Collision early warning method, device and equipment for electric unmanned vehicle

Abstract

The invention belongs to the technical field of intelligent automobiles, and discloses a collision early warning method, device and equipment for an electric unmanned vehicle. The method comprises the following steps: acquiring environmental information data around a vehicle in real time through a sensor; preprocessing the acquired environment information data to obtain effective data which can be used for subsequent analysis; calculating a predicted track of the vehicle and a motion track of the target object according to the obtained effective data; calculating the probability of collision by using a collision probability model according to a calculation result; when the calculated collision probability exceeds a preset threshold value, an early warning signal is sent to a vehicle control system; and after the early warning signal is received, the vehicle control system takes corresponding emergency braking measures to avoid collision. According to the invention, by using the sensor and the collision probability model, comprehensive monitoring of the surrounding environment of the electric unmanned vehicle is realized, the potential collision risk is predicted in advance, and the safety of vehicle driving is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent vehicles, and in particular relates to a collision warning method, device and equipment for an electric unmanned vehicle. Background Art

[0002] With the continuous advancement of science and technology and people's demand for travel, electric driverless vehicles have become a hot topic in current research and development. As a means of transportation with highly autonomous and intelligent features, electric driverless vehicles have great potential in improving travel efficiency, reducing traffic accidents, and alleviating environmental pollution. However, electric driverless vehicles still face many challenges in practical applications, one of which is how to effectively predict and avoid collision accidents. Collision accidents can threaten traffic safety and the safety of passengers' lives and property. Therefore, developing a reliable collision warning method is crucial to the safety of electric driverless vehicles.

[0003] In traditional collision warning systems, only sensors (such as radars and cameras) are usually used to sense the surrounding environment to determine whether there is a risk of collision. However, due to the high degree of autonomy of electric unmanned vehicles, relying solely on sensor data cannot meet the accuracy and real-time requirements of collision warning in complex road environments. The present invention proposes a collision warning method, device and equipment for electric unmanned vehicles, which adopts a probability calculation model based on a Bayesian network, combines vehicle dynamic information and environmental perception data, and uses a Kalman filter algorithm to update the speed and direction of the target object, thereby achieving comprehensive monitoring of the surrounding environment of the electric unmanned vehicle, improving the accuracy and real-time performance of the collision warning, and ensuring the safety of vehicle driving by predicting potential collision risks in advance. Summary of the invention

[0004] In view of the deficiencies of the above-mentioned existing methods, the present invention provides a collision warning method, device and equipment for an electric unmanned vehicle, aiming to solve the core problem that traditional collision warning methods cannot meet the requirements of collision warning accuracy and real-time performance.

[0005] To achieve the above-mentioned purpose, the present invention is implemented by the following technical solutions: A collision warning method for an electric unmanned vehicle, the method comprising the following steps: Step S10: acquiring environmental information data and road condition information data around the vehicle in real time through sensors installed on the electric unmanned vehicle; Step S20: pre-processing the environmental information data and road condition information data collected in real time to obtain valid data that can be used for subsequent analysis; Step S30: Calculate the predicted driving trajectory of the vehicle according to the obtained valid data, and track the motion trajectory of the target object; Step S40: Calculate the probability of collision using a collision probability model according to the calculation result of step S30 and the tracked motion trajectory of the target object; Step S50: When the calculated collision probability exceeds a preset threshold, a warning signal is sent to the vehicle control system; Step S60: After receiving the warning signal, the vehicle control system takes corresponding emergency braking measures to avoid collision.

[0006] Preferably, in step S10, the sensor refers to a variety of sensors installed on the electric unmanned vehicle, including cameras, lidars, millimeter-wave radars and ultrasonic sensors, etc.; the environmental information data around the vehicle includes the location, speed, shape, etc. of obstacles, and the road condition information data on the vehicle includes the degree of vehicle congestion, visibility, weather conditions, etc.

[0007] Preferably, the data preprocessing step in step S20 includes noise removal, missing value filling and feature extraction.

[0008] Preferably, in step S30, the predicted driving trajectory of the vehicle is calculated by formula (1): (1) in, x ( t ) is the time at which the vehicle t The position at x 0 is the initial position, v x is the initial velocity, a x is the acceleration.

