information is inconsistent with the autonomous judgment result of traffic light status, the safety status of the pedestrian crossing area is adopted first, and the traffic light status adopted first is obtained.

[0011] Through this technical solution, this application can further improve the accuracy and robustness of traffic light status judgment by cross-validating multi-source sensing data. Especially when external information is unreliable, it can make judgments that are more consistent with the actual traffic conditions, avoiding misjudgments or omissions, thereby improving the reliability of the warning system.

[0012] Based on the above, this application further proposes that, according to the traffic light status that is given priority and combined with the pedestrian movement trajectory prediction received from the crosswalk area, the steps for determining a pedestrian's red-light running behavior include: Acquire individual pedestrian movement data in crosswalk areas; Acquire data on non-pedestrian moving objects in the pedestrian crossing area; Based on individual pedestrian movement data and non-pedestrian moving object data, calculate the distance between pedestrian and vehicle travel paths, as well as the speed at which pedestrians cross boundaries; Analyze local crowd behavior patterns in pedestrian crossing areas based on individual pedestrian movement data; The pedestrian's red-light running behavior is determined based on the priority traffic signal status, the distance between the pedestrian and vehicle paths, the pedestrian's speed when crossing the boundary, and the behavior patterns of local crowds.

[0013] Through this technical solution, this application can comprehensively consider individual pedestrians, non-pedestrian moving objects, and crowd behavior patterns to make a more comprehensive and detailed judgment on pedestrians' red-light running behavior, thereby improving the accuracy of judgment and the timeliness of early warning, and effectively avoiding false alarms or missed alarms caused by judgment based on a single factor.

[0014] As an optional approach, the steps to determine a pedestrian's red-light violation behavior, based on the prioritized traffic light status and the received pedestrian movement trajectory prediction for the crosswalk area, include: Acquire data on non-pedestrian moving objects in the pedestrian crossing area; Analyze the relative motion between pedestrians and non-pedestrian moving objects based on non-pedestrian moving object data; Monitor the movement trajectory of pedestrians within the crosswalk area; Analyze the micro-behavioral characteristics of pedestrians; Based on the traffic light status, non-pedestrian moving object data, relative motion between pedestrians and non-pedestrian moving objects, motion trajectory and micro-behavioral characteristics, the pedestrian's red light running behavior is determined.

[0015] Through this technical solution, this application can gain a deeper understanding of pedestrians' intentions and potential risks by analyzing the relative motion between pedestrians and non-pedestrian moving objects, as well as the micro-behavioral characteristics of pedestrians. This allows for earlier and more accurate identification of red-light running behavior, providing valuable reaction time for the warning system.

[0016] In some preferred embodiments, the steps for determining a pedestrian's red-light violation based on the prioritized traffic light status and the received pedestrian movement trajectory prediction for the crosswalk area include: Acquire individual pedestrian motion data and pedestrian posture data; Based on individual pedestrian motion data and pedestrian posture data, analyze the relative distance and approach speed between pedestrians and the pedestrian crossing boundary; Analyze pedestrian gait and speed patterns based on individual pedestrian motion data; Analyze pedestrian head orientation and gaze focus based on pedestrian posture data; Based on individual pedestrian motion data and pedestrian posture data, identify pedestrian edge activities that are not intended to cross the boundary. The pedestrian's red-light running behavior is determined based on the priority traffic signal status, the relative distance and approach speed between the pedestrian and the crosswalk boundary, the pedestrian's gait pattern, the pedestrian's speed pattern, the pedestrian's head orientation, the pedestrian's visual focus, and the pedestrian's non-crossing edge activities.

[0017] Through this technical solution, this application can capture early signals of pedestrians' crossing intentions by conducting refined analysis of individual pedestrian movement and posture data, such as gait, speed, head orientation, and gaze focus. This allows for earlier prediction and judgment of pedestrians' red-light running behavior before they actually enter the danger zone, greatly improving the timeliness and effectiveness of early warning.

[0018] Furthermore, based on the traffic light status (which is the most frequently accepted factor), the relative distance and approach speed between the pedestrian and the crosswalk boundary, the pedestrian's gait pattern, speed pattern, head orientation, visual focus, and non-crossing edge movement, the steps to determine a pedestrian's red-light violation include: Acquire environmental sensor data; Based on environmental sensor data, the confidence levels of individual pedestrian motion data and pedestrian posture data are adjusted to obtain adjusted individual pedestrian motion data and adjusted pedestrian posture data. Based on the adjusted individual pedestrian motion data, the pedestrian's gait and speed patterns are analyzed to obtain the latest analyzed gait and speed patterns; Based on the adjusted pedestrian posture data, the pedestrian's head orientation and gaze focus are analyzed to obtain the latest analyzed head orientation and gaze focus; Based on the adjusted individual pedestrian motion data and adjusted pedestrian posture data, the non-crossing purpose edge activities of pedestrians are identified, and the latest identified non-crossing purpose edge activities are obtained. The pedestrian's red-light running behavior is determined based on the priority traffic signal status, the relative distance and approach speed between the pedestrian and the crosswalk boundary, the latest analyzed gait and speed patterns, the latest analyzed head orientation and gaze focus, and the latest identified non-crossing edge activities.

[0019] Through this technical solution, this application can dynamically adjust the confidence level of pedestrian movement and posture data by introducing environmental sensor data, thereby improving the accuracy and reliability of data analysis under complex and ever-changing environmental conditions, making the judgment of red-light running behavior more accurate, and reducing misjudgments caused by environmental interference.

[0020] Based on this, this application further proposes that, according to the traffic signal status (which is given priority), the relative distance and approach speed between the pedestrian and the crosswalk boundary, the latest analyzed gait and speed patterns, the latest analyzed head orientation and gaze focus, and the latest identified non-crossing edge activity, the steps for determining a pedestrian's red-light running behavior include: Adjust the confidence weights of individual pedestrian motion data and pedestrian posture data based on environmental sensor data; Based on the individual pedestrian motion data and pedestrian posture data after adjusting the confidence weight, the corrected relative distance and approach speed between the pedestrian and the pedestrian crossing boundary are calculated. Based on the individual pedestrian motion data after adjusting the confidence weights, we analyze the pedestrian gait patterns and speed patterns to obtain the corrected gait patterns and speed patterns. Based on pedestrian posture data after adjusting confidence weights, we analyze pedestrian head orientation and visual focus to obtain corrected head orientation and visual focus. Based on the pedestrian individual motion data and pedestrian posture data after adjusting the confidence weight, the non-crossing target edge activity of pedestrians is identified, and the corrected non-crossing target edge activity is obtained. The weight of each judgment criterion in the red light violation judgment is adjusted based on the priority traffic signal status, the corrected relative distance and approach speed between pedestrians and crosswalk boundaries, the corrected gait and speed patterns, the corrected head orientation and gaze focus, the corrected non-crossing edge activity, and the real-time confidence of each judgment criterion. The pedestrian's red-light running behavior is determined based on the traffic light status that is given priority, the corrected relative distance and approach speed between the pedestrian and the crosswalk boundary, the corrected gait and speed patterns, the corrected head orientation and gaze focus, the corrected non-crossing edge activity, and the adjusted weights of each judgment criterion in the red-light running judgment.

[0021] Through this technical solution, this application can dynamically adjust the weight of each judgment criterion, enabling the red light violation judgment model to adaptively optimize according to the real-time environment and data quality, thereby maintaining high-precision judgment capability in different scenarios and further improving the intelligence and robustness of the system.

[0022] To enhance functionality, the steps to trigger an alert based on red-light running behavior include: Obtain the status information of the warning device; Adjust and execute the warning information distribution strategy based on the status information of the warning devices and the red light violation; After implementing the warning information distribution strategy, obtain visual feedback information of the warning area; Based on the visual feedback information from the warning area, the activation method of the warning device in the warning information distribution strategy is adjusted to obtain the revised warning information distribution strategy. Activate the warning device according to the revised warning information distribution strategy.

[0023] Through this technical solution, this application can dynamically adjust the warning information distribution strategy and device activation method according to the status of the warning device and the visual feedback of the warning area, thereby realizing the intelligence and personalization of the warning, ensuring that the warning information can reach the target pedestrian in the most effective way, and improving the actual effect of the warning.

