A method and system for anti-collision warning when a vehicle door is opened

Abstract

The present invention belongs to the technical field of vehicle safety, and discloses a method and system for anti-collision warning when a vehicle door is opened; it includes: collecting intention analysis data; analyzing the intention analysis data to identify the door-opening intention of a passenger; obtaining vehicle characteristic data; predicting a driving trajectory according to the vehicle characteristic data; analyzing the driving trajectory to evaluate a collision risk; giving corresponding warnings to the passenger according to the evaluation result of the collision risk and the door-opening intention; the present invention can accurately identify the door-opening risk and timely remind the vehicle passengers, reduce the accident risk caused by vehicle blind spots, improve the safety of vehicle door opening, and effectively guarantee the traffic safety of pedestrians and cyclists.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle safety, and more specifically, to a method and system for anti-collision warning when a vehicle door is opened. Background Art

[0002] With the rapid growth of the number of automobiles, the urban road traffic environment has become increasingly complex. The potential safety hazards in the blind areas at the rear side of a vehicle during low-speed driving or parking have gradually emerged. In particular, "door-opening collisions" caused by opening vehicle doors occur frequently, seriously threatening the lives of pedestrians and cyclists. Traditional rearview mirror observations are difficult to fully cover the blind areas at the rear side, and the driver's judgment is easily restricted by the environment and visual blind areas, resulting in uncontrollable risks. Therefore, there is an urgent need for an intelligent vehicle door opening warning system to effectively improve the safety when a vehicle door is opened.

[0003] The patent application with the publication number CN116238416A discloses a method, device, equipment and medium for anti-collision warning when a vehicle door is opened, including: obtaining the perception information around the current vehicle, where the perception information at least includes image information and ultrasonic echo information; respectively identifying the target object types and distances in the image information and the ultrasonic echo information; fusing the target object types and distances of the image information and the ultrasonic echo information to obtain the target object types and target object distances of each; determining whether the current vehicle meets the door opening condition according to the target object types and target object distances of each. If the current vehicle meets the door opening condition, the vehicle door is opened; if the current vehicle does not meet the door opening condition, the vehicle door is locked and an anti-collision warning is issued. This application avoids the occurrence of collision accidents caused by human negligence and realizes the anti-collision warning for manual or automatic opening of the vehicle door.

[0004] Although the above technology can realize the anti-collision warning when a vehicle door is opened, it only roughly describes the recognition results by combining the radar data, image data, and ultrasonic echo data in three aspects to identify the obstacles with collision risks around the vehicle, and does not elaborate on the specific technical solutions. As a result, in actual applications, the ability to identify obstacles in complex environments is insufficient, and it is unable to effectively identify the obstacles with collision risks around the vehicle, making it impossible for the driver to obtain the necessary safety information in a timely manner when opening the door, thus affecting the accuracy and reliability of the warning, increasing the risk of accidents, and unable to fully guarantee the safety of pedestrians and cyclists.

[0005] In view of this, the present invention proposes a method and system for anti-collision warning when a vehicle door is opened to solve the above problems. Summary of the Invention

[0006] In order to overcome the above-mentioned defects of the prior art and achieve the above object, the present invention provides the following technical solution: A method for anti-collision warning when a vehicle door is opened, including:

[0007] Collect intention analysis data;

[0008] Analyze the intention analysis data to identify the passenger's door opening intention;

[0009] Obtain vehicle characteristic data;

[0010] Predict the driving trajectory based on the vehicle characteristic data;

[0011] Analyze the driving trajectory to evaluate the collision risk;

[0012] Give corresponding warnings to the passengers according to the evaluation result of the collision risk and the door opening intention.

[0013] Furthermore, the intention analysis data includes vehicle speed and door distance; the door distance is the distance between the passenger and the inner side of the door;

[0014] The method for identifying the passenger's door opening intention includes:

[0015] Preset intention thresholds, where the intention thresholds include a vehicle speed threshold and a vehicle distance threshold; compare the vehicle speed with the vehicle speed threshold, and compare the door distance with the vehicle distance threshold; if the vehicle speed is less than the vehicle speed threshold and the door distance is less than the vehicle distance threshold, then generate a door opening intention instruction; if the vehicle speed is greater than or equal to the vehicle threshold, or the door distance is greater than or equal to the vehicle distance threshold, then no door opening intention instruction is generated;

[0016] The vehicle characteristic data includes a target distance and a relative speed; the target distance is the distance between the vehicle and a moving obstacle, and the relative speed is the moving speed of the moving obstacle relative to the vehicle.

[0017] Furthermore, the steps for predicting the driving trajectory include:

[0018] Step S101: Take the vehicle characteristic data corresponding to the same moving obstacle as a set of data, and each set of data corresponds to a moving obstacle one by one;

[0019] Step S102: Predict the relative speed of each moving obstacle at a future moment according to the relative speed in each set of data, and mark it as the predicted speed, and add each predicted speed to the corresponding set of data;

[0020] Step S103: Predict the target distance of each moving obstacle at a future moment according to the predicted speed and the target distance in each set of data, and mark it as the predicted distance, and add each predicted distance to the corresponding set of data;

[0021] Step S104: Based on the predicted distance in each set of data, construct the driving trajectory corresponding to each moving obstacle.

