Pedestrian safety device

Abstract

The invention relates to the technical field of pedestrian safety protection, and discloses a pedestrian safety device which comprises a camera used for acquiring image information of a rear coming vehicle in real time; the radar is used for detecting the distance, the speed and the direction of a rear coming vehicle; the vibration device is used for providing a vibration warning for a wearer when the approaching of a rear vehicle is detected; the sound prompting device is used for providing a sound alarm for the wearer when the approaching of the rear vehicle is detected; and the narrowband communication module is used for sending alarm information to surrounding vehicles or traffic systems when the rear coming vehicle approaches. According to the invention, a multi-sensor fusion technology is adopted, and equipment such as a camera, a radar and a GPS are combined to obtain information of a rear coming vehicle in real time, so that high-precision collision risk prediction is achieved; compared with a scheme which only depends on a single sensor in the prior art, the method solves the problem that the error of the single sensor is large, and improves the stability and accuracy of the pedestrian safety device in a complex environment.

Description

Technical Field

[0001] The present invention relates to the technical field of pedestrian safety protection, and specifically to a pedestrian safety device. Background Art

[0002] Pedestrian safety devices are an important part of intelligent transportation systems, aiming to improve pedestrian safety in traffic by real-time monitoring of the traffic environment and timely warning of potential collision risks. With the increasing density of urban traffic, the traffic safety threats faced by pedestrians are constantly intensifying. To effectively prevent traffic accidents, especially collisions between vehicles and pedestrians, pedestrian safety devices have emerged. Modern pedestrian safety devices usually include various sensors and alarm mechanisms that can issue alarms when a vehicle approaches to remind pedestrians of the approaching vehicle.

[0003] Existing pedestrian safety devices all use a single sensor or a limited combination of multiple sensors in actual applications to achieve collision warning, and judge whether there is a collision risk by calculating the relative distance between the approaching vehicle and the pedestrian. When the system determines that the collision risk is high, it usually issues a warning through a vibration device or a sound alarm device to remind pedestrians of the approaching vehicles around.

[0004] However, the existing pedestrian safety devices have a lack of flexibility in their fixed alarm triggering mechanisms. They judge whether to issue an alarm based on fixed distance thresholds and speed thresholds. This method cannot be dynamically adjusted according to the specific traffic environment or the actual motion state of the vehicle, lacks dynamic adjustment in different situations, fails to optimize the alarm timing and intensity according to environmental changes, and lacks robustness to uncertain factors, making the system may not be able to make accurate judgments in the presence of noise or sensor errors, affecting the coordinated response between vehicles and pedestrians. The present invention provides a pedestrian safety device to solve the deficiencies existing in the prior art. Summary of the Invention

[0005] In view of the deficiencies of the prior art, the present invention provides a pedestrian safety device, which solves the problems in the prior art of lacking dynamic adjustment in different situations, failing to optimize the alarm timing and intensity according to environmental changes, and lacking robustness to uncertain factors, making the system may not be able to make accurate judgments in the presence of noise or sensor errors.

[0006] To achieve the above objectives, the present invention is realized through the following technical solutions: A pedestrian safety device, characterized by comprising: A camera for real-time acquisition of image information of the approaching vehicle from behind; A radar for detecting the distance, speed and direction of the approaching vehicle from behind; A vibration device for providing a vibration warning to the wearer when it detects the approach of the vehicle from behind; A sound alert device for providing a sound warning to the wearer when a vehicle approaching from behind is detected; A narrowband communication module for sending alert messages to surrounding vehicles and traffic systems when a vehicle approaching from behind is detected; A processing unit for receiving input information from a camera and a radar sensor, analyzing it, and triggering alert and control instructions.

[0007] Preferably, the camera and the radar work in a collaborative manner. Based on the movement trajectory and speed of the vehicle, the collision risk of the vehicle approaching from behind is judged through the fusion of image recognition and radar data.

[0008] Preferably, the output signal of the radar is used to assist the image judgment of the camera. The processing unit uses radar data and image data to predict the future movement trajectory of the approaching vehicle. The predicted trajectory is calculated based on the movement state model of the vehicle, and this movement state model satisfies the following movement equations: ; ; Wherein, is the tangential displacement along the trajectory, is the offset perpendicular to the trajectory, is the speed of the approaching vehicle, is the yaw angle of the approaching vehicle, is the time variable.