[0024] To improve the system, the steps to trigger a warning based on running a red light include: Obtain visual feedback information from pedestrians regarding the warning; Based on visual feedback information, it is possible to identify whether pedestrians are wearing hearing-impairing devices and whether they are using visual distraction devices, thus obtaining the identification results of hearing-impairing devices and visual distraction devices. Adjust the warning information distribution method based on visual feedback information, hearing impairment device recognition results, and visual distraction device recognition results; Activate the warning devices according to the adjusted warning information distribution method.

[0025] Through this technical solution, this application can intelligently adjust the distribution method of warning information based on the pedestrian's visual feedback to the warning and whether they are wearing hearing-impairing devices or using visual distraction devices, thereby providing more targeted and effective warnings based on the perceptual characteristics of different pedestrians, significantly improving the reach and effect of the warnings.

[0026] Secondly, this application also discloses a pedestrian red-light violation warning system based on the Internet of Things for intelligent transportation, used to execute pedestrian red-light violation warnings based on the Internet of Things for intelligent transportation, including: The perception data acquisition module is used to acquire multi-source local perception data at traffic intersections; The traffic light status judgment module is used to determine the current status of traffic lights based on multi-source local sensing data and obtain the autonomous judgment result of traffic light status. The priority acceptance determination module is used to acquire and monitor the status information of external traffic lights. When the update of the external traffic light status information is delayed or the external traffic light status information is inconsistent with the autonomous judgment result of the traffic light status, the autonomous judgment result of the traffic light status is given priority to obtain the traffic light status that is given priority. The red light violation judgment module is used to judge the pedestrian's red light violation behavior based on the priority traffic signal status and the pedestrian movement trajectory prediction received in the pedestrian crossing area. The warning judgment trigger module is used to trigger a warning based on the act of running a red light.

[0027] Through this technical solution, this application can provide a system that integrates multi-source perception, intelligent judgment and adaptive warning functions. It solves the problems of delay and inaccuracy of traffic light information in the prior art from both hardware and software levels, and realizes timely and accurate early warning of pedestrians running red lights, thereby comprehensively improving the level of traffic safety management.

[0028] Beneficial Effects: This application discloses a pedestrian red-light violation warning method based on the Internet of Things for intelligent transportation. It acquires multi-source local sensing data at traffic intersections and autonomously judges the current state of traffic lights based on this data, obtaining an autonomous judgment result for the traffic light status. Simultaneously, the system acquires and monitors external traffic light status information. When external information updates are delayed or inconsistent with the autonomous judgment result, the autonomous judgment result is prioritized, thus obtaining the prioritized traffic light status. Based on this, combined with pedestrian movement trajectory prediction in the crosswalk area, the system judges the pedestrian's red-light violation behavior and triggers a warning according to the judgment result.

[0029] Through the above technical solution, this application effectively solves the problem in existing technologies where unstable communication between the new traffic light controller and the IoT platform leads to delays or irregular lags in the transmission of traffic light status information, thus affecting the accuracy and timeliness of the warning system's judgment. Specifically, this application introduces local multi-source sensing data and an autonomous judgment mechanism, breaking the over-reliance on a single external information source. Even when network congestion causes delays or inconsistencies in external information, it can ensure the real-time and accuracy of traffic light status judgment. This enables the warning system to promptly and accurately identify and trigger warnings before or at the initial stage of pedestrian red-light running, i.e., within the critical time window before pedestrians enter the danger zone, thereby avoiding the drawback of delayed warnings that significantly reduce the warning effect in existing technologies. Therefore, this application significantly improves the timeliness, accuracy, and effectiveness of pedestrian red-light running warnings, providing more reliable technical support for urban traffic safety management. Attached Figure Description

[0030] Figure 1 This is a flowchart of a pedestrian red-light violation warning method based on the Internet of Things for intelligent transportation, as one embodiment of the present invention. Figure 2 This is a flowchart of a pedestrian red-light violation warning method based on the Internet of Things for intelligent transportation, according to another embodiment of the present invention. Figure 3 This is a system block diagram of a pedestrian red-light violation warning system based on the Internet of Things for intelligent transportation, according to another embodiment of the present invention. Explanation of reference numerals in the attached figures: 1. Pedestrian red-light violation warning system based on intelligent transportation IoT; 11. Sensing data acquisition module; 12. Traffic light status judgment module; 13. Priority data acquisition determination module; 14. Red-light violation behavior judgment module; 15. Warning judgment trigger module. Detailed Implementation

[0031] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0032] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0033] A pedestrian red-light violation warning method based on the Internet of Things (IoT) for intelligent transportation addresses the issue that traditional intelligent transportation IoT pedestrian red-light violation warning systems in urban traffic management typically rely on real-time traffic signal status information transmitted from the city traffic control center to determine pedestrian red-light violation behavior and trigger warnings. However, the introduction of new traffic signal controllers and congestion of public wireless network channels have led to irregular delays in the transmission of traffic signal status information, causing the warning system's judgment of the traffic signal status to become disconnected from the actual physical world state. This delayed judgment, especially during the critical time window when pedestrians rush to cross crosswalks, causes the warning system to miss the best opportunity to effectively dissuade pedestrians before they enter the danger zone, thus reducing the warning effect.

[0034] In response, this application proposes a pedestrian red-light violation warning method based on the Internet of Things for intelligent transportation, combined with... Figure 1 As shown, it includes: S1, acquire multi-source local sensing data of traffic intersections; S2, based on multi-source local sensing data, determine the current state of the traffic lights and obtain the autonomous judgment result of the traffic light state; S3, acquire and monitor external traffic light status information. When the external traffic light status information is delayed or inconsistent with the autonomous judgment result of the traffic light status, the autonomous judgment result of the traffic light status is adopted first to obtain the traffic light status that is adopted first. S4. Based on the priority traffic light status and combined with the pedestrian movement trajectory prediction received from the crosswalk area, determine the pedestrian's red light violation behavior. S5 triggers a warning based on the act of running a red light.

[0035] To better understand the method proposed in this application, some key terms and implementation environments involved will be explained first.

[0036] "Multi-source local sensing data" refers to real-time data acquired at traffic intersections through various sensor devices, including but not limited to video surveillance data, radar data, lidar data, and geomagnetic sensor data. This data can provide detailed information about various traffic participants and the environment at traffic intersections, such as vehicles, pedestrians, and traffic lights.

[0037] "Autonomous judgment result of traffic signal status" refers to the result obtained by the system independently analyzing and judging the current color (red, yellow, green) and status (e.g., whether it is flashing) of the traffic signal based on local sensing data.

[0038] "External traffic light status information" refers to traffic light status data transmitted by the city traffic control center or other external systems.

[0039] "Priority traffic light status" refers to the traffic light status that is finally determined by the system after decision-making when the local autonomous judgment result is inconsistent with external information or when external information is delayed, and is used for subsequent judgment.

[0040] "Pedestrian trajectory prediction" refers to predicting the movement path and location of pedestrians in the future based on their historical movement data and current movement trends.

[0041] "Running a red light" refers to the act of a pedestrian entering or attempting to enter a pedestrian crossing area when the traffic light is red.

[0042] This method is typically deployed in an intelligent transportation IoT platform, which integrates various sensors, data processing units, communication modules, and warning devices to perceive traffic conditions in real time, process data, make decisions, and execute warning actions.

[0043] The pedestrian red-light violation warning method proposed in this application is based on the Internet of Things for intelligent transportation. Its core lies in achieving accurate judgment and timely warning of pedestrian red-light violation behavior through a series of collaborative steps.

[0044] First, the method includes "acquiring multi-source local sensing data at traffic intersections." Specifically, data can be collected by deploying various sensors at traffic intersections. For example, high-definition cameras can be installed to acquire video images of the intersection, capturing the real-time dynamics of vehicles and pedestrians; millimeter-wave radar or lidar can be configured to accurately measure the speed, position, and distance information of vehicles and pedestrians; in addition, geomagnetic sensors or pressure sensors can be used to detect the presence of pedestrians or vehicles in crosswalk areas. These sensors can operate independently or cooperate with each other to build a comprehensive traffic environment sensing network. For example, video data can be used to identify the color of traffic lights, while radar data can be used to track the precise location and speed of pedestrians.