[0022] Further, in the step S102, the method for predicting the relative speed of each moving obstacle at a future moment includes:

[0023] Regarding the relative speed in each data set as a speed set, with the speed sets corresponding one-to-one to the data sets; inputting each speed set into the trained speed prediction model to predict the relative speed of each moving obstacle at a future moment, where the speed prediction model is an RNN neural network model; the training method of the speed prediction model includes:

[0024] Pre-collect consecutive relative speeds and construct a speed training set, where is an integer greater than 1; based on the speed training set, train a speed prediction model for predicting the relative speed at a future moment;

[0025] Preset a sliding step length L and a sliding window length D; use the sliding window method to convert the relative speeds in the speed training set into multiple training samples, with each training sample including D relative speeds; take each training sample as the input of the speed prediction model, predict the relative speed after the sliding step length L as the output, use the last L relative speeds of each training sample as the prediction target, evaluate the model accuracy using the mean absolute percentage error MAPE for the prediction result, and when the calculated MAPE is less than the preset MAPE, the speed prediction model corresponding to the relative speed is trained; generate a speed prediction model for predicting the relative speed at a future moment based on the relative speed.

[0026] Further, in the step S103, the method for predicting the target distance of each moving obstacle at a future moment includes:

[0027] Obtain the acquisition moment corresponding to each set of vehicle feature data, sort the acquisition moments corresponding to the vehicle feature data in the same data set from the latest to the earliest, and obtain a moment sorting table corresponding to each data set; take the vehicle feature data corresponding to the acquisition moment ranked first in each moment sorting table as the real-time feature data of the corresponding data set; input each set of real-time feature data into the trained distance prediction model to predict the target distance at a future moment; replace the target distance in each set of real-time feature data with the target distance at a corresponding future moment, and input the replaced real-time feature data into the trained distance prediction model to predict the target distance at two future moments; and so on, and predict the target distance of each moving obstacle at a future moment in total;

[0028] The training process of the distance prediction model includes:

[0029] Pre-collect a set of real-time feature data, for a corresponding future distance is set for each set of real-time feature data, and the future distance is the target distance at a future moment. is an integer greater than 1. The real-time feature data and the corresponding future distance are converted into a corresponding set of feature vectors; each set of feature vectors is used as the input of the distance prediction model. The distance prediction model takes a set of predicted future distances corresponding to each set of real-time feature data as the output, and takes the actual future distance corresponding to each set of real-time feature data as the prediction target. The actual future distance is the pre-set future distance corresponding to the real-time feature data; minimizing the sum of the prediction errors of all real-time feature data is used as the training target; the distance prediction model is trained until the sum of the prediction errors reaches convergence and then the training stops; the distance prediction model is a deep neural network model;

[0030] In the step S104, the method for constructing the driving trajectory corresponding to each moving obstacle is: sorting the predicted distances corresponding to each moving obstacle from the earliest to the latest according to the corresponding prediction order, and obtaining a distance sorting table corresponding to each moving obstacle; according to the distance sorting table, constructing the driving trajectory corresponding to each moving obstacle.

[0031] Further, the method for evaluating the collision risk includes:

[0032] An optimization algorithm is used to set the collision threshold, and the predicted distances in the driving trajectories corresponding to each moving obstacle are respectively compared with the collision threshold; if there is a predicted distance less than or equal to the collision threshold, the corresponding moving obstacle is marked as a collision obstacle; if all predicted distances are greater than the collision threshold, the corresponding moving obstacle is not marked; the target azimuth angle at the current moment is obtained and marked as the current azimuth angle; in the driving trajectory corresponding to the collision obstacle, the first predicted distance less than or equal to the collision threshold is marked as the collision distance; the future moment corresponding to the collision distance is marked as the collision moment, and according to the current moment and the collision moment, a time interval is constructed; the predicted speed and predicted distance corresponding to the collision obstacle within the time interval are used as the prediction set, and the prediction set and the current azimuth angle are input into the trained angle prediction model to predict the target azimuth angle corresponding to the collision moment and mark it as the collision azimuth angle;

[0033] The collision azimuth angle is compared with 0 to determine the spatial location of the collision obstacle; if the collision azimuth angle is greater than 0, the collision obstacle is on the right side of the vehicle; if the collision azimuth angle is less than 0, the collision obstacle is on the left side of the vehicle; the door corresponding to the door opening intention command is marked as the door to be opened, and the spatial location of the door to be opened is obtained; the spatial location of the door to be opened is compared with the spatial location of the collision obstacle; if the door to be opened and the collision obstacle are located on the same side, it is assessed that there is a collision risk with the corresponding collision obstacle; if the door to be opened and the collision obstacle are not located on the same side, it is assessed that there is no collision risk with the corresponding collision obstacle.