[0009] Preferably, the narrowband communication module is used for information exchange with nearby vehicles and traffic management systems. The narrowband communication module sends alert messages to nearby vehicles, and the vehicles make active braking and avoidance decisions based on the received alert messages.

[0010] Preferably, the processing unit calculates the optimal alert trigger moment based on dynamic game theory and issues an appropriate alert signal according to the calculation result. The game model includes the following cost functions: ; ; Wherein, is the safety cost function of the pedestrian, is the current position of the vehicle, is the current position of the pedestrian, is the control strategy adopted by the pedestrian, is the weight coefficient, is the safety cost function of the vehicle, is the speed of the vehicle, is the speed of the pedestrian, is the control strategy of the vehicle, is the weight coefficient, is the weight coefficient.

[0011] Preferably, the processing unit determines whether to trigger an alarm based on the distance between the future trajectory of the vehicle and the current position of the pedestrian. The future trajectory prediction uses the geodesic method based on Riemannian geometry, and the geodesic equation is: ; where represents the acceleration of the vehicle at time , is the Christoffel symbol, and respectively represent the speeds of the vehicle in the and directions.

[0012] Preferably, the processing unit optimizes the response of the vehicle to the pedestrian safety alarm through robust control, and the optimization process is realized by the following control law: ; where is the optimal control gain matrix, is the state vector of the vehicle, is the optimal control input of the vehicle.

[0013] Preferably, when the alarm is triggered, the vibration device and the sound prompting device adjust the working intensity and prompting method based on the relative distance, speed and direction of the vehicle coming from behind, and adjust the vibration intensity emitted by the vibration device and the prompting volume of the sound according to the approaching speed of the vehicle.

[0014] Preferably, the narrowband communication module performs two-way communication with the vehicle through wireless communication technology, and based on the position information of the person, sends alarm information to the vehicle in motion through the narrowband communication module. The alarm information includes the current position of the pedestrian, the speed of the vehicle coming from behind, and the direction information.

[0015] Preferably, the pedestrian safety device composed of the camera, radar and narrowband communication module further includes a GPS positioning module. The GPS positioning module can obtain the geographical location information of the wearer in real time, and through networking with the camera, radar and narrowband communication module, realizes the sharing of safety information and early warning within the area.

[0016] The present invention provides a pedestrian safety device, which has the following beneficial effects: 1. The present invention adopts multi-sensor fusion technology, combines devices such as cameras, radars and GPS to obtain information about the vehicle coming from behind in real time, achieving high-precision collision risk prediction; compared with the prior art that only relies on a single sensor, the present invention solves the problem of large errors in a single sensor and improves the stability and accuracy of the pedestrian safety device in a complex environment.

[0017] 2. The present invention optimizes the alarm triggering timing through dynamic game, and the system can calculate the optimal alarm moment according to the relative motion state of pedestrians and vehicles; compared with the prior art that uses fixed triggering conditions, the present invention effectively reduces the situation of triggering alarms too early or too late, ensures that the alarm can timely and accurately remind pedestrians, thereby improving safety.

[0018] 3. The present invention adopts robust control technology to optimize the reaction process of the vehicle, ensuring that the vehicle can still make accurate responses even under the interference of communication delays or system noises; compared with the control methods in the prior art that lack adaptability to external disturbances, the present invention provides more stable performance and ensures the reliability of the system in an uncertain environment.

[0019] 4. The present invention continuously adjusts the alarm strategy through an intelligent optimization algorithm, and the system can optimize the risk assessment in real time in a changing traffic environment; different from the prior art that adheres to static alarm schemes, the present invention solves the disadvantage of poor adaptability of traditional methods to dynamic traffic environments, can automatically adjust the alarm intensity and timing according to different traffic conditions, and ensures the continuous safety of pedestrians. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is the architecture diagram of the pedestrian safety device of the present invention; Figure 2 It is the operation step diagram of the pedestrian safety device of the present invention; DETAILED DESCRIPTION OF THE INVENTION

[0021] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0022] Please refer to the attached Figure 1 , the present invention provides a pedestrian safety device, including: A camera for real-time acquisition of image information of oncoming vehicles from behind; A radar for detecting the distance, speed and direction of oncoming vehicles from behind; A vibration device for providing a vibration warning to the wearer when it detects that an oncoming vehicle is approaching; A sound prompting device for providing a sound warning to the wearer when it detects that an oncoming vehicle is approaching; A narrowband communication module for sending alarm information to surrounding vehicles or traffic systems when an oncoming vehicle is approaching; A processing unit, configured to receive input information from a camera and a radar sensor, analyze the information, and trigger an alarm or a control instruction.