[0045] Secondly, this method involves "determining the current state of traffic lights based on multi-source local sensing data and obtaining an autonomous judgment result of the traffic light state." After acquiring multi-source local sensing data, the system uses this data to autonomously determine the current state of traffic lights. For example, it can perform image recognition and processing on video images, analyzing changes in the color and brightness of the traffic lights to determine whether they are red, yellow, or green. It can also indirectly infer the state of the traffic lights by analyzing traffic flow data, such as vehicle traffic patterns and pedestrian waiting area clustering. This autonomous judgment mechanism allows the system to have a basic understanding of the traffic light state even when there is no external traffic light information or the external information is unreliable.

[0046] Furthermore, the method also includes "acquiring and monitoring external traffic light status information. When the update of external traffic light status information is delayed or inconsistent with the autonomous judgment result of the traffic light status, the autonomous judgment result of the traffic light status is given priority, and the traffic light status is given priority." The system continuously receives traffic light status information from the city traffic control center or other external systems. Simultaneously, the system monitors the update frequency and timeliness of this external information. When it is found that external information has not been updated for a long time (i.e., update is delayed), or when external information contradicts the local autonomous judgment result, the system will activate the priority acceptance mechanism. In this case, the system will choose to accept the traffic light status determined locally, rather than the external information. For example, if external information shows that the current light is green, but the local video analysis result clearly shows that the light has turned red, and the external information has not been updated for a long time, the system will give priority to accepting the local judgment of the red light status.

[0047] Next, the method includes "determining whether a pedestrian has run a red light based on the prioritized traffic light status and the received pedestrian trajectory prediction for the crosswalk area." After determining the prioritized traffic light status, the system combines the predicted pedestrian trajectory for the crosswalk area to determine whether a red light violation has occurred. For example, if the prioritized traffic light status is red, and the system predicts that a pedestrian is about to enter the crosswalk area, or has already entered the crosswalk area, then the system will determine that the pedestrian has run a red light. Pedestrian trajectory prediction can be achieved by analyzing information such as the pedestrian's historical movement data, current speed, direction, and distance from the crosswalk boundary.

[0048] Finally, the method includes "triggering warnings based on red-light running behavior." Once the system detects red-light running, it immediately triggers the corresponding warning. The warning can take many forms; for example, it can broadcast a warning message "Please do not run red lights" via a voice announcement system; it can display warning information on LED screens on both sides of the pedestrian crossing; or it can project warning images in front of pedestrians using ground projection equipment. The purpose of the warning is to promptly remind pedestrians to pay attention to safety and avoid traffic accidents.

[0049] The pedestrian red-light violation warning method based on the Internet of Things for intelligent transportation proposed in this application effectively solves the problem of delayed warning caused by the delay of traffic light information in traditional systems by integrating multi-source local sensing data and external traffic light status information and introducing an autonomous judgment and priority acceptance mechanism. Optional, combined Figure 2As shown, S3 acquires and monitors external traffic light status information. When the update of external traffic light status information is delayed or the external traffic light status information is inconsistent with the autonomous judgment result of traffic light status, the autonomous judgment result of traffic light status is adopted first. The steps to obtain the traffic light status that is adopted first include: S31, Obtain visual status information of vehicle signal lights; S32, obtain information on the actual traffic status of vehicles in the pedestrian crossing area; S33, based on visual status information and actual vehicle traffic status information, determines the safety status of the area crossed by the traveler; S34. When the update of external traffic light status information is delayed or the external traffic light status information is inconsistent with the autonomous judgment result of traffic light status, the safety status of the pedestrian crossing area shall be given priority, and the traffic light status given priority shall be obtained.

[0050] Specifically, "visual status information of vehicle traffic lights" refers to the status data obtained through real-time identification and analysis of the color, flashing frequency, etc., of vehicle traffic lights at traffic intersections using visual sensors (such as cameras). Its purpose is to directly perceive visual indications of vehicle passage rights. "Actual vehicle passage status information in pedestrian crossing areas" can be understood as real-time monitoring of data such as whether vehicles are crossing in the pedestrian crossing area, their speed, direction, and distance from the pedestrian crossing boundary, using sensors such as radar, lidar, or high-precision visual recognition. Its purpose is to obtain the true dynamics of vehicles in the pedestrian crossing area to assess potential risks to pedestrians. In practical applications, "safety status of pedestrian crossing areas" specifically refers to the real-time assessment of whether pedestrians can safely cross the pedestrian crossing, considering both the visual status of vehicle traffic lights and the actual vehicle passage situation in the pedestrian crossing area. For example, when the traffic light is red and there are no vehicles crossing the pedestrian crossing area or the vehicles have come to a complete stop, the safety status of the pedestrian crossing area can be judged as safe; conversely, when the traffic light is green or a vehicle is speeding across the pedestrian crossing area, the safety status of the pedestrian crossing area is judged as unsafe.

[0051] In some preferred embodiments, a specific example is given below. Suppose at a traffic intersection, the external traffic light control system experiences a communication failure, resulting in a delay in updating its status information, or its reported status is inconsistent with the status determined by local perception data (e.g., traffic light colors identified by an intersection camera). In this case, prioritizing only the local autonomous judgment may pose certain risks. According to the solution of this application, the system first acquires the visual status information of vehicle traffic lights, for example, by capturing and analyzing images of vehicle traffic lights in real time through a high-definition camera installed at the intersection, identifying that the current vehicle traffic light is red. Simultaneously, the system acquires the actual vehicle traffic status information of the pedestrian crossing area, for example, by detecting, through geomagnetic sensors, radar, or visual recognition technology, that there are currently no vehicles crossing the pedestrian crossing area, or that all vehicles have come to a complete stop before the stop line. Based on this visual status information and the actual vehicle traffic status information, the system can determine that the safety status of the pedestrian crossing area is "safe." At this time, even if the update of the external traffic light status information is delayed or inconsistent with the autonomous judgment result, the system will prioritize this "safe status of the pedestrian crossing area" determined based on actual traffic flow. For example, if an external traffic light displays a green light (but may be expired or incorrect), while the local system determines it to be red, but visual and actual vehicle traffic monitoring confirms that the vehicle's traffic light is indeed red and there are no vehicles at the crosswalk, then the system will accept the "safety status of the pedestrian crossing area" as the priority traffic light status, guiding subsequent judgments and warnings regarding pedestrian red-light jaywalking. This approach ensures that, in critical moments, the system can make decisions based on data closest to the actual safety situation, thereby effectively preventing potential traffic accidents.

[0052] Optionally, the steps for determining a pedestrian's red-light violation based on the prioritized traffic light status and the received pedestrian movement trajectory prediction for the crosswalk area include: Acquire individual pedestrian movement data in crosswalk areas; Acquire data on non-pedestrian moving objects in the pedestrian crossing area; Based on individual pedestrian movement data and non-pedestrian moving object data, calculate the distance between pedestrian and vehicle travel paths, as well as the speed at which pedestrians cross boundaries; Analyze local crowd behavior patterns in pedestrian crossing areas based on individual pedestrian movement data; The pedestrian's red-light running behavior is determined based on the priority traffic signal status, the distance between the pedestrian and vehicle paths, the pedestrian's speed when crossing the boundary, and the behavior patterns of local crowds.

[0053] Pedestrian individual motion data refers to the dynamic information of a single pedestrian within the crosswalk area, including their position, speed, acceleration, and direction. This data can be acquired through various sensors, such as high-definition cameras, LiDAR, and millimeter-wave radar, and processed by target detection and tracking algorithms. Its purpose is to accurately grasp the real-time dynamics of each pedestrian. Non-pedestrian moving object data refers to data on non-pedestrian targets moving within or near the crosswalk area, such as vehicles, bicycles, and electric vehicles. This data can also be acquired through visual or radar sensors and used to identify potential sources of traffic conflict. Its purpose is to comprehensively assess the complexity of the traffic environment within the crosswalk area.

[0054] Furthermore, calculating the distance between pedestrian and vehicle paths, as well as the pedestrian's speed across the boundary, can be achieved using path prediction algorithms based on acquired individual pedestrian motion data and non-pedestrian moving object data. This estimates the minimum distance between the pedestrian's expected path and the vehicle's expected path. Simultaneously, the distance between the pedestrian's current position and the crosswalk boundary can be calculated, and combined with their velocity vector, the pedestrian's speed across the crosswalk boundary can be predicted. The aim is to quantify the proximity of pedestrians to potential hazards and the urgency of pedestrians entering the danger zone. In addition, analyzing local crowd behavior patterns in crosswalk areas involves comprehensively analyzing the motion data of multiple individual pedestrians within the crosswalk area to identify patterns such as overall crowd movement trends, aggregation and dispersion, following behavior, and avoidance behavior. For example, when a group of pedestrians collectively moves towards the crosswalk boundary, even if an individual pedestrian has not yet fully entered the danger zone, it may indicate a potential risk of running a red light. The purpose is to assist in judging red-light running behavior from a macro perspective, improving the accuracy and predictability of the judgment.