[0034] Furthermore, the steps of setting the collision threshold using the optimization algorithm include:

[0035] Step S201: Preset initialization temperature , minimum temperature , cooling coefficient and the maximum number of iterations , and let the current temperature ;

[0036] Step S202: preset threshold range; randomly set a feasible solution , feasible solution is a value within the threshold range, the feasible solution The range is the threshold range;

[0037] Step S203: determining the fitness function;

[0038] Step S204: Calculate feasible solutions The corresponding fitness ; feasible solution For the current point, perform random perturbations in the neighborhood of the current point to obtain a new feasible solution , and calculate the new feasible solution The corresponding fitness ;

[0039] Step S205: Calculate the fitness difference If the fitness difference , then let , which will be the new feasible solution The value of is given to the feasible solution If the fitness difference , then calculate the probability , according to the probability make ;

[0040] Step S206: Loop steps S204 to S205 until the number of loops reaches the maximum number of iterations. When the cycle ends, let the current temperature , make the maximum number of iterations ;

[0041] Step S207: Loop through steps S204 to S206 until the current temperature When it is, the loop ends, and the numerical value corresponding to the feasible solution is obtained and used as the collision threshold.

[0042] Furthermore, in the above-mentioned step S203, the expression of the fitness function is: ;

[0043] In the formula, is the fitness, is the accuracy; the accuracy is obtained through a preset threshold set, and the threshold set includes the accuracies corresponding to different collision thresholds;

[0044] In the above-mentioned step S205, the fitness difference has the expression ;

[0045] The probability has the expression: ; where e is the natural constant.

[0046] Furthermore, the method for giving corresponding warnings to passengers includes:

[0047] If the vehicle speed is less than the vehicle speed threshold and there is a collision with an obstacle, the warning lights on the vehicle rearview mirror and the door corresponding to the spatial position of the collision obstacle light up;

[0048] If an open door intention instruction is generated and there is a collision with an obstacle, the buzzer in the vehicle gives a warning, and the warning lights on the vehicle rearview mirror and the door to be opened flash;

[0049] If the vehicle speed is greater than or equal to the vehicle speed threshold and there is a collision with an obstacle, the warning light on the vehicle rearview mirror lights up.

[0050] A door opening anti-collision warning system that implements the above-mentioned door opening anti-collision warning method includes:

[0051] A data acquisition module for collecting intention analysis data;

[0052] An intention recognition module for analyzing the intention analysis data to recognize the passenger's door opening intention;

[0053] A data acquisition module for obtaining vehicle characteristic data;

[0054] A trajectory prediction module that predicts the driving trajectory based on the vehicle characteristic data;

[0055] A collision analysis module, configured to analyze the driving trajectory and evaluate the collision risk;

[0056] A risk warning module, configured to give corresponding warnings to passengers according to the evaluation result of the collision risk and the door opening intention.

[0057] The technical effects and advantages of a method and system for anti-collision warning when opening a vehicle door according to the present invention:

[0058] By collecting intention analysis data, the door opening intention of passengers can be accurately identified; and by obtaining vehicle characteristic data, the driving trajectory of moving obstacles can be effectively predicted, so as to accurately analyze the collision risk between the vehicle door and the moving obstacles, and then take differential warnings, making full use of in-vehicle sensor data and data analysis algorithms to realize intelligent collision warning when the vehicle door is opened; at the same time, technologies such as adaptively adjusting the collision threshold and combining spatial relationships for accurate risk assessment are adopted to effectively avoid false alarms or missed alarms, improve the accuracy and reliability of the warning; it can accurately identify the door opening risk and timely remind vehicle passengers, reduce the accident risk caused by vehicle blind spots, improve the safety of opening the vehicle door, and effectively guarantee the traffic safety of pedestrians and cyclists. Brief Description of the Drawings

[0059] Figure 1 It is a schematic diagram of a system for anti-collision warning when opening a vehicle door according to Embodiment 1 of the present invention;

[0060] Figure 2 It is a flowchart of a method for predicting a driving trajectory according to Embodiment 1 of the present invention;

[0061] Figure 3 It is a flowchart of a method for anti-collision warning when opening a vehicle door according to Embodiment 2 of the present invention. Detailed Embodiments

[0062] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments 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.

[0063] Embodiment 1

[0064] Please refer to Figure 1 As shown, the system for anti-collision warning when opening a vehicle door described in this embodiment includes a data collection module, an intention recognition module, a data acquisition module, a trajectory prediction module, a collision analysis module, and a risk warning module; each module is connected by wired and / or wireless means to realize data transmission between modules.

[0065] The data acquisition module is used to acquire intention analysis data.

[0066] The intention analysis data includes vehicle speed and door distance; the vehicle speed is the driving speed of the vehicle, and the vehicle speed is obtained by a vehicle speed sensor installed at the bottom of the vehicle; the door distance is the distance between the passenger and the inner side of the door, and the door distance is obtained by a laser ranging sensor installed on the inner side of the door.

[0067] The intention recognition module is used to analyze the intention analysis data and recognize the passenger's door opening intention.