[0023] For the camera, in this embodiment, the camera obtains real-time image information of the vehicle coming from behind in real time and processes the information to facilitate subsequent determination of the distance, speed, direction of the vehicle, and whether there is a collision risk. This step, through cooperation with other sensor modules such as the radar sensor, provides dynamic perception between the vehicle and pedestrians, providing key data support for the subsequent alarm triggering of the system. By collecting and analyzing the image information of the vehicle coming from behind, the system can timely identify its motion state and potential danger when the vehicle approaches.

[0024] Specifically, the camera is mainly responsible for image acquisition of the vehicle coming from behind and can obtain real-time image data of the rear environment at a certain sampling frequency. The camera uses a common imaging technology (such as a CMOS image sensor) to capture the environment through a built-in image sensor. Generally, the camera can work normally under all-weather and various lighting conditions and has strong image recognition capabilities.

[0025] Generally, the camera captures an image in real time and transmits it to the processing unit, and the processing unit further performs image analysis to extract key features of the vehicle coming from behind from the image. The image processing module usually includes technologies such as license plate recognition, head / tail recognition of the vehicle, and detection of the size and contour of the vehicle. As an option, in order to improve the accuracy of image analysis, deep learning algorithms can also be combined for image classification and object detection.

[0026] During the process of image acquisition, the resolution and frame rate of the camera are key factors affecting its accuracy and response speed. Specifically, for a vehicle traveling at high speed, the sampling frequency of the camera should be high enough to ensure that complete and clear vehicle information can be obtained under fast motion conditions, reducing the impact of blurred images.

[0027] In a possible implementation manner, the image collected by the camera will be subjected to image analysis by the processing unit, with the goal of extracting key information such as the motion trajectory, speed, and direction of the vehicle coming from behind from the image. Image analysis mainly relies on image recognition technologies, especially the convolutional neural network (CNN) model based on AI technology, for vehicle detection and recognition.

[0028] During the image recognition process, the system first preprocesses the image, removes noise, and filters the image. Then, through methods such as edge detection and contour extraction, it identifies the areas in the image that may belong to vehicles. On this basis, by extracting features such as the front of the vehicle, the rear of the vehicle, and the license plate, the system further identifies whether there is an oncoming vehicle in the image. In some embodiments, the processing unit determines the motion state of the vehicle, including the motion trajectory and speed of the vehicle, by calculating the contour and direction of the target object in the image.

[0029] In some embodiments, by combining radar data (such as distance, speed, etc. information), the error of image recognition can be further corrected, and the recognition accuracy of vehicle dynamics can be improved. Through the fusion of image and radar data, the system can accurately obtain the speed, distance, and driving direction of the vehicle.

[0030] Specifically, assume that the lateral position of the vehicle image recognized by the camera at a certain moment is , the longitudinal position is , and the yaw angle of the vehicle is . Then, through the fusion of image recognition and the position sensor, the system can estimate the motion trajectory of the vehicle , and its motion equation can be approximated as: ; ; where is the initial position of the vehicle, is the vehicle speed, is the real-time direction angle of the vehicle, is the time. According to the above equations, the camera can estimate the future trajectory of the vehicle within a given time, and the processing unit uses these data to determine whether there is a collision risk.

[0031] As an option, the pedestrian safety device of the present invention can combine the image data obtained by the camera with the distance, speed, etc. information provided by the radar sensor through data fusion technology. The image data can be used to confirm the position, contour, etc. of the oncoming vehicle, while the radar data provides more accurate distance and speed information. Through multi-sensor data fusion, the system can improve the prediction accuracy of the motion trajectory of the oncoming vehicle, thereby enhancing the ability to judge the collision risk.

[0032] Based on the fused data, the system calculates the relative position and speed between the oncoming vehicle and the pedestrian to predict the future collision risk. If the system determines that the motion trajectory of the oncoming vehicle will intersect with the pedestrian, and the time distance at the intersection is short, the system will trigger an alarm. At this time, the processing unit will send control signals to the vibration device and the sound prompt device for warning; at the same time, the narrowband communication module will transmit the alarm information to nearby vehicles to remind the driver to take evasive or braking operations.