[0009] Preferably, the process of tracking the motion trajectory of the target object in step S30 includes: Target tracking: Use multi-target tracking algorithms to track surrounding dynamic objects; Target correction: Update the speed and direction of the target object through the Kalman filter algorithm.

[0010] Preferably, the update formula of the Kalman filter algorithm is as shown in formula (2): (2) in, is the current estimated state, is the Kalman gain, is the measured value, H is the measurement matrix, is the estimated state at the previous moment.

[0011] Preferably, the collision probability model in step S40 adopts a Bayesian network model, which integrates information data of the vehicle's surrounding environment collected by various sensors through the Bayesian network and evaluates the probability of collision. The steps of constructing the collision probability model include: Define variables: Determine the variables that can affect the probability of a collision, including the vehicle's surrounding environment information, vehicle status, and obstacle information; Formulate hypotheses: Make assumptions about the relationships between variables based on the impact of the vehicle's speed and the location of obstacles on the vehicle's collision; Determine the conditional probability table: Based on the selected variables and assumptions, construct a conditional probability table to describe the dependencies between variables; Build Bayesian network structures: Use Bayesian network graphs to visualize variables and their dependencies; Parameter learning: Using existing data, the parameters of the constructed Bayesian network are estimated through maximum likelihood estimation or Bayesian Learning; Reasoning and prediction: Based on the existing data and learned parameters, reasoning and prediction are performed, and the probability of a collision at a certain point in the future is predicted based on the information data of the vehicle's surrounding environment collected by sensors; Update and optimize: After accumulating more data, the parameters of the Bayesian network are continuously updated.

[0012] Preferably, in the continuously updating the parameters of the Bayesian network, the updating formula of the Bayesian network parameters is as shown in formula (3): (3) in, P ( Collision | Data ) represents the collision probability under the given vehicle surrounding environment information data collected by the sensor, P ( Data | Collision ) represents the conditional probability of the vehicle surrounding environment information data collected when a collision occurs, P ( Collision ) is the prior probability of collision, P ( Data ) represents the prior probability of the information data of the vehicle's surrounding environment collected by the sensor.

[0013] In addition, to achieve the above-mentioned purpose, the present invention also proposes a collision warning device for an electric unmanned vehicle, which is characterized by: Data acquisition module: used to obtain environmental information data around the vehicle in real time through sensors; Data preprocessing module: used to preprocess the collected environmental information data to obtain valid data that can be used for subsequent analysis; Trajectory prediction and tracking module: used to calculate the predicted trajectory of the vehicle based on the valid data obtained, and track the motion trajectory of the target object; Collision probability calculation module: used to calculate the probability of collision using the collision probability model based on the calculation results and the tracked target object motion trajectory; Warning module: used to send a warning signal to the vehicle control system when the calculated collision probability exceeds a preset threshold; Warning response module: After receiving the warning signal, the vehicle control system takes corresponding emergency braking measures to avoid collision.

[0014] In addition, to achieve the above-mentioned purpose, the present invention also proposes a collision warning device for an electric unmanned vehicle, wherein the collision warning device for an electric unmanned vehicle comprises: a memory, a processor, and a collision warning program for an electric unmanned vehicle stored in the memory and executable on the processor, wherein when the collision warning program for an electric unmanned vehicle is executed by the processor, the steps of a collision warning method for an electric unmanned vehicle as described above are implemented.

[0015] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer program product, which includes a collision warning program for an electric unmanned vehicle. When the collision warning program for an electric unmanned vehicle is executed by a processor, it implements a collision warning method for an electric unmanned vehicle as described above.

[0016] The advantages and effects of the present invention are: The present invention adopts a probability calculation model based on a Bayesian network, combines vehicle dynamic information and environmental perception data, and uses a Kalman filter algorithm to update the speed and direction of the target object, thereby achieving comprehensive monitoring of the environment around the electric unmanned vehicle, improving the accuracy and real-time performance of collision warnings, and ensuring the safety of vehicle driving by predicting potential collision risks in advance. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 The present invention is a flowchart of a collision warning method for an electric unmanned vehicle.

[0019] Figure 2 It is a schematic diagram of the steps of constructing a collision probability model of the present invention.