[0068] The method for recognizing the passenger's door opening intention includes:

[0069] Preset intention thresholds, where the intention thresholds include a vehicle speed threshold and a vehicle distance threshold. Both the vehicle speed threshold and the vehicle distance threshold are preset by those skilled in the art according to the actual situation. In this embodiment, the preferred vehicle speed threshold is , and the vehicle distance threshold is ;

[0070] Compare the vehicle speed with the vehicle speed threshold, and compare the door distance with the vehicle distance threshold; if the vehicle speed is less than the vehicle speed threshold and the door distance is less than the vehicle distance threshold, then generate a door opening intention instruction, indicating that the passenger has a door opening intention; if the vehicle speed is greater than or equal to the vehicle threshold, or the door distance is greater than or equal to the vehicle distance threshold, then no door opening intention instruction is generated, indicating that the passenger does not have a door opening intention.

[0071] The data acquisition module is used to acquire vehicle feature data.

[0072] The vehicle feature data includes a target distance and a relative speed; the target distance is the distance between the vehicle and a moving obstacle, and the relative speed is the moving speed of the moving obstacle relative to the vehicle; the moving obstacle is a moving object in the traffic environment that can interfere with or hinder the opening of the vehicle door, such as pedestrians, cyclists, other vehicles, etc.; the vehicle feature data is obtained by a millimeter-wave radar installed on the rear bumper of the vehicle.

[0073] The trajectory prediction module predicts the driving trajectory according to the vehicle feature data.

[0074] As Figure 2 shown, the steps for predicting the driving trajectory include:

[0075] Step S101: Take the vehicle feature data corresponding to the same moving obstacle as a set of data, and each set of data corresponds to a moving obstacle one by one; by aggregating the vehicle feature data of the same moving obstacle, it helps to capture the dynamic characteristics of different moving obstacles, reduce the possibility of data mixing, and improve the accuracy and effectiveness of the subsequent prediction process;

[0076] Step S102: According to the relative speeds in each data set, predict the relative speeds of each moving obstacle at corresponding future moments, mark them as predicted speeds, and add each predicted speed to the corresponding data set; By predicting the relative speeds at future moments, it helps to timely understand the future relative speeds of the moving obstacles, thereby improving the accuracy of predicting the future target distances subsequently;

[0077] Step S103: According to the predicted speeds and target distances in each data set, predict the target distances of each moving obstacle at corresponding future moments, mark them as predicted distances, and add each predicted distance to the corresponding data set; By combining the predicted speeds and target distances, it can more accurately predict the predicted distances at future moments, thereby obtaining the future driving trajectories of the moving obstacles, which helps to evaluate potential collision risks;

[0078] Step S104: Based on the predicted distances in each data set, construct the driving trajectories corresponding to each moving obstacle.

[0079] In the above Step S102, the method for predicting the relative speed of each moving obstacle at a corresponding future moment includes:

[0080] Take the relative speeds in each data set as a speed set, and the speed sets correspond to the data sets one by one; Input each speed set into the trained speed prediction model to predict the relative speeds of each moving obstacle at corresponding future moments, and the speed prediction model is an RNN neural network model; In this embodiment, preferably the moment is 5s; The training method of the speed prediction model includes:

[0081] Pre-collect consecutive relative speeds and construct a speed training set, where is an integer greater than 1; Based on the speed training set, train the speed prediction model for predicting the relative speed at future moments;

[0082] Preset the sliding step length L and the sliding window length D according to the actual experience of those skilled in the art; Use the sliding window method to convert the relative speeds in the speed training set into multiple training samples, and each training sample includes D relative speeds; Take each training sample as the input of the speed prediction model, predict the relative speed after the sliding step length L as the output, and take the last L relative speeds of each training sample as the prediction target. Use the mean absolute percentage error MAPE to evaluate the model accuracy of the prediction result. When the calculated MAPE is less than the preset MAPE, the speed prediction model corresponding to the relative speed is trained; Among them, the calculation formula of the mean absolute percentage error MAPE is:

[0083] ; In the formula, is the prediction target corresponding to the th predicted relative speed, is the th predicted relative speed, is the number of predicted relative speeds; generate a speed prediction model for predicting the relative speed at a future moment based on the relative speed.

[0084] Exemplarily, the speed training set A contains 10 relative speeds, , is the th relative speed, , define the length of the sliding window as 3 and the sliding step length L as 1, use the sliding window to construct 8 training samples, each training sample contains 3 consecutive relative speeds, and use the next relative speed of the 3 consecutive relative speeds as the prediction target; for example:

[0085] Training sample , training sample The corresponding prediction target is ; training sample , training sample The corresponding prediction target is ; and so on, for training the speed prediction model corresponding to the relative speed.