[0033] Specifically, based on the oncoming vehicle trajectory calculated by the processing unit, the system will give an early warning and select an appropriate time to issue an alarm according to multi-dimensional information such as distance, speed, and direction, reducing the probability of accidents.

[0034] For the radar, by analyzing and predicting the detected oncoming vehicle data, the movement trajectory of the oncoming vehicle is calculated, and it is judged whether there is a collision risk according to the calculation results. Through accurate trajectory prediction, the system can evaluate the relative distance, speed, and collision possibility between the pedestrian and the oncoming vehicle, so as to decide whether to trigger an alarm.

[0035] In this embodiment, the prediction of the movement trajectory adopts a calculation method based on the vehicle motion state model. The camera provides the image information of the oncoming vehicle, while the radar provides the accurate distance, speed, and direction data of the vehicle. The processing unit calculates the future movement trajectory of the vehicle by fusing the data from different sensors. Through this data fusion technology, the error caused by a single sensor can be effectively reduced, and a more accurate prediction result can be provided.

[0036] Generally, the movement trajectory of the oncoming vehicle can be predicted through a basic physical motion model. Assume the initial position of the vehicle is , the speed is , the yaw angle is , then the tangential displacement along the path and the offset perpendicular to the trajectory can be calculated respectively by the following formulas: ; ; where, is the tangential displacement along the trajectory, is the offset perpendicular to the trajectory, is the speed of the oncoming vehicle, is the yaw angle of the oncoming vehicle, is the time variable.

[0037] Specifically, according to the motion state of the vehicle, the system can predict the future movement trajectory of the vehicle by measuring the position (x, y) and speed of the vehicle at a certain moment. Using the above formulas, the system can estimate the trajectory points of the vehicle within the future time , so as to obtain the future position and movement direction of the vehicle.

[0038] As an option, after estimating the trajectory of the oncoming vehicle, the system needs to determine whether the trajectory intersects with the current position of the pedestrian. If the predicted trajectory coincides with the position of the pedestrian, the system considers that there is a collision risk. To accurately judge this, the system calculates the relative distance between the oncoming vehicle and the pedestrian, and further predicts the collision time point and collision intensity based on this distance and data such as the speed and direction of the oncoming vehicle.

[0039] In a possible implementation, the system calculates the relative distance between the oncoming vehicle and the pedestrian to evaluate the collision risk. The calculation formula for the relative distance is as follows: ; Wherein, represents the relative distance between the pedestrian and the vehicle, and respectively represent the axis and axis coordinates of the vehicle at time axis coordinates, and respectively represent the axis and axis coordinates of the pedestrian at time axis coordinates, is the time variable.

[0040] In some embodiments, if the calculated relative distance is less than the preset safety distance and the speed of the vehicle exceeds a certain threshold , the system determines that there is a high-risk collision. At this time, the system immediately triggers the alarm mechanism, emits vibration and sound warnings, and transmits the alarm information to nearby vehicles through the narrowband communication module.

[0041] Specifically, to further improve the accuracy of collision risk judgment, the system also adopts a method based on dynamic game theory to dynamically calculate the optimal alarm trigger moment. By introducing game theory, the behaviors of pedestrians and vehicles are regarded as a decision-making process of mutual interaction. In this process, pedestrians need to react according to the behavior of the oncoming vehicle, and the vehicle will also make decisions according to the reaction of the pedestrian. The system evaluates the costs and benefits of triggering the alarm at different moments through a dynamic game model, so as to calculate the optimal alarm moment.

[0042] In the game model, the cost functions of pedestrians and vehicles respectively represent the safety distances and movement trajectories of pedestrians and vehicles: ; ; Wherein, is the safety cost function of the pedestrian, is the current position of the vehicle, is the current position of the pedestrian, is the control strategy adopted by the pedestrian, is the weight coefficient, is the safety cost function of the vehicle, is the speed of the vehicle, is the speed of the pedestrian, is the control strategy of the vehicle, is the weight coefficient, is the weight coefficient.

[0043] When dealing with collision risk assessment, the system also needs to take into account various uncertain factors, such as communication delay, environmental changes, sensor errors, etc. As an option, the system uses robust control methods to optimize the vehicle's response. Through the robust control algorithm, the system can ensure that when the vehicle receives an alarm, it can stably perform braking or avoidance operations, and still maintain a high response accuracy even in the presence of system disturbances or noises. The control law is as follows: ; where, is the optimal control gain matrix, is the state vector of the vehicle, is the optimal control input of the vehicle. Through robust control, the vehicle can accurately perform actions such as braking or decelerating according to the calculation results of the optimal control gain matrix, effectively avoiding collisions with pedestrians.