[0055] Optionally, the steps for determining a pedestrian's red-light violation based on the prioritized traffic light status and the received pedestrian movement trajectory prediction for the crosswalk area include: Acquire data on non-pedestrian moving objects in the pedestrian crossing area; Analyze the relative motion between pedestrians and non-pedestrian moving objects based on non-pedestrian moving object data; Monitor the movement trajectory of pedestrians within the crosswalk area; Analyze the micro-behavioral characteristics of pedestrians; Based on the traffic light status, non-pedestrian moving object data, relative motion between pedestrians and non-pedestrian moving objects, motion trajectory and micro-behavioral characteristics, the pedestrian's red light running behavior is determined.

[0056] Specifically, acquiring data on non-pedestrian moving objects in pedestrian crossing areas refers to the perception data generated by all moving objects within the pedestrian crossing area, excluding pedestrians, such as bicycles, electric bikes, scooters, pets, and even larger objects blown by the wind. This data can be acquired using devices such as visual sensors, radar sensors, or lidar, with the aim of comprehensively perceiving the dynamic environment within the pedestrian crossing area.

[0057] Analyzing the relative motion between pedestrians and non-pedestrian moving objects based on their data can be understood as calculating parameters such as the rate of change of distance, relative speed, and relative direction between pedestrians and these non-pedestrian moving objects. For example, when a bicycle rapidly approaches a pedestrian on a crosswalk, its relative motion data will reveal the potential collision risk or impact on pedestrian behavior. The purpose is to assess the immediate risk to the pedestrian and the potential interference of other moving objects on the pedestrian's decision-making.

[0058] In practical applications, monitoring the movement trajectory of pedestrians within crosswalk areas involves continuously tracking their location information and mapping their movement paths over time. This can be achieved using high-precision positioning technology, visual tracking algorithms, or multi-sensor fusion technology. The goal is to obtain information on the continuity and directionality of pedestrian movement, providing a foundation for subsequent behavioral analysis.

[0059] Furthermore, analyzing pedestrians' micro-behavioral characteristics involves recognizing and analyzing subtle movements, postures, head orientation, gaze focus, and gait changes. For example, a pedestrian's sudden pause, facing the intersection, frequent head movements to observe vehicles, or increased walking speed may all indicate an intention to run a red light or hesitation. These characteristics can be extracted using high-resolution cameras combined with artificial intelligence visual analysis algorithms, with the aim of gaining a deeper understanding of pedestrians' potential intentions and risk propensities.

[0060] In some preferred embodiments, a specific example is given below. Suppose at a traffic intersection, the traffic light is red, prohibiting pedestrians from crossing. At this time, a pedestrian is standing on the edge of a crosswalk, and their movement trajectory shows a tendency to move towards the center of the intersection. If only the pedestrian's movement trajectory is used for prediction, the system may determine that there is a risk of running a red light. However, the solution of this application further acquires data on non-pedestrian moving objects in the crosswalk area. For example, the system detects an electric bicycle approaching the pedestrian quickly from behind, with a tendency to turn towards the crosswalk. The system analyzes the relative motion between the pedestrian and the electric bicycle, finding that the electric bicycle is approaching the pedestrian at a high speed, and its path may intersect with the pedestrian's potential crossing path. At the same time, the system monitors the pedestrian's movement trajectory and analyzes their micro-behavioral characteristics, such as the pedestrian's head facing the direction of the electric bicycle, accompanied by a slight backward movement of the body, or their gait pattern showing hesitation. By comprehensively considering the traffic light status (red light), non-pedestrian moving object data (electric bicycles), the relative motion between pedestrians and non-pedestrian moving objects, the pedestrian's movement trajectory, and the pedestrian's micro-behavioral characteristics, the system can more accurately determine whether a pedestrian has a potential risk of running a red light, or whether their behavior has become uncertain due to the influence of an electric bicycle. This allows the system to trigger warnings in a timely manner, reminding pedestrians and electric bicycle drivers to pay attention to safety and effectively avoid potential traffic accidents.

[0061] Optionally, the steps for determining a pedestrian's red-light violation based on the prioritized traffic light status and the received pedestrian movement trajectory prediction for the crosswalk area include: Acquire individual pedestrian motion data and pedestrian posture data; Based on individual pedestrian motion data and pedestrian posture data, analyze the relative distance and approach speed between pedestrians and the pedestrian crossing boundary; Analyze pedestrian gait and speed patterns based on individual pedestrian motion data; Analyze pedestrian head orientation and gaze focus based on pedestrian posture data; Based on individual pedestrian motion data and pedestrian posture data, identify pedestrian edge activities that are not intended to cross the boundary. The pedestrian's red-light running behavior is determined based on the priority traffic signal status, the relative distance and approach speed between the pedestrian and the crosswalk boundary, the pedestrian's gait pattern, the pedestrian's speed pattern, the pedestrian's head orientation, the pedestrian's visual focus, and the pedestrian's non-crossing edge activities.

[0062] Specifically, pedestrian motion data can be understood as data describing the dynamic characteristics of pedestrians, such as spatial position, speed, and acceleration. This data can be acquired through sensors such as high-definition cameras, LiDAR, or millimeter-wave radar deployed at traffic intersections. For example, target tracking algorithms can be used to continuously locate pedestrians in video streams, thereby obtaining their trajectory, instantaneous speed, and direction over a period of time. The purpose is to provide quantitative information on pedestrian movement behavior.

[0063] Pedestrian posture data refers to data describing the static and dynamic posture characteristics of pedestrians, such as the relative positions of various body parts, joint angles, head orientation, and gaze direction. It is typically obtained through computer vision techniques, such as deep learning-based human posture estimation models, by analyzing pedestrians in images or videos. Its purpose is to reveal the pedestrian's body language and underlying intentions.

[0064] In practical applications, analyzing the relative distance and approach speed between pedestrians and the crosswalk boundary involves calculating the Euclidean distance between the pedestrian's current position and the crosswalk's starting or stopping line, and combining this with the pedestrian's direction and speed of movement to predict whether they will enter the crosswalk area in the near future and at what speed. For example, when a pedestrian is close to the boundary and moving towards it at a relatively fast speed, it may indicate an intention to cross. The purpose is to quantify the physical proximity of pedestrians to dangerous areas.

[0065] Furthermore, analyzing pedestrian gait and speed patterns involves extracting characteristics of pedestrian movement, such as stride frequency, stride length, average speed, and rate of change of speed, through time-series analysis of individual pedestrian motion data. For example, a brisk gait or sudden acceleration may indicate that a pedestrian is eager to cross. The aim is to assess the urgency and intention of pedestrians from a dynamic behavioral perspective.

[0066] Furthermore, analyzing pedestrian head orientation and gaze focus involves using pedestrian posture data to determine whether a pedestrian's head is facing the direction of traffic flow or traffic lights, and whether their gaze is focused on the traffic environment. For example, pedestrians whose heads are facing traffic lights and whose gaze is focused on vehicles typically have a higher level of attention to traffic conditions. The purpose is to assess pedestrian attention distribution and perception of the traffic environment.

[0067] As a preferred implementation method, identifying pedestrians' non-crossing-intention edge activities refers to distinguishing between activities such as waiting at the edge of a crosswalk, talking to others, checking a mobile phone, or observing the surrounding environment—activities not intended for crossing the road—and genuinely preparing to cross, by comprehensively analyzing individual pedestrian movement data and pedestrian posture data. For example, a pedestrian who lingers near the boundary for an extended period with their head facing a direction other than traffic may be engaged in activities not intended for crossing. The aim is to eliminate false positives and improve the accuracy of judgment.

[0068] Therefore, by comprehensively considering the traffic light status, the relative distance and approach speed between the pedestrian and the crosswalk boundary, the pedestrian's gait pattern, the pedestrian's speed pattern, the pedestrian's head orientation, the pedestrian's gaze focus, and the pedestrian's non-crossing edge activities, a multi-feature fusion judgment model can be constructed, such as a machine learning-based classifier or expert rule system, to make a final judgment on the pedestrian's red light violation.