[0020] Figure 3 It is a schematic structural diagram of a collision warning device for an electric unmanned vehicle of the present invention.

[0021] Figure 4 The present invention is a schematic block diagram of the collision warning electronic device for an electric unmanned vehicle. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] The present invention provides a collision warning method for an electric unmanned vehicle, the method is as follows: Figure 1 As shown, the following steps are included: Step S10: Acquire environmental information data and road condition information data around the vehicle in real time through sensors installed on the electric driverless vehicle.

[0024] Sensors refer to a variety of sensors installed on electric driverless vehicles, including cameras, lidar, millimeter-wave radar and ultrasonic sensors; the environmental information data around the vehicle includes information such as the position, speed and shape of other vehicles, pedestrians and obstacles; the road condition information data on the vehicle includes information such as vehicle congestion, visibility, weather conditions, etc.

[0025] The camera is a common vehicle-mounted camera and may also have an infrared night vision searchlight function. It is mainly used for driving records and photographing the vehicle's surroundings. The laser radar has high resolution and long-distance detection capabilities. It is mainly used for vehicles to detect long-distance obstacles, other vehicles, pedestrians and animals, so that the vehicle control system can make timely collision warnings. The millimeter-wave radar has strong penetration and is suitable for severe weather conditions. When encountering severe weather conditions such as heavy rain, blizzards, dust and fog, the strong penetration of the millimeter-wave radar can detect obstacles, other vehicles, pedestrians and animals that may appear in front of the vehicle, collect target motion data in time, calculate the predicted trajectory of the target and track the motion trajectory of the target object, and then calculate the probability of collision between the vehicle and the target, and issue a timely warning when there is a risk of collision. The ultrasonic sensor uses a sound wave with a frequency higher than the human hearing range (usually greater than 20 kHz), usually called ultrasonic wave. The sensor emits ultrasonic pulses and receives the reflected echo to measure the distance. It is widely used in various occasions, especially in autonomous driving vehicles and robotics technology, for detecting obstacles, measuring distance or environmental perception. By comprehensively utilizing the above devices, the stability and accuracy of data acquisition of electric driverless vehicles can be significantly improved, so that the collision probability model can more accurately calculate the collision probability and issue warnings in a timely manner.

[0026] Step S20: Preprocessing the environmental information data and road condition information data collected in real time, including removing noise, filling missing values ​​and extracting features, to obtain valid data that can be used for subsequent analysis.

[0027] Noise refers to irrelevant, meaningless or erroneous information in the data. In data preprocessing, some methods are needed to remove noise to prevent negative impact on the model. Common noise removal methods include smoothing, filtering and outlier detection. Smoothing is to reduce the volatility of data by averaging, weighted averaging or sliding window operations. For example, moving average or exponential smoothing can be used to smooth time series data; filtering is to remove noise components in the frequency domain through filters. Common filtering methods include low-pass filtering, high-pass filtering and band-pass filtering. For example, Fourier transform can be used to convert signals to the frequency domain for processing; outliers refer to data points that are significantly different from other observations. Statistical methods (such as the 3σ principle) or distance and density-based algorithms (such as LOF, DBSCAN) can be used to detect and remove outliers.

[0028] Missing values ​​refer to the situation where some variables or observations in the data are missing. Before conducting data analysis, missing values ​​need to be processed to avoid affecting the training and prediction capabilities of the model. Common methods for filling missing values ​​include deletion, interpolation, and model prediction. Deletion means that if the proportion of missing values ​​is very small and randomly distributed, you can choose to directly delete samples or features with missing values; interpolation means estimating missing values ​​through some methods and replacing them with estimated values. Common interpolation methods include mean interpolation, median interpolation, mode interpolation, and regression interpolation; model prediction means that if the number of missing values ​​is large, other features can be used as prediction targets to build a model, and then the model can be used to predict and fill in the missing values.