[0044] For the vibration device and the sound prompting device, in this embodiment, when the system detects that the relative distance between the oncoming vehicle and the pedestrian is less than the safety threshold and there is a collision risk, the alarm signal will be immediately triggered. The intensity and type (vibration, sound) of the alarm signal will be dynamically adjusted according to the speed, distance of the oncoming vehicle, and the system's assessment of the collision risk.

[0045] Generally, when the oncoming vehicle is close to the pedestrian and has a high speed, the system will select strong vibration prompts and high-frequency sound alarms to immediately warn the pedestrian to avoid. As an option, if the oncoming vehicle has a low speed or is far away, the alarm intensity will be correspondingly reduced to avoid excessive interference to the wearer.

[0046] Specifically, when the relative distance calculated by the system is less than the preset safety distance and the speed of the oncoming vehicle exceeds a certain threshold , the system will consider the collision risk to be high and immediately trigger high-intensity vibration and sound alarms. On the contrary, if the relative distance is far or the oncoming vehicle speed is slow, the system will adopt a milder alarm method.

[0047] In terms of the formula, assume the relative speed between the pedestrian and the oncoming vehicle is , and the relative distance is . Then the alarm trigger condition can be expressed as: and ; where represents the relative distance between the pedestrian and the vehicle, is the safety distance threshold, represents the relative speed between the pedestrian and the vehicle, is the threshold of the relative speed. When the condition is met, the alarm will be triggered, and the system will remind the wearer through a vibration device and a sound prompt device.

[0048] In a possible implementation, the system not only reminds the pedestrian through vibration and sound prompts, but also triggers corresponding control mechanisms based on data such as the speed and distance of the oncoming vehicle to ensure that the pedestrian can take evasive measures as soon as possible when the collision risk is high. If the system determines that a collision is inevitable, it will issue an emergency alarm to prompt the pedestrian to stop or avoid immediately.

[0049] To further optimize the reaction, the system uses dynamic game theory for behavior prediction and decision-making. The interaction between the pedestrian and the vehicle can be regarded as a dynamic game, and the decisions of both sides are affected by the behavior of the other side. The system calculates the optimal alarm trigger timing through a game model to ensure that the reactions of the pedestrian and the vehicle are the most efficient.

[0050] Through game theory, the system will calculate the time point when the alarm is triggered and optimize the intensity and manner of the reaction based on that moment. The cost function of the game model is as follows: ; ; where is the safety cost function of the pedestrian, is the current position of the vehicle, is the current position of the pedestrian, is the control strategy taken by the pedestrian, is the weight coefficient, is the safety cost function of the vehicle, is the speed of the vehicle, is the speed of the pedestrian, is the control strategy of the vehicle, is the weight coefficient, is the weight coefficient.

[0051] As an option, the system also applies robust control technology to optimize the vehicle's response. The purpose of robust control is to ensure that the vehicle can still make a correct response after receiving an alert when there are uncertainties in the system, such as sensor errors, communication delays, etc. The control law is as follows: ; where, is the optimal control gain matrix, is the state vector of the vehicle, is the optimal control input of the vehicle.

[0052] Through the optimal control gain matrix , the system can calculate the optimal braking force or acceleration force to ensure that the vehicle can make a quick and accurate response in an emergency. In some embodiments, robust control enables the vehicle to make a correct avoidance action according to the optimal control law even when there is a certain delay in the communication between the vehicle and the pedestrian safety device.

[0053] For the narrowband communication module, in this embodiment, the transmission process of the alert information is implemented through the narrowband communication module. The role of this communication module is to transmit the alert information to nearby vehicles or the traffic management system through wireless communication technology when the pedestrian safety device detects an oncoming vehicle and triggers an alert. In this way, the vehicle can receive the alert from the pedestrian safety device in real time, ensuring that the vehicle can take timely action to avoid colliding with pedestrians.

[0054] Generally, the narrowband communication module exchanges information with surrounding vehicles or the traffic management system through communication protocols such as low-power wide-area network (LPWAN). The advantage of this wireless communication protocol lies in its low energy consumption and long communication distance, which is suitable for real-time information transmission between vehicles and pedestrian safety devices.