[0086] In the above step S103, the method for predicting the target distance of each moving obstacle corresponding to a future moment includes:

[0087] Obtain the acquisition moment corresponding to each set of vehicle feature data, sort the acquisition moments corresponding to the vehicle feature data in the same data set from the latest to the earliest, and obtain the moment sorting table corresponding to each data set; use the vehicle feature data corresponding to the acquisition moment ranked first in each moment sorting table as the real-time feature data of the corresponding data set; input each set of real-time feature data into the trained distance prediction model respectively to predict the target distance at a future moment; replace the target distance in each set of real-time feature data with the target distance at a corresponding future moment, and input the replaced real-time feature data into the trained distance prediction model respectively to predict the target distance at two future moments; and so on, and predict the target distance of each moving obstacle corresponding to the future moments;

[0088] The training process of the distance prediction model includes:

[0089] Pre-collect sets of real-time feature data, set corresponding future distances for sets of real-time feature data, and the future distance is the target distance at a future moment, Let \(n\) be an integer greater than 1. Convert the real-time feature data and the corresponding future distances into a corresponding set of feature vectors. The future distances corresponding to the real-time feature data are collected by those skilled in the art during the process of predicting the target distances of each moving obstacle at corresponding future times. Collect a set of real-time feature data, and sequentially collect the future distances corresponding to each set of real-time feature data. Sequentially set the corresponding future distances for each set of real-time feature data.

[0090] Take each set of feature vectors as the input of the distance prediction model. The distance prediction model outputs a set of predicted future distances corresponding to each set of real-time feature data, and takes the actual future distance corresponding to each set of real-time feature data as the prediction target. The actual future distance is the pre-set future distance corresponding to the real-time feature data. Take minimizing the sum of the prediction errors of all real-time feature data as the training objective. Among them, the calculation formula of the prediction error is where \(e_i\) is the prediction error, \(i\) is the group number of the feature vector corresponding to the real-time feature data, the \(\hat{d}_i\) is the predicted future distance corresponding to the \(i\)-th set of real-time feature data, the \(d_i\) is the actual future distance corresponding to the \(i\)-th set of real-time feature data. Train the distance prediction model until the sum of the prediction errors reaches convergence and then stop training.

[0091] The above distance prediction model is specifically a deep neural network model, which includes an input layer, a hidden layer, and an output layer. Each hidden layer includes multiple neurons, and there are connections between each neuron and the neurons in the next layer. The connections contain weights, which determine the importance and influence of data transmission in the neural network. An activation function is applied to each neuron between the hidden layer and the output layer. The activation function introduces non-linearity and allows the network to learn more complex patterns and features.

[0092] In the above step S104, the method for constructing the driving trajectory corresponding to each moving obstacle is as follows: Sort the predicted distances corresponding to each moving obstacle from earliest to latest according to the corresponding prediction order to obtain a distance sorting table for each moving obstacle. Construct the driving trajectory corresponding to each moving obstacle according to the distance sorting table.

[0093] The collision analysis module is used to analyze the driving trajectory and evaluate the collision risk.

[0094] The methods for evaluating the collision risk include:

[0095] Set the collision threshold using an optimization algorithm, and compare the predicted distances in the driving trajectories corresponding to each moving obstacle with the collision threshold respectively; if there is a predicted distance less than or equal to the collision threshold, mark the corresponding moving obstacle as a collision obstacle; if all predicted distances are greater than the collision threshold, do not mark the corresponding moving obstacle; obtain the target azimuth angle at the current moment and mark it as the current azimuth angle; the target azimuth angle is the angle of the collision obstacle relative to the vehicle's forward direction, and the target azimuth angle is obtained by a millimeter-wave radar installed on the vehicle's rear bumper; mark the first predicted distance less than or equal to the collision threshold in the driving trajectory corresponding to the collision obstacle as the collision distance; mark the future moment corresponding to the collision distance as the collision moment, and construct a time interval based on the current moment and the collision moment; use the predicted speed and predicted distance corresponding to the collision obstacle within the time interval as a prediction set, and input the prediction set and the current azimuth angle into a trained angle prediction model to predict the target azimuth angle corresponding to the collision moment and mark it as the collision azimuth angle;

[0096] Compare the collision azimuth angle with 0 to determine the spatial positioning of the collision obstacle; if the collision azimuth angle is greater than 0, the collision obstacle is on the right side of the vehicle, and if the collision azimuth angle is less than 0, the collision obstacle is on the left side of the vehicle; mark the door corresponding to the door opening intention instruction as the door to be opened, and obtain the spatial positioning of the door to be opened; the spatial positioning is either on the left side or the right side of the vehicle; compare the spatial positioning of the door to be opened with the spatial positioning of the collision obstacle; if the door to be opened and the collision obstacle are on the same side, evaluate that there is a collision risk for the corresponding collision obstacle; if the door to be opened and the collision obstacle are not on the same side, evaluate that there is no collision risk for the corresponding collision obstacle.

[0097] The steps of setting the collision threshold using an optimization algorithm include:

[0098] Step S201: Preset the initial temperature , the lowest temperature , the cooling coefficient and the maximum number of iterations , and let the current temperature ; By initializing the parameters, it helps to control the convergence speed of the algorithm, avoid premature convergence to the local optimal solution, and improve the global search ability;

[0099] Step S202: Preset the threshold range, which is pre-set by those skilled in the art according to the actual situation to ensure that the feasible solution is within a reasonable boundary, improving the practicality and reliability of the algorithm; randomly set a feasible solution , the feasible solution is a value within the threshold range, and the feasible solution The range is the threshold range; randomly generating feasible solutions can introduce diversity and increase the exploration opportunities, thereby enhancing the possibility of finding the optimal solution;