[0029] Feature extraction is to convert the original data into a more representative feature representation to facilitate subsequent model training and prediction. Common feature extraction methods include principal component analysis (PCA), linear discriminant analysis (LDA), feature selection and feature construction. Principal component analysis is an unsupervised dimensionality reduction method that maps high-dimensional data to low-dimensional space through linear transformation and retains the most important features; linear discriminant analysis is a supervised dimensionality reduction method that finds the best projection direction so that the projection points of samples of the same type are as close as possible and the projection points of samples of different types are as far apart as possible; feature selection is to select a feature subset from the original features that has a better predictive ability for the target variable. Common feature selection methods include filtering, wrapping and embedded.

[0030] Step S30: Calculate the predicted driving trajectory of the vehicle based on the obtained valid data, and track the motion trajectory of the target object.

[0031] The predicted driving trajectory of the vehicle is calculated by formula (1): (1) in, x ( t ) is the time at which the vehicle t The position at x 0 is the initial position, v x is the initial velocity of the vehicle, a x is the acceleration of the vehicle.

[0032] The process of tracking the motion trajectory of the target object includes: Target tracking: Use multi-target tracking algorithms to track surrounding dynamic objects; Target correction: Update the speed and direction of the target object through the Kalman filter algorithm.

[0033] Among them, the Multi-Object Tracking (MOT) algorithm refers to an algorithm that continuously tracks and identifies multiple targets in a video sequence. Its goal is to accurately detect and track all targets of interest from the video captured by the on-board camera, and re-identify the target when it is lost or occluded. The multi-object tracking algorithm can be divided into two main steps: target detection and target association.

[0034] Object detection refers to finding the location of the object of interest in each frame of a video sequence by using an object detection algorithm. Commonly used object detection algorithms include deep learning-based methods such as Faster R-CNN, YOLO, and SSD, as well as traditional feature extraction and machine learning methods such as HOG+SVM. Target association refers to matching and associating targets detected in adjacent frames to form trajectories, which usually includes target matching and trajectory management. Target matching aims to determine the correspondence between the target detected in the current frame and the target that already existed in the previous frame. Common target matching methods include Kalman filtering, Hungarian algorithm, association filter and correlation filter. Trajectory management is responsible for maintaining the trajectory information of the tracked target and predicting and updating it according to the target's motion model. Common trajectory management methods include Kalman filtering, particle filtering and multi-hypothesis tracking.

[0035] In the target correction step, the update formula of the Kalman filter algorithm used is shown in formula (2): (2) in, is the current estimated state, is the Kalman gain, is the measured value, H is the measurement matrix, is the estimated state at the previous moment.

[0036] Step S40: Calculate the probability of collision using a collision probability model based on the calculation result of step S30 and the tracked motion trajectory of the target object.

[0037] The collision probability model uses a Bayesian network model, which integrates multiple sensors and evaluates the probability of collision. The steps to construct the collision probability model are as follows: Figure 2 As shown, including: Define variables: Identify variables that can affect the probability of a collision, including sensor data, environmental information, vehicle status, and obstacle information; Formulate hypothesis: Make assumptions about the relationship between variables based on the actual situation and problem requirements; Determine the conditional probability table: Based on the selected variables and assumptions, construct a conditional probability table to describe the dependencies between variables; Build Bayesian network structures: Use Bayesian network graphs to visualize variables and their dependencies; Parameter learning: Using existing data, the parameters of the constructed Bayesian network are estimated through maximum likelihood estimation or Bayesian Learning; Reasoning and prediction: Based on the existing data and learned parameters, reasoning and prediction are performed, and the probability of a collision at a certain point in the future is predicted based on the real-time data from the sensor; Update and optimize: After accumulating more data, the parameters of the Bayesian network are continuously updated.

[0038] Preferably, in the continuously updating the parameters of the Bayesian network, the updating formula of the Bayesian network parameters is as shown in formula (3): (3) in, P ( Collision | Data ) represents the collision probability under the given vehicle surrounding environment information data collected by the sensor, P ( Data | Collision ) represents the conditional probability of the vehicle surrounding environment information data collected when a collision occurs, P ( Collision ) is the prior probability of collision, P ( Data ) represents the prior probability of the information data of the vehicle's surrounding environment collected by the sensor.

[0039] Step S50: When the calculated collision probability exceeds a preset threshold, a warning signal is sent to the vehicle control system.