[0055] As an option, when the alert information is transmitted to the vehicle, the receiving module of the vehicle can receive the warning signal from the pedestrian safety device, which contains key information such as the current position of the pedestrian, the speed and direction of the oncoming vehicle, etc. The vehicle decides whether to take safety measures such as active braking, deceleration or avoidance based on the received information and its own real-time data. At this time, the vehicle's autonomous driving system or the driver's active response will be triggered, thus avoiding potential traffic accidents.

[0056] In specific implementations, the format and content of the alert information need to be standardized to ensure that vehicles and the traffic management system can accurately receive and understand the signal. The alert information includes but is not limited to the following: The current geographical location coordinates of the pedestrian, usually obtained through the GPS module; including information such as the current speed and movement direction (angle) of the oncoming vehicle; Based on the predicted motion trajectory, the relative positions between pedestrians and vehicles, and the calculated collision risks, the system provides suggestions on whether to trigger braking, avoidance, etc.

[0057] In a possible implementation, the alert information is not only transmitted to the vehicle, but also can be transmitted to the traffic management system. The traffic management system can, based on the received information, comprehensively judge the behaviors of multiple pedestrians and vehicles through the central processing unit. If multiple collision risk events occur in a certain area, the system can initiate a global safety warning and guide or adjust the traffic flow in this area in advance.

[0058] In some embodiments, the traffic management system can calculate high-risk areas based on the information summary of multiple pedestrian safety devices, and issue relevant warnings to drivers through traffic lights or road signs, etc. The system can also adjust the timing of traffic lights according to the real-time traffic conditions to ensure the safer flow of pedestrians and vehicles.

[0059] As an option, when transmitting the alert information, the system not only sends signals unidirectionally, but also includes a feedback mechanism. When the vehicle receives the alert information, if the on-vehicle system (such as the AEB automatic braking system) is activated and takes actions, the vehicle can send a confirmation signal to the pedestrian safety device, feedbacking that it has activated the braking or avoidance behavior, so that the pedestrian safety device can make dynamic adjustments according to the feedback information of the vehicle and make a more intelligent reaction according to the current traffic environment.

[0060] For the processing unit, in this embodiment, the system continuously monitors and analyzes data from various sensors such as cameras, radars, and GPS positioning to judge whether there is still a collision risk. Generally, when the pedestrian safety device issues an alert, the system does not immediately stop monitoring the environment. Instead, the system will continue to evaluate the speed, direction, position of the vehicle behind and the movement state of the pedestrian itself to confirm whether it is necessary to update the alert information or perform other operations.

[0061] Specifically, after the initial alert, the system will regularly obtain real-time data from cameras and radars, and update and analyze this data through the processing unit. If the system judges that the relative position between the oncoming vehicle and the pedestrian has changed, or the speed and direction of the oncoming vehicle have changed, the processing unit will recalculate the future collision risk and decide whether it is necessary to re-trigger the alert or adjust the intensity of the alert.

[0062] As an option, when the distance between the pedestrian and the vehicle gradually shortens or the speed of the oncoming vehicle suddenly increases, the system will adjust the alarm intensity in real time through a feedback mechanism. If it detects that the oncoming vehicle is moving too fast and the approaching distance is short, the system can increase the intensity of the alarm to ensure that the pedestrian is reminded in time. The monitoring and alarm triggering in this process are dynamically changing and respond in real time according to the changes in the environment.

[0063] To improve the response ability and reliability of the system, this embodiment adopts an intelligent optimization algorithm. This optimization algorithm is based on the analysis of machine learning and historical data, and can predict the collision risk in different situations and make corresponding adjustments. In a possible implementation, the system will continuously optimize the prediction model according to the actual situation of the traffic environment. By continuously learning and optimizing, the system can improve the accuracy of collision risk prediction based on past experience data, and further optimize the alarm triggering timing and intensity.

[0064] For example, in some scenarios (such as heavy traffic, poor weather conditions, etc.), the system may perceive potential risks in advance and give an alarm in advance to avoid collisions caused by slow reaction within a short time. Through learning historical data, the system can intelligently adjust the alarm strategy to make the alarm more sensitive and effective.

[0065] In some embodiments, this optimization algorithm can also take into account other factors, such as the relative motion patterns of pedestrians and vehicles, road curvature, weather changes, etc. These factors may affect the accuracy of the prediction model. Therefore, the system will correct the prediction according to the real-time traffic conditions to provide a more accurate collision risk assessment.