[0069] In some preferred embodiments, this application is implemented as follows. Assume that at a certain intersection, the traffic light status autonomously determines that the current pedestrian crossing is red.

[0070] Specifically, when a pedestrian approaches the edge of a crosswalk, the system first acquires the pedestrian's individual motion and posture data. If the pedestrian's individual motion data shows that they linger near the crosswalk boundary for an extended period, maintaining a constant relative distance to the boundary or moving slowly parallel to it, while their approach speed is close to zero; and their posture data shows that their head is facing away from the roadway, their gaze is focused on a mobile phone screen, and their gait and speed patterns both indicate stillness or slow movement, and the system identifies that they are engaging in edge-moving activity unrelated to crossing the crosswalk, then, despite being near the crosswalk, the system will determine that the pedestrian does not intend to jaywalk, thus avoiding triggering unnecessary warnings.

[0071] For example, when another pedestrian moves rapidly toward the crosswalk boundary, the system acquires their individual motion and posture data. If the pedestrian's individual motion data shows that their relative distance to the crosswalk boundary is rapidly decreasing, their approach speed is high, and their gait and speed patterns indicate a rapid, brisk walk; simultaneously, their posture data shows that their head is facing the roadway, their gaze is focused on an oncoming vehicle or traffic light, and no non-crossing edge activity is detected. In this case, even if the traffic light is red, the system will determine based on these combined characteristics that the pedestrian has run a red light and immediately trigger a warning to alert pedestrians and passing vehicles to be aware of safety.

[0072] Optionally, the steps to determine a pedestrian's red-light running behavior based on the preferred traffic light status, the relative distance and approach speed between the pedestrian and the crosswalk boundary, the pedestrian's gait pattern, the pedestrian's speed pattern, the pedestrian's head orientation, the pedestrian's visual focus, and the pedestrian's non-crossing edge activity include: Acquire environmental sensor data; Based on environmental sensor data, the confidence levels of individual pedestrian motion data and pedestrian posture data are adjusted to obtain adjusted individual pedestrian motion data and adjusted pedestrian posture data. Based on the adjusted individual pedestrian motion data, the pedestrian's gait and speed patterns are analyzed to obtain the latest analyzed gait and speed patterns; Based on the adjusted pedestrian posture data, the pedestrian's head orientation and gaze focus are analyzed to obtain the latest analyzed head orientation and gaze focus; Based on the adjusted individual pedestrian motion data and adjusted pedestrian posture data, the non-crossing purpose edge activities of pedestrians are identified, and the latest identified non-crossing purpose edge activities are obtained. The pedestrian's red-light running behavior is determined based on the priority traffic signal status, the relative distance and approach speed between the pedestrian and the crosswalk boundary, the latest analyzed gait and speed patterns, the latest analyzed head orientation and gaze focus, and the latest identified non-crossing edge activities.

[0073] Acquiring environmental sensor data refers to collecting real-time data reflecting the current environmental conditions using environmental sensors deployed at traffic intersections, such as light sensors, rain sensors, fog sensors, and wind speed sensors. This data is used to assess environmental factors affecting the quality of pedestrian motion and posture data collection.

[0074] Based on environmental sensor data, the confidence levels of individual pedestrian motion and posture data are adjusted to obtain adjusted individual pedestrian motion and posture data. This means that when environmental sensor data indicates unfavorable conditions for data acquisition, such as insufficient lighting or reduced visibility due to rain or fog, the system will correspondingly lower the confidence levels of the individual pedestrian motion and posture data acquired through visual sensors. Conversely, in favorable environments, the confidence levels remain high. This adjustment process can be based on preset rules or machine learning models.

[0075] Based on the adjusted individual pedestrian motion data, the gait and speed patterns of pedestrians are analyzed to obtain the latest analyzed gait and speed patterns. This means that after the original individual pedestrian motion data has been adjusted for confidence, the gait and speed patterns are re-analyzed using this adjusted data. Because the reliability of the input data is improved, the obtained gait and speed patterns will more accurately reflect the pedestrian's true movement intentions.

[0076] Based on the adjusted pedestrian posture data, the head orientation and gaze focus of pedestrians are analyzed to obtain the latest analyzed head orientation and gaze focus. This means that after the original pedestrian posture data has been adjusted for confidence, the head orientation and gaze focus are re-analyzed using this adjusted data. This helps to more accurately determine whether pedestrians are observing traffic conditions or have the intention to cross.

[0077] Based on adjusted individual pedestrian movement and posture data, the system identifies pedestrians' non-crossing-purpose edge activities. This latest identification of non-crossing-purpose edge activities refers to re-identifying non-crossing activities of pedestrians in the edge area of ​​a crosswalk, such as waiting, observing, or making phone calls, based on the adjusted data. This helps distinguish between genuine red-light running and simply stopping at the roadside.

[0078] As a specific implementation method, a concrete example is given below. Suppose that in rainy or foggy weather conditions, the light and rain sensors at a traffic intersection detect low visibility and rainfall. In this situation, the environmental sensor data acquired by the system indicates poor environmental conditions. Based on this data, the system reduces the confidence level of the pedestrian motion and posture data acquired by the camera. For example, if the confidence level is normally 0.9, it might be adjusted to 0.6 in rainy or foggy weather. Subsequently, the system uses this adjusted, lower-confidence pedestrian motion and posture data to reanalyze pedestrian gait patterns, speed patterns, head orientation, gaze focus, and non-crossing edge activities. For example, in low visibility conditions, the system might be more cautious in judging subtle changes in pedestrian posture, avoiding misjudgments due to image blur. Ultimately, based on the prioritized traffic light status, the relative distance and approach speed between pedestrians and the crosswalk boundary, and these newly analyzed, more robust behavioral characteristics, the system can more accurately determine pedestrians' red-light running behavior, maintaining high accuracy even in inclement weather.

[0079] Optionally, the steps for determining a pedestrian's red-light running behavior, based on the preferred traffic light status, the relative distance and approach speed between the pedestrian and the crosswalk boundary, the latest analyzed gait and speed patterns, the latest analyzed head orientation and gaze focus, and the latest identified non-crossing edge activity, include: Adjust the confidence weights of individual pedestrian motion data and pedestrian posture data based on environmental sensor data; Based on the individual pedestrian motion data and pedestrian posture data after adjusting the confidence weight, the corrected relative distance and approach speed between the pedestrian and the pedestrian crossing boundary are calculated. Based on the individual pedestrian motion data after adjusting the confidence weights, we analyze the pedestrian gait patterns and speed patterns to obtain the corrected gait patterns and speed patterns. Based on pedestrian posture data after adjusting confidence weights, we analyze pedestrian head orientation and visual focus to obtain corrected head orientation and visual focus. Based on the pedestrian individual motion data and pedestrian posture data after adjusting the confidence weight, the non-crossing target edge activity of pedestrians is identified, and the corrected non-crossing target edge activity is obtained. The weight of each judgment criterion in the red light violation judgment is adjusted based on the priority traffic signal status, the corrected relative distance and approach speed between pedestrians and crosswalk boundaries, the corrected gait and speed patterns, the corrected head orientation and gaze focus, the corrected non-crossing edge activity, and the real-time confidence of each judgment criterion. The pedestrian's red-light running behavior is determined based on the traffic light status that is given priority, the corrected relative distance and approach speed between the pedestrian and the crosswalk boundary, the corrected gait and speed patterns, the corrected head orientation and gaze focus, the corrected non-crossing edge activity, and the adjusted weights of each judgment criterion in the red-light running judgment.

[0080] Specifically, adjusting the confidence weights of individual pedestrian motion and posture data involves assigning different reliability or importance coefficients to the original individual pedestrian motion and posture data based on environmental sensor data (such as light intensity, weather conditions, and visibility). For example, at night or in rainy or foggy weather, the confidence level of pedestrian posture data acquired by visual sensors may be low, and its weight will be adjusted accordingly, while the weight of motion data acquired by radar or lidar may be increased. From this, the corrected relative distance and approach speed between the pedestrian and the crosswalk boundary, the corrected gait and speed patterns, the corrected head orientation and gaze focus, and the corrected non-crossing edge activity can be calculated. These corrected data and patterns reflect the actual reliability of each judgment criterion under the current environmental conditions.