[0100] Step S203: Determine the fitness function; it helps to perform quantitative calculations on the evaluation criteria of feasible solutions, thereby judging the quality of feasible solutions;

[0101] Step S204: Calculate the feasible solution corresponding fitness ; taking the feasible solution as the current point, perform random perturbation within the neighborhood of the current point to obtain a new feasible solution , and calculate the fitness corresponding to the new feasible solution ; it helps to discover better solutions and enhance the local search ability of the algorithm;

[0102] Step S205: Calculate the fitness difference ; if the fitness difference , then let , that is, assign the value of the new feasible solution to the feasible solution ; if the fitness difference , then calculate the probability , and according to the probability let ; allowing the algorithm to dynamically adjust the feasible solution according to the quality of the solution helps to achieve a more flexible search process, and accepting inferior solutions according to the probability helps to avoid falling into local optima and increase the possibility of global search;

[0103] Step S206: Loop steps S204 to S205 until the number of loops reaches the maximum number of iterations ; when it reaches, the loop ends; let the current temperature , that is, cool down the current temperature in step S201 and assign the cooled value to the current temperature; let the maximum number of iterations , that is, assign the reduced value of the maximum number of iterations to the maximum number of iterations; if the reduced maximum number of iterations is not an integer, then round up the reduced maximum number of iterations to make it an integer; gradually cooling down helps the algorithm to gradually converge, reduce the acceptance of inferior solutions, and finally find the optimal solution;

[0104] Step S207: Loop steps S204 to S206 until the current temperature ; when it reaches, the loop ends, obtain the value corresponding to the feasible solution , and use it as the collision threshold.

[0105] In the above step S203, the expression of the fitness function is: ;

[0106] In the formula, is the fitness, is the accuracy; the accuracy is obtained through a preset set of thresholds, and the set of thresholds includes the accuracy corresponding to different collision thresholds, and the set of thresholds is preset by those skilled in the art through collision tests and in combination with the actual situation.

[0107] In the above step S205, the fitness difference has the expression of ;

[0108] The probability has the expression of: ; in the formula, e is the natural constant.

[0109] The risk warning module gives corresponding warnings to passengers according to the evaluation result of the collision risk and the door opening intention.

[0110] The method of giving corresponding warnings to passengers includes:

[0111] If the vehicle speed is less than the vehicle speed threshold and there is a collision with an obstacle, the warning lights installed on the vehicle rearview mirror and the door with the same spatial positioning as the collision obstacle will light up;

[0112] If a door opening intention instruction is generated and there is a collision with an obstacle, the buzzer in the vehicle will give a warning, that is, it will sound three times continuously, and the warning lights on the vehicle rearview mirror and the door to be opened will flash;

[0113] If the vehicle speed is greater than or equal to the vehicle speed threshold and there is a collision with an obstacle, the warning light on the vehicle rearview mirror will light up.

[0114] In this embodiment, by collecting intention analysis data, the door opening intention of passengers is accurately identified; and by obtaining vehicle characteristic data, the driving trajectory of moving obstacles is effectively predicted, so as to accurately analyze the collision risk between the vehicle door and the moving obstacle, and then adopt differential warnings, making full use of in-vehicle sensor data and data analysis algorithms to achieve intelligent collision warning when the vehicle door is opened; at the same time, technologies such as adaptively adjusting the collision threshold and combining spatial relationships for accurate risk assessment are adopted to effectively avoid false alarms or missed alarms, improve the accuracy and reliability of the warning; it can accurately identify the door opening risk and timely remind the vehicle passengers, reduce the accident risk caused by vehicle blind spots, improve the safety of door opening, and effectively guarantee the traffic safety of pedestrians and cyclists.

[0115] Embodiment 2

[0116] Please refer to Figure 3As shown, for the parts not described in detail in this embodiment, refer to the description in Embodiment 1. A method for anti-collision warning when opening a vehicle door is provided. The method includes:

[0117] Collect intention analysis data;

[0118] Analyze the intention analysis data to identify the passenger's intention to open the door;

[0119] Obtain vehicle characteristic data;

[0120] Predict the driving trajectory according to the vehicle characteristic data;

[0121] Analyze the driving trajectory to evaluate the collision risk;

[0122] Give corresponding warnings to the passengers according to the evaluation result of the collision risk and the intention to open the door.

[0123] Embodiment 3

[0124] This application also provides an electronic device. The electronic device may include one or more processors and one or more memories. Among them, computer-readable code is stored in the memory. When the computer-readable code is run by one or more processors, it can execute an anti-collision warning method for opening a vehicle door as described above.

[0125] The method or system according to the embodiment of the present application can also be implemented by means of the architecture of the electronic device shown in the present application. The electronic device may include a bus, one or more CPUs, a ROM, a RAM, a communication port connected to the network, an input / output, a hard disk, etc. The storage device in the electronic device, such as ROM or hard disk, can store an anti-collision warning method for opening a vehicle door provided by the present application. Further, the electronic device may also include a user interface. Of course, the architecture shown in the present application is only exemplary. When implementing different devices, one or more components shown in the electronic device of the present application can be omitted according to actual needs.