[0066] During the continuous monitoring process, the system needs to continuously evaluate the relative position and speed between the oncoming vehicle and the pedestrian. To conduct a more accurate risk assessment, the system uses some kinematic formulas to describe the relative motion between the pedestrian and the oncoming vehicle. The relative distance can be calculated by the following formula: ; where represents the relative distance between the pedestrian and the vehicle, and respectively represent the -axis and -axis coordinates of the vehicle at time , and respectively represent the -axis and -axis coordinates of the pedestrian at time , is the time variable.

[0067] This formula can calculate the real-time relative distance between pedestrians and vehicles at each moment. Specifically, based on the relative speed between the vehicle and the pedestrian and the current relative distance , the system can predict the future time to collision , thereby determining the risk level of collision: ; Among them, is the relative distance, is the risk level of collision, is the relative speed between the vehicle and the pedestrian. If is less than the preset safety time , the system considers that there is a high risk of collision and will re-trigger the alarm.

[0068] In some embodiments, the system dynamically adjusts the triggering timing of the alarm according to this formula to ensure maximum protection for pedestrians in the shortest time. Through this continuous monitoring and real-time evaluation, the pedestrian safety device can cope with various complex traffic environments and ensure that pedestrians can obtain timely safety tips under different circumstances.

[0069] The operation process of the pedestrian safety device described below can be mutually referred to the pedestrian safety device described above.

[0070] Please refer to the appendix Figure 2 , in the embodiments of the present invention, the pedestrian safety device improves the traffic safety of pedestrians through real-time monitoring, prediction and alarm mechanisms. The overall process combines various sensor data (such as cameras, radars, GPS, etc.) with intelligent algorithms, can real-time evaluate the collision risk between pedestrians and vehicles, and take appropriate response measures, including the following operation process: For step S1, collect, obtain and analyze the image information of the oncoming vehicle from behind. The camera is responsible for obtaining the visual data of the traffic environment behind, using image recognition technology to identify the features of the oncoming vehicle (such as the front of the vehicle, the rear of the vehicle, the license plate, etc.). Through this image data, the system can real-time obtain information such as the speed and direction of the oncoming vehicle, providing basic data for subsequent trajectory prediction and collision risk assessment. This step is the basis of the entire system to ensure the accurate acquisition of information about the oncoming vehicle from behind.

[0071] For step S2, the prediction of the moving trajectory of the vehicle coming from behind and the judgment of the collision risk. Based on the acquired image information and radar data, the system predicts the moving trajectory of the oncoming vehicle. By fusing image recognition and radar data (such as distance, speed, etc.), the system calculates the future moving trajectory of the oncoming vehicle. The system uses a physical motion model and a geodesic method based on Riemannian geometry to predict the driving path of the vehicle, and evaluates the relative position and speed between the pedestrian and the oncoming vehicle. According to the predicted trajectory, the system determines whether there is a collision risk and calculates the likelihood of a collision occurring. If it is calculated that there is an intersection between the trajectories of the pedestrian and the oncoming vehicle, the system further determines the risk level of the collision.

[0072] For step S3, alarm triggering and reaction control. After determining that there is a collision risk, it is responsible for triggering and controlling the reaction of the alarm signal. When the system calculates that the relative distance between the vehicle and the pedestrian is less than the safe distance and the collision risk is high, the system will trigger the vibration device and the sound prompt device to warn the pedestrian to pay attention to the vehicle behind. The intensity and type (vibration, sound) of the alarm will be adjusted according to the speed, distance and risk level of the oncoming vehicle. In addition, the system also uses dynamic game theory to calculate the optimal alarm triggering time to ensure that the alarm is triggered at the most appropriate time and avoid premature or late reactions.

[0073] For step S4, alarm signal transmission and intelligent control feedback. In the process of transmitting the alarm signal to surrounding vehicles and the traffic management system, when the system detects a collision risk, the narrowband communication module transmits the alarm information to nearby vehicles through wireless communication technology to remind the driver to avoid or brake. The alarm information includes data such as the current position of the pedestrian, the speed and direction of the oncoming vehicle, providing real-time basis for the reaction of the vehicle. In addition, the traffic management system can receive this information and make adjustments through the traffic signal system to provide more safety guarantees for the vehicle.