[0081] The real-time confidence level of each judgment criterion refers to the degree to which each factor used to judge red-light running behavior (such as the distance between the pedestrian and the boundary, gait pattern, head orientation, etc.) is considered accurate or reliable at a specific moment. For example, when the pedestrian's head orientation data is blurred due to strong backlighting, its confidence level will decrease. Adjusting the weight of each judgment criterion in red-light running judgment means dynamically changing the influence of each judgment criterion in the final decision-making process based on these real-time confidence levels. For example, when the confidence level of gait pattern is high, its weight in the judgment will be increased, while when the confidence level of gaze focus is low, its weight will be decreased. In this way, it can be ensured that the system can make judgments on red-light running behavior more flexibly and accurately under different environmental conditions.

[0082] In some preferred embodiments, a specific example is given below. Suppose at a traffic intersection, at night, with insufficient light and light fog. In this situation, the confidence level of pedestrian posture data (such as head orientation and gaze focus) acquired by visual sensors will significantly decrease due to reduced visibility. Based on environmental sensor data (such as light sensor and visibility sensor data), the system will adjust the confidence weight of the pedestrian posture data to reduce its influence in subsequent analysis. Simultaneously, the confidence level of individual pedestrian motion data acquired by radar sensors (such as relative distance and approach speed to the crosswalk boundary, gait pattern, and speed pattern) is relatively high, and its weight will be increased accordingly. When determining whether a pedestrian has run a red light, the system will calculate the corrected judgment criteria based on this adjusted data. For example, the corrected head orientation and gaze focus may have less impact on the final judgment due to lower confidence levels, while the corrected relative distance and approach speed between the pedestrian and the crosswalk boundary, gait pattern, and speed pattern will have greater decision weight due to higher confidence levels. In addition, the system will evaluate the confidence level of these revised judgment criteria in real time. For example, if a pedestrian's gait pattern can still be clearly identified in fog, its real-time confidence level will remain high, thus playing a more important role in the final judgment. Ultimately, the system will combine the traffic light status that is given priority and the various revised judgment criteria that have been dynamically weighted to arrive at the final judgment on the pedestrian's red-light running behavior.

[0083] Optionally, the steps to trigger a warning based on the act of running a red light include: Obtain the status information of the warning device; Adjust and execute the warning information distribution strategy based on the status information of the warning devices and the red light violation; After implementing the warning information distribution strategy, obtain visual feedback information of the warning area; Based on the visual feedback information from the warning area, the activation method of the warning device in the warning information distribution strategy is adjusted to obtain the revised warning information distribution strategy. Activate the warning device according to the revised warning information distribution strategy.

[0084] Specifically, acquiring warning device status information refers to the system's real-time monitoring and collection of the operational status of various warning devices (such as audible and visual warning devices, ground projection lights, and voice broadcasters) deployed at traffic intersections or pedestrian crossings. This status information can include parameters such as the device's online or offline status, battery level, fault codes, brightness, and volume, with the aim of ensuring that subsequent warning information can be effectively conveyed through normally functioning devices.

[0085] The strategy of adjusting and executing warning information distribution based on the status information of warning devices and the red-light running behavior can be understood as follows: after detecting a pedestrian running a red light, the system does not simply trigger a preset warning, but first dynamically selects the most appropriate combination of warning devices, warning intensity, and distribution timing based on the actual operating status of the current warning devices. For example, if a voice broadcaster is malfunctioning, the system will prioritize other normally functioning audio-visual warning devices or adjust their activation priority. This strategy aims to improve the initial effectiveness of the warnings.

[0086] In practical applications, after implementing the warning information distribution strategy, obtaining visual feedback information of the warning area refers to the system using visual sensors (such as cameras) to monitor the warning area in real time to capture pedestrians' reactions to the issued warning. This visual feedback information can include the pedestrian's head orientation, gait changes, whether they pause, whether they look at the warning source, etc., with the aim of evaluating the actual effect of the current warning.

[0087] Furthermore, based on visual feedback information from the warning area, the activation method of the warning devices in the warning information distribution strategy is adjusted, resulting in a revised warning information distribution strategy. This means that the system iteratively optimizes the original warning strategy based on pedestrians' real-time reactions to the warnings. For example, if visual feedback shows that pedestrians do not react significantly to the current warning, the system may adjust the activation method of the warning devices, such as increasing the warning volume, changing the flashing frequency of the lights, activating more types of warning devices, or projecting the warning information to a location that is more easily noticed by pedestrians, in order to enhance the penetration and attractiveness of the warning.

[0088] Therefore, by activating the warning devices according to the revised warning information distribution strategy, the final warning is evaluated and optimized in real time, which can attract pedestrians' attention to the greatest extent and prompt them to stop running red lights.

[0089] In some preferred embodiments, a specific example is given below. Suppose at a busy traffic intersection, the system determines the traffic light is red based on sensor data and detects a pedestrian preparing to jaywalk. At this point, the system first obtains the status information of all warning devices at the intersection. For example, it finds that the voice announcer on the left side of the intersection is offline due to a line fault, while the ground projection light and audible / visual warning device on the right side are working normally. Based on this, the system adjusts its warning information distribution strategy, prioritizing the activation of the ground projection light and audible / visual warning device, projecting a "No Crossing" warning pattern onto the ground in the direction the pedestrian is traveling, and simultaneously emitting a warning sound. After the warning is executed, the system obtains visual feedback information from the warning area via a camera, finding that although the pedestrian heard the warning sound, their head did not turn towards the warning source, and their gait did not change significantly, indicating a continued tendency to move forward. Based on this visual feedback, the system determined that the initial warning effect was insufficient. Therefore, it further adjusted the activation method of the warning devices in the warning information distribution strategy. For example, it increased the brightness of the ground projection lights to maximum and changed their projected pattern to a more impactful flashing red "Stop" sign. Simultaneously, it increased the volume and flashing frequency of the audible and visual sirens and activated the LED display screen above the intersection to play a warning animation. Ultimately, according to the revised warning information distribution strategy, the system activated these devices, attracting pedestrians with a stronger warning message and ultimately stopping them from running the red light.

[0090] Optionally, the steps to trigger a warning based on the act of running a red light include: Obtain visual feedback information from pedestrians regarding the warning; Based on visual feedback information, it is possible to identify whether pedestrians are wearing hearing-impairing devices and whether they are using visual distraction devices, thus obtaining the identification results of hearing-impairing devices and visual distraction devices. Adjust the warning information distribution method based on visual feedback information, hearing impairment device recognition results, and visual distraction device recognition results; Activate the warning devices according to the adjusted warning information distribution method.

[0091] Specifically, "acquiring pedestrians' visual feedback to warnings" refers to capturing visual data such as facial expressions, body movements, and gaze direction of pedestrians before and after receiving warning information through visual sensors (such as high-definition cameras) deployed at traffic intersections or pedestrian crossings. This data can reflect the degree of pedestrians' perception and reaction to the warning.

[0092] The "identification of whether pedestrians are wearing hearing-impairing devices" can be understood as using image recognition or deep learning algorithms to analyze visual feedback information and detect whether there are devices such as headphones or earmuffs in the pedestrian's head area that may obstruct hearing. Its purpose is to assess the pedestrian's ability to receive auditory warnings. Simultaneously, "identification of whether pedestrians are using visually distracting devices" refers to using similar algorithms to detect whether pedestrians are using devices such as mobile phones or tablets that may distract their visual attention, or whether their gaze has been deviated from the traffic environment for an extended period. Its purpose is to assess the pedestrian's ability to receive visual warnings and their current state of attention. Thus, the results of auditory-impairing device identification and visually distracting device identification can be obtained.

[0093] In practical applications, "adjusting the distribution method of warning information" specifically involves dynamically selecting or combining different warning methods based on the aforementioned visual feedback information, the results of auditory obstruction device identification, and the results of visual distraction device identification. For example, if it is identified that a pedestrian is wearing an auditory obstruction device, enhanced visual warnings (such as brighter warning lights with a higher flashing frequency) or tactile warnings (such as ground vibrations) can be prioritized; if it is identified that a pedestrian is using a visual distraction device, more intrusive visual warnings (such as warning information projected into the pedestrian's line of sight) or combined with auditory warnings (such as directional voice broadcasts) can be used. The aim is to ensure that the warning information reaches the target pedestrian in the most effective way.

[0094] "Activating warning devices" refers to controlling the corresponding warning devices (such as LED displays, ground projection lights, directional speakers, haptic feedback devices, etc.) to issue warnings according to the adjusted warning information distribution method.