[0126] Embodiment 4

[0127] An embodiment of the present application discloses a computer-readable storage medium. Computer-readable instructions are stored on the computer-readable storage medium. When the computer-readable instructions are run by a processor, an anti-collision warning method for opening a vehicle door according to the embodiment of the present application as described with reference to the above drawings can be executed. The storage medium includes but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory, etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0128] In addition, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium storing machine-readable instructions that can be run by a processor to execute instructions corresponding to the method steps provided by the present application, such as: a method for anti-collision warning when opening a vehicle door. When the computer program is executed by a central processing unit (CPU), the above functions defined in the method of the present application are executed.

[0129] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

[0130] Finally: The above is only a preferred embodiment of the present invention and is not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should all be included within the protection scope of the present invention.

Claims

1. A method for anti-collision warning when opening a vehicle door, characterized in that, Including: Collecting intention analysis data; Analyzing the intention analysis data to identify the passenger's door-opening intention; Obtaining vehicle characteristic data; Predicting the driving trajectory according to the vehicle characteristic data; The steps of predicting the driving trajectory include: Step S101: Taking the vehicle characteristic data corresponding to the same moving obstacle as a set of data, and the data sets correspond one-to-one with the moving obstacles; Step S102: Predicting the relative speed of each moving obstacle at a future moment according to the relative speed in each data set, and marking it as the predicted speed, and adding each predicted speed to the corresponding data set; Step S103: Predicting the target distance of each moving obstacle at a future moment according to the predicted speed and the target distance in each data set, and marking it as the predicted distance, and adding each predicted distance to the corresponding data set; Step S104: Constructing the driving trajectory corresponding to each moving obstacle based on the predicted distance in each data set; In the step S102, the method for predicting the relative speed of each moving obstacle at a future moment includes: Take the relative speed in each data set as a speed set, and the speed sets correspond to the data sets one by one; input each speed set into the trained speed prediction model to predict the relative speed of each moving obstacle corresponding to the future relative speeds at several moments, and the speed prediction model is an RNN neural network model; the training method of the speed prediction model includes: Pre-acquire consecutive relative velocities and construct a velocity training set, where is an integer greater than 1; Based on the velocity training set, train a velocity prediction model to predict the relative velocity at a future time; Presetting a sliding step length L and a sliding window length D; converting the relative speeds in the speed training set into multiple training samples by using the sliding window method, and each training sample includes D relative speeds; taking each training sample as the input of the speed prediction model, predicting the relative speed after the sliding step length L as the output, taking the last L relative speeds of each training sample as the prediction target, evaluating the model accuracy by using the mean absolute percentage error MAPE for the prediction result, and when the calculated MAPE is less than the preset MAPE, the speed prediction model corresponding to the relative speed is trained; generating a speed prediction model for predicting the relative speed at a future moment according to the relative speed; Analyzing the driving trajectory to evaluate the collision risk; Giving corresponding warnings to the passengers according to the evaluation result of the collision risk and the door-opening intention.

2. The anti-collision warning method for door opening according to claim 1, wherein The intention analysis data includes the vehicle speed and the door distance; The door distance is the distance between the passenger and the inner side of the door; The method for identifying the passenger's door-opening intention includes: Presetting an intention threshold, and the intention threshold includes a vehicle speed threshold and a vehicle distance threshold; comparing the vehicle speed with the vehicle speed threshold, and comparing the door distance with the vehicle distance threshold; if the vehicle speed is less than the vehicle speed threshold and the door distance is less than the vehicle distance threshold, then generating a door-opening intention instruction; if the vehicle speed is greater than or equal to the vehicle threshold, or the door distance is greater than or equal to the vehicle distance threshold, then not generating a door-opening intention instruction; The vehicle characteristic data includes the target distance and the relative speed; the target distance is the distance between the vehicle and the moving obstacle, and the relative speed is the moving speed of the moving obstacle relative to the vehicle.

3. The anti-collision warning method for door opening according to claim 2, wherein In the step S103, the method for predicting the target distance of each moving obstacle at a future moment includes: Obtain the acquisition time corresponding to each set of vehicle feature data, sort the acquisition times corresponding to the vehicle feature data in the same data set from late to early, and obtain the time sorting table corresponding to each data set; Use the vehicle feature data corresponding to the acquisition time ranked first in each time sorting table as the real-time feature data of the corresponding data set; Input each set of real-time feature data into the trained distance prediction model respectively to predict the target distance at the next moment; Replace the target distance in each set of real-time feature data with the corresponding target distance at the next moment, and input the replaced real-time feature data into the trained distance prediction model respectively to predict the target distance at the next two moments; And so on, predict the target distances corresponding to each moving obstacle at the next moments; The training process of the distance prediction model includes: Pre-collect a set of real-time feature data, for each set of real-time feature data, set a corresponding future distance, where the future distance is the target distance at a future moment, k is an integer greater than 1, convert the real-time feature data and the corresponding future distance into a corresponding set of feature vectors; use each set of feature vectors as the input of the distance prediction model, where the distance prediction model outputs a set of predicted future distances corresponding to each set of real-time feature data, and uses the actual future distance corresponding to each set of real-time feature data as the prediction target, and the actual future distance is the pre-set future distance corresponding to the real-time feature data; use minimizing the sum of prediction errors of all real-time feature data as the training target; train the distance prediction model until the sum of prediction errors reaches convergence and then stop training; the distance prediction model is a deep neural network model; In step S104, the method for constructing the driving trajectory corresponding to each moving obstacle is as follows: sort the predicted distances corresponding to each moving obstacle from the earliest to the latest according to the corresponding prediction order to obtain a distance sorting table for each moving obstacle; construct the driving trajectory corresponding to each moving obstacle according to the distance sorting table.