[0074] For step S5, continuous monitoring of collision risk and system optimization. It is responsible for continuously optimizing and monitoring the system. After the initial alarm, the system will still continuously track the moving trajectory of the vehicle coming from behind and update the collision risk assessment in real time according to the new sensor data. If the speed of the oncoming vehicle changes or other factors are affected, the system will re-evaluate the collision risk and decide whether to re-trigger the alarm or adjust the alarm intensity. Through intelligent algorithms and machine learning, the system will also optimize the alarm strategy to ensure that in a complex traffic environment, the pedestrian safety device can respond flexibly according to real-time data and minimize the probability of a collision occurring to the greatest extent.

[0075] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A pedestrian safety device, characterized in that: include: Camera, used to obtain image information of vehicles coming from behind in real time; Radar, used to detect the distance, speed and direction of vehicles coming from behind; A vibration device for providing a vibration warning to the wearer when a vehicle approaching from behind is detected; An audio warning device is used to provide an audio warning to the wearer when a vehicle approaching from behind is detected; A narrowband communication module for sending warning messages to surrounding vehicles and traffic systems when a vehicle is approaching from behind; The processing unit is used to receive input information from the camera and radar sensors, analyze it and trigger alarms and control instructions.

2. The pedestrian safety device according to claim 1, characterized in that: The camera and radar work in a collaborative manner to determine the risk of collision with a vehicle coming from behind based on the vehicle's motion trajectory and speed through the fusion of image recognition and radar data.

3. The pedestrian safety device according to claim 1, characterized in that: The output signal of the radar is used to assist the camera in image judgment. The processing unit uses the radar data and the image data to predict the future motion trajectory of the vehicle. The predicted trajectory is calculated based on the vehicle's motion state model, which satisfies the following motion equation: ; ; in, is the tangential displacement along the trajectory, is the vertical trajectory offset, is the speed of the oncoming vehicle, is the yaw angle of the oncoming vehicle, is the time variable.

4. The pedestrian safety device according to claim 1, characterized in that: The narrowband communication module is used to exchange information with nearby vehicles and traffic management systems. The narrowband communication module sends alarm information to nearby vehicles, and the vehicles make active braking and avoidance decisions based on the received alarm information.

5. The pedestrian safety device according to claim 1, characterized in that: The processing unit calculates the optimal alarm triggering time based on dynamic game theory and issues an appropriate alarm signal according to the calculation result. The game model includes the following cost function: ; ; in, is the pedestrian safety cost function, is the current position of the vehicle, is the current position of the pedestrian, Control strategies for pedestrians, is the weight coefficient, is the safety cost function of the vehicle, is the speed of the vehicle, is the speed of pedestrians, For the vehicle control strategy, is the weight coefficient, is the weight coefficient.

6. The pedestrian safety device according to claim 1, characterized in that: The processing unit determines whether to trigger an alarm based on the distance between the future trajectory of the vehicle and the current position of the pedestrian. The future trajectory prediction adopts a geodesic method based on Riemannian geometry. The geodesic equation is: ; in, Indicates at time The acceleration of the vehicle, is the Christofel symbol, and Respectively indicate that the vehicle is and Speed ​​in direction.

7. The pedestrian safety device according to claim 1, characterized in that: The processing unit optimizes the vehicle's response to the pedestrian safety alert through robust control, and the optimization process is implemented through the following control law: ; in, is the optimal control gain matrix, is the state vector of the vehicle, is the optimal control input for the vehicle.

8. The pedestrian safety device according to claim 1, characterized in that: When the alarm is triggered, the vibration device and the sound prompt device adjust the working intensity and prompt mode based on the relative distance, speed and direction of the vehicle coming from behind. According to the speed of the approaching vehicle, the vibration intensity emitted by the vibration device is adjusted and the sound prompt volume is adjusted.

9. The pedestrian safety device according to claim 1, characterized in that: The narrowband communication module performs two-way communication with the vehicle through wireless communication technology. Based on the location information of the person, the narrowband communication module sends alarm information to the moving vehicle. The alarm information includes the current location of the pedestrian, the speed of the oncoming vehicle, and the direction information.

10. The pedestrian safety device according to claim 1, characterized in that: The pedestrian safety device composed of the camera, radar and narrowband communication module further includes a GPS positioning module, which can obtain the wearer's geographic location information in real time, and realize safety information sharing and early warning in the area by networking with the camera, radar and narrowband communication module.

Images