[0095] In some preferred embodiments, a specific example is given below. Suppose at a traffic intersection, the system determines through perception data that a pedestrian is about to cross the red light and triggers a warning. At this point, the system first obtains the pedestrian's visual feedback to the warning via a high-definition camera deployed at the intersection. For example, the system analyzes the image using an image recognition algorithm and finds that the pedestrian is wearing in-ear headphones and their gaze is focused on the screen of their smartphone. Based on this, the system identifies that the pedestrian is wearing a hearing-impairing device and using a visually distracting device. Based on these identification results, the system adjusts the way the warning information is distributed. Specifically, the system may reduce the priority of traditional auditory warnings and instead activate the ground projection lights in the pedestrian crossing area, projecting a conspicuous red "Stop" sign directly onto the ground in front of the pedestrian, and simultaneously activate a directional voice broadcast system, issuing a "Do not cross the red light" voice warning to the pedestrian at a volume slightly higher than the ambient noise level. In this way, even if the pedestrian's auditory and visual attention are distracted, they can still effectively receive the warning information through more targeted visual and auditory warnings, thereby avoiding crossing the red light in time.

[0096] This application also discloses a pedestrian red-light violation warning system based on the Internet of Things for intelligent transportation, used to execute pedestrian red-light violation warnings based on the Internet of Things for intelligent transportation, combined with... Figure 3 As shown, the pedestrian red-light violation warning system 1 based on the intelligent transportation Internet of Things includes: The perception data acquisition module 11 is used to acquire multi-source local perception data at traffic intersections; Traffic light status judgment module 12 is used to judge the current status of traffic lights based on multi-source local sensing data and obtain the autonomous judgment result of traffic light status. The priority acceptance determination module 13 is used to acquire and monitor the status information of external traffic lights. When the update of the status information of external traffic lights is delayed or the status information of external traffic lights is inconsistent with the autonomous judgment result of the status of traffic lights, the autonomous judgment result of the status of traffic lights is given priority to obtain the priority traffic light status. The red light violation judgment module 14 is used to judge the pedestrian's red light violation behavior based on the priority traffic signal light status and the pedestrian movement trajectory prediction received in the pedestrian crossing area. The warning judgment trigger module 15 is used to trigger a warning based on the act of running a red light.

[0097] To better understand the system proposed in this application, the following provides a detailed description of each module involved and its implementation.

[0098] First, the system includes a sensing data acquisition module. The acquisition of multi-source local sensing data at traffic intersections has already been described in the above embodiments and will not be repeated here. It is important to emphasize that the sensing data acquisition module can be configured to acquire data in various ways. For example, this module can integrate multiple sensor interfaces for connecting to high-definition cameras, millimeter-wave radar, lidar, geomagnetic sensors, or pressure sensors, etc. In one implementation, the sensing data acquisition module can be a hardware interface unit responsible for collecting and initially processing raw analog or digital signals from different types of sensors, such as performing analog-to-digital conversion or data format encapsulation, and then transmitting them to subsequent processing modules. In another implementation, the sensing data acquisition module can be a software component running on an edge computing device, receiving sensing data streams from distributed sensors via a network interface and performing preliminary data cleaning and preprocessing.

[0099] Secondly, the system includes a traffic light status determination module. The above embodiments have already described how the current status of traffic lights is determined based on multi-source local sensing data, resulting in an autonomous determination of the traffic light status; this will not be repeated here. It is important to emphasize that the traffic light status determination module can be configured to perform complex analysis tasks. For example, this module can be an embedded processor running an image recognition algorithm to directly identify the color and status of traffic lights from video data. Alternatively, this module can be a rule-based expert system that indirectly infers the status of traffic lights by analyzing traffic flow data from radar or geomagnetic sensors. In some implementations, this module can be a separate computing unit dedicated to executing deep learning models to extract features from multi-source sensing data and determine the traffic light status.

[0100] Furthermore, the system includes a priority acceptance determination module. The above embodiments have already described the acquisition and monitoring of external traffic light status information. When the update of external traffic light status information is delayed or inconsistent with the autonomous judgment result of the traffic light status, the autonomous judgment result is prioritized, resulting in the prioritized traffic light status information, which will not be repeated here. It is important to emphasize that the priority acceptance determination module can be configured to implement intelligent decision-making logic. For example, this module can be a communication interface unit responsible for receiving external traffic light status information and includes a timer to monitor the information update delay. Simultaneously, this module can have a built-in comparator to compare the external information with the autonomous judgment result of the traffic light status. When a delay or inconsistency is detected, this module can select to accept the autonomous judgment result according to preset priority rules. In one implementation, this module can be a software service running on the central processing unit, responsible for coordinating the reception of external information, the acquisition of local judgment results, and the final decision output.

[0101] Next, the system includes a red-light violation judgment module. The above implementation has already described how to determine a pedestrian's red-light violation behavior based on the prioritized traffic light status and the received pedestrian trajectory predictions for the crosswalk area; this will not be repeated here. It is important to emphasize that the red-light violation judgment module can be configured to perform complex behavior analysis and prediction. For example, this module can be a high-performance computing unit running pedestrian trajectory prediction algorithms, such as Kalman filtering or deep learning-based prediction models, to predict the future position of pedestrians in the crosswalk area. Simultaneously, this module can include a rule engine that combines the prioritized traffic light status and the predicted pedestrian trajectory to determine whether red-light violation has occurred. In some implementations, this module can be a distributed processing unit capable of processing the movement data of multiple pedestrians in parallel and outputting the red-light violation judgment result in real time.

[0102] Finally, the system includes a warning judgment trigger module. The content of triggering warnings based on red-light running has already been described in the above embodiments and will not be repeated here. It is important to emphasize that the warning judgment trigger module can be configured to flexibly control various warning devices. For example, this module can be an output control unit, connected via wired or wireless means to warning devices such as voice broadcasting systems, LED displays, and ground projection equipment, and send corresponding activation commands based on the judgment result of red-light running. In one implementation, this module can be a programmable logic controller (PLC), which executes a preset warning strategy based on the received red-light running signal, such as adjusting the intensity or duration of the warning. In another implementation, this module can be a software application that sends control commands to remote warning devices via IoT protocols.

[0103] The pedestrian red-light violation warning system based on the Internet of Things for intelligent transportation proposed in this application has a core innovation: through modular design, it effectively combines local autonomous judgment of traffic light status with external information monitoring, and introduces a priority acceptance mechanism. Compared to traditional systems that rely solely on information from external traffic control centers, this system, through a sensing data acquisition module and a traffic light status judgment module, empowers the system to independently perceive and judge the status of traffic lights. Even in cases of delays or inconsistencies in external information transmission, the priority acceptance determination module can decisively accept the local autonomous judgment results, ensuring the timeliness and accuracy of the system's judgment of traffic light status. Therefore, the red-light violation judgment module can accurately judge the pedestrian's red-light violation behavior based on the most timely and accurate traffic light status, combined with pedestrian trajectory prediction, and the warning judgment triggering module will trigger the warning at the optimal time. This design effectively overcomes the warning lag problem caused by delays in external traffic light information in existing technologies, significantly improving the timeliness and effectiveness of warnings, thereby maximizing pedestrian traffic safety.

[0104] The above are merely embodiments of this application and are not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for warning pedestrians who run red lights based on the Internet of Things for intelligent transportation, characterized in that, include: Acquire multi-source local sensing data at traffic intersections; Based on the multi-source local sensing data, the current state of the traffic lights is determined, and the autonomous judgment result of the traffic light state is obtained. Acquire and monitor external traffic light status information. When the external traffic light status information is delayed or inconsistent with the autonomous judgment result of the traffic light status, the autonomous judgment result of the traffic light status is adopted first to obtain the traffic light status that is adopted first. Based on the priority traffic light status and combined with the pedestrian movement trajectory prediction received from the crosswalk area, the pedestrian's red light violation behavior is determined. A warning is triggered based on the described red-light violation.

2. The pedestrian red-light violation warning method based on intelligent transportation Internet of Things according to claim 1, characterized in that, The step of acquiring and monitoring external traffic light status information, and prioritizing the adoption of the traffic light status autonomous judgment result when the external traffic light status information is delayed or inconsistent with the external traffic light status information, includes: Obtain visual status information of vehicle signal lights; Obtain information on the actual traffic status of vehicles in pedestrian crossing areas; Based on the visual state information and the actual traffic status information of the vehicle, the safety status of the area crossed by the traveler is determined. When the update of the external traffic light status information is delayed or the external traffic light status information is inconsistent with the autonomous judgment result of the traffic light status, the safety status of the pedestrian crossing area is adopted first, and the traffic light status adopted first is obtained.