4. A method for anti-collision warning when a vehicle door is opened according to claim 3, characterized in that, The method for evaluating the collision risk includes: Set the collision threshold using an optimization algorithm, and compare the predicted distances in the driving trajectory corresponding to each moving obstacle with the collision threshold respectively; if there is a predicted distance less than or equal to the collision threshold, mark the corresponding moving obstacle as a collision obstacle; if all predicted distances are greater than the collision threshold, do not mark the corresponding moving obstacle; obtain the target azimuth angle at the current moment and mark it as the current azimuth angle; mark the first predicted distance less than or equal to the collision threshold in the driving trajectory corresponding to the collision obstacle as the collision distance; mark the future moment corresponding to the collision distance as the collision moment, and construct a time interval according to the current moment and the collision moment; use the predicted speed and predicted distance corresponding to the collision obstacle within the time interval as a prediction set, and input the prediction set and the current azimuth angle into the trained angle prediction model to predict the target azimuth angle corresponding to the collision moment and mark it as the collision azimuth angle; Compare the collision azimuth angle with 0 to determine the spatial location of the collision obstacle; if the collision azimuth angle is greater than 0, the collision obstacle is on the right side of the vehicle, and if the collision azimuth angle is less than 0, the collision obstacle is on the left side of the vehicle; mark the door corresponding to the door opening intention instruction as the door to be opened, and obtain the spatial location of the door to be opened; compare the spatial location of the door to be opened with the spatial location of the collision obstacle; if the door to be opened and the collision obstacle are on the same side, evaluate that there is a collision risk for the corresponding collision obstacle; if the door to be opened and the collision obstacle are not on the same side, evaluate that there is no collision risk for the corresponding collision obstacle.

5. A method for anti-collision warning when a vehicle door is opened according to claim 4, characterized in that, The steps for setting the collision threshold using an optimization algorithm include: Step S201: Preset the initial temperature , the minimum temperature , the temperature reduction coefficient and the maximum number of iterations , and set the current temperature ; Step S202: Preset a threshold range; randomly set a feasible solution , where the feasible solution is a value within the threshold range, and the range of the feasible solution is the threshold range; Step S203: Determine the fitness function; Step S204: Calculate the feasible solution The corresponding fitness ; Take the feasible solution as the current point, perform a random perturbation within the neighborhood of the current point to obtain a new feasible solution , and calculate the fitness of the new feasible solution corresponding thereto ; Step S205: Calculate the fitness difference ; If the fitness difference , then let , that is, assign the value of the new feasible solution to the feasible solution ; If the fitness difference , then calculate the probability , and according to the probability let ; Step S206: Loop through steps S204 to S205 until the number of loops reaches the maximum number of iterations At this time, the loop ends; let the current temperature , let the maximum number of iterations ; Step S207: Loop through steps S204 to S206 until the current temperature when the loop ends, obtain the value corresponding to the feasible solution and use it as the collision threshold.

6. The anti-collision warning method for opening a vehicle door according to claim 5, characterized in that, In the above-mentioned step S203, the expression of the fitness function is as follows: ; wherein, is the fitness, is the accuracy; the accuracy is obtained through a preset threshold set, and the threshold set includes accuracies corresponding to different collision thresholds; In the above step S205, the fitness difference has the expression ; Probability The expression is as follows: ; where e is the natural constant.

7. A method for anti-collision warning when opening a vehicle door according to claim 6, characterized in that The method for giving corresponding warnings to passengers includes: If the vehicle speed is less than the vehicle speed threshold and there is a collision obstacle, the warning lights on the vehicle rearview mirror and the door with the same spatial location as the collision obstacle will light up; If a door opening intention instruction is generated and there is a collision obstacle, the buzzer in the vehicle will give a warning, and the warning lights on the vehicle rearview mirror and the door to be opened will flash; If the vehicle speed is greater than or equal to the vehicle speed threshold and there is a collision obstacle, the warning light on the vehicle rearview mirror will light up.

8. A door opening anti-collision warning system, which is applied to the door opening anti-collision warning method described in any one of claims 1-7, is characterized in that, It includes: A data acquisition module for collecting intention analysis data; An intention recognition module for analyzing the intention analysis data to recognize the passenger's door opening intention; A data acquisition module for obtaining vehicle feature data; A trajectory prediction module for predicting the driving trajectory according to the vehicle feature data; A collision analysis module for analyzing the driving trajectory to evaluate the collision risk; A risk warning module for giving corresponding warnings to passengers according to the evaluation result of the collision risk and the door opening intention.

Images