[0040] Determining the preset threshold is an important decision and needs to be adjusted according to specific situations and needs. The present invention adjusts the threshold through acceptable risk level, performance balance and data-driven.

[0041] The acceptable risk level refers to the threshold value determined according to the acceptable degree of risk in the application scenario. This needs to take into account the consequences and impacts of possible collisions. For example, in a road traffic system, vehicles on highways may allow for a lower collision probability threshold, while a higher threshold is required in urban areas or densely populated areas; Performance balance refers to determining the threshold by weighing the accuracy and false alarm rate of the alarm. Too frequent alarms may cause the driver to ignore the real danger, while too conservative alarms will increase the false alarm rate. Therefore, indicators such as ROC curves, precision and recall are used to evaluate the performance under different thresholds, and the appropriate threshold is selected to achieve a balance; Data-driven means using existing data to determine the threshold, analyzing historical data and testing based on real scenarios to obtain performance under different collision probabilities, and selecting the appropriate threshold based on demand.

[0042] It should be understood that the preset threshold is not fixed and may be dynamically adjusted according to factors such as application environment, system performance requirements and user feedback. Therefore, in actual application, the present invention monitors and adjusts the threshold according to real-time conditions to ensure the accuracy and reliability of the system.

[0043] Step S60: After receiving the warning signal, the vehicle control system takes corresponding emergency braking measures to avoid collision.

[0044] After the vehicle control system confirms that there is a potential risk of collision, it will calculate the required braking force and send corresponding instructions to the braking system. The calculation of the required braking force involves factors such as the dynamic model, road friction coefficient, and vehicle mass, which need to be considered comprehensively. After receiving the emergency braking instruction, the braking system immediately performs the corresponding braking operation, which may trigger the anti-lock braking system (ABS) or electronic stability control system (ESC) to ensure vehicle stability and optimal braking effect. In order to further optimize the braking effect, the vehicle control system may adjust the braking force distribution of each wheel, and through the dynamic brake force distribution system (EBD) or traction control system (TCS), it can make real-time adjustments according to different road conditions and vehicle dynamic characteristics. At the same time, the vehicle control system will also send a warning signal to the driver, reminding him of the occurrence of an emergency through visual, sound and other means, which helps to attract the driver's attention and enable him to make an appropriate response.

[0045] In addition, the present invention also proposes a collision warning device for an electric unmanned vehicle, please refer to Figure 3 , the collision warning device for an electric unmanned vehicle comprises: Data acquisition module: used to obtain environmental information data around the vehicle in real time through sensors; Data preprocessing module: used to preprocess the collected environmental information data to obtain valid data that can be used for subsequent analysis; Trajectory prediction and tracking module: used to calculate the predicted trajectory of the vehicle based on the valid data obtained, and track the motion trajectory of the target object; Collision probability calculation module: used to calculate the probability of collision using the collision probability model based on the calculation results and the tracked target object motion trajectory; Warning module: used to send a warning signal to the vehicle control system when the calculated collision probability exceeds a preset threshold; Warning response module: After receiving the warning signal, the vehicle control system takes corresponding emergency braking measures to avoid collision.

[0046] The collision warning device for an electric unmanned vehicle provided by the present application adopts the collision warning method for an electric unmanned vehicle in the above-mentioned embodiment, which can solve the technical problem that the traditional collision warning method cannot meet the accuracy and real-time performance of the collision warning. Compared with the prior art, the beneficial effects of the collision warning device for an electric unmanned vehicle provided by the present application are the same as the beneficial effects of the collision warning method for an electric unmanned vehicle provided by the above-mentioned embodiment, and the other technical features of the collision warning device for an electric unmanned vehicle are the same as the features disclosed in the above-mentioned embodiment method, which will not be described in detail here.

[0047] The present application provides a collision warning device for an electric unmanned vehicle, the collision warning device for an electric unmanned vehicle comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the collision warning method for the electric unmanned vehicle in the above-mentioned embodiment one.

[0048] Reference below Figure 4 , which shows a schematic diagram of the structure of a collision warning device for an electric unmanned vehicle suitable for implementing an embodiment of the present application. The collision warning device for an electric unmanned vehicle in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Pesonal Digital Assistants: Personal Digital Assistants), PADs (Portable Application Descriptions