3. The pedestrian red-light violation warning method based on intelligent transportation Internet of Things according to claim 1, characterized in that, The step of determining a pedestrian's red-light violation based on the priority traffic light status and the received pedestrian movement trajectory prediction for the crosswalk area includes: Acquire individual pedestrian movement data in crosswalk areas; Acquire data on non-pedestrian moving objects in the pedestrian crossing area; Based on the individual pedestrian movement data and the non-pedestrian moving object data, calculate the distance between the pedestrian and vehicle travel paths, as well as the pedestrian's speed when crossing the boundary; Based on the individual pedestrian movement data, analyze the local crowd behavior patterns in the pedestrian crossing area; The pedestrian's red-light running behavior is determined based on the priority traffic signal status, the distance between the pedestrian and vehicle paths, the pedestrian's speed when crossing the boundary, and the behavior patterns of local crowds.

4. The pedestrian red-light violation warning method based on intelligent transportation Internet of Things according to claim 1, characterized in that, The step of determining a pedestrian's red-light violation based on the priority traffic light status and the received pedestrian movement trajectory prediction for the crosswalk area includes: Acquire data on non-pedestrian moving objects in the pedestrian crossing area; Based on the non-pedestrian moving object data, analyze the relative motion between pedestrians and non-pedestrian moving objects; Monitor the movement trajectory of pedestrians within the crosswalk area; Analyze the micro-behavioral characteristics of pedestrians; The pedestrian's red-light running behavior is determined based on the traffic light status, non-pedestrian moving object data, the relative motion between the pedestrian and the non-pedestrian moving object, the motion trajectory, and the micro-behavioral characteristics.

5. A method for warning pedestrians from running red lights based on intelligent transportation IoT as described in claim 1, characterized in that, The step of determining a pedestrian's red-light violation based on the priority traffic light status and the received pedestrian movement trajectory prediction for the crosswalk area includes: Acquire individual pedestrian motion data and pedestrian posture data; Based on the individual pedestrian motion data and the pedestrian posture data, analyze the relative distance and approach speed between the pedestrian and the pedestrian crossing boundary; Based on the individual pedestrian motion data, analyze the pedestrian's gait and speed patterns; Based on the pedestrian posture data, analyze the pedestrian's head orientation and gaze focus; Based on the individual pedestrian motion data and the pedestrian posture data, identify the non-crossing edge activities of pedestrians; The pedestrian's red-light running behavior is determined based on the traffic signal status, the relative distance and approach speed between the pedestrian and the crosswalk boundary, the pedestrian's gait pattern, the pedestrian's speed pattern, the pedestrian's head orientation, the pedestrian's line of sight focus, and the pedestrian's non-crossing edge activities.

6. A method for warning pedestrians from running red lights based on the Internet of Things for intelligent transportation, as described in claim 5, is characterized in that... The steps for determining a pedestrian's red-light running behavior based on the prioritized traffic signal status, the relative distance and approach speed between the pedestrian and the crosswalk boundary, the pedestrian's gait pattern, the pedestrian's speed pattern, the pedestrian's head orientation, the pedestrian's visual focus, and the pedestrian's non-crossing edge activity include: Acquire environmental sensor data; Based on the environmental sensor data, the confidence levels of the individual pedestrian motion data and the pedestrian posture data are adjusted to obtain the adjusted individual pedestrian motion data and the adjusted pedestrian posture data. Based on the adjusted individual pedestrian motion data, the pedestrian's gait and speed patterns are analyzed to obtain the latest analyzed gait and speed patterns; Based on the adjusted pedestrian posture data, the head orientation and gaze focus of pedestrians are analyzed to obtain the latest analyzed head orientation and gaze focus; Based on the adjusted individual pedestrian motion data and adjusted pedestrian posture data, the non-crossing purpose edge activities of pedestrians are identified, and the latest identified non-crossing purpose edge activities are obtained. The pedestrian's red-light running behavior is determined based on the priority traffic signal status, the relative distance and approach speed between the pedestrian and the crosswalk boundary, the latest analyzed gait and speed patterns, the latest analyzed head orientation and gaze focus, and the latest identified non-crossing edge activities.

7. A method for warning pedestrians from running red lights based on the Internet of Things for intelligent transportation, as described in claim 6, is characterized in that... The steps for determining a pedestrian's red-light running behavior based on the prioritized traffic light status, the relative distance and approach speed between the pedestrian and the crosswalk boundary, the latest analyzed gait and speed patterns, the latest analyzed head orientation and gaze focus, and the latest identified non-crossing edge activity include: Based on the environmental sensor data, adjust the confidence weights of the individual pedestrian motion data and the pedestrian posture data; Based on the individual pedestrian motion data and pedestrian posture data after adjusting the confidence weight, the corrected relative distance and approach speed between the pedestrian and the pedestrian crossing boundary are calculated. Based on the individual pedestrian motion data after adjusting the confidence weights, we analyze the pedestrian gait patterns and speed patterns to obtain the corrected gait patterns and speed patterns. Based on pedestrian posture data after adjusting confidence weights, we analyze pedestrian head orientation and visual focus to obtain corrected head orientation and visual focus. Based on the pedestrian individual motion data and pedestrian posture data after adjusting the confidence weight, the non-crossing target edge activity of pedestrians is identified, and the corrected non-crossing target edge activity is obtained. The weight of each judgment criterion in the red light violation judgment is adjusted based on the priority traffic signal status, the corrected relative distance and approach speed between pedestrians and crosswalk boundaries, the corrected gait and speed patterns, the corrected head orientation and gaze focus, the corrected non-crossing edge activity, and the real-time confidence of each judgment criterion. The pedestrian's red-light running behavior is determined based on the traffic light status that is given priority, the corrected relative distance and approach speed between the pedestrian and the crosswalk boundary, the corrected gait and speed patterns, the corrected head orientation and gaze focus, the corrected non-crossing edge activity, and the adjusted weights of each judgment criterion in the red-light running judgment.

8. A method for warning pedestrians from running red lights based on intelligent transportation IoT according to claim 1, characterized in that, The steps for triggering a warning based on the red-light violation include: Obtain the status information of the warning device; Based on the status information of the warning device and the red-light violation, the warning information distribution strategy is adjusted and executed. After executing the warning information distribution strategy, obtain visual feedback information of the warning area; Based on the visual feedback information of the warning area, the activation method of the warning device in the warning information distribution strategy is adjusted to obtain the revised warning information distribution strategy. Activate the warning device according to the revised warning information distribution strategy.

9. A method for warning pedestrians from running red lights based on intelligent transportation IoT according to claim 1, characterized in that, The steps for triggering a warning based on the red-light violation include: Obtain visual feedback information from pedestrians regarding the warning; Based on the visual feedback information, it is determined whether the pedestrian is wearing a hearing-impairing device and whether the pedestrian is using a visual distraction device, thus obtaining the hearing-impairing device identification result and the visual distraction device identification result. Based on the visual feedback information, the recognition results of the hearing impairment device, and the recognition results of the visual distraction device, the warning information distribution method is adjusted; Activate the warning devices according to the adjusted warning information distribution method.

10. A pedestrian red-light violation warning system based on intelligent transportation Internet of Things, used to execute pedestrian red-light violation warnings based on intelligent transportation Internet of Things, characterized in that, include: The perception data acquisition module is used to acquire multi-source local perception data at traffic intersections; The traffic light status judgment module is used to determine the current status of the traffic lights based on the multi-source local sensing data and obtain the autonomous judgment result of the traffic light status. The priority acceptance determination module is used to acquire and monitor the status information of external traffic lights. When the update of the external traffic light status information is delayed or the external traffic light status information is inconsistent with the autonomous judgment result of the traffic light status, the autonomous judgment result of the traffic light status is given priority to obtain the traffic light status that is given priority. The red light violation judgment module is used to judge the pedestrian's red light violation behavior based on the priority traffic signal status and the pedestrian movement trajectory prediction received in the pedestrian crossing area. The warning judgment trigger module is used to trigger a warning based on the red light violation.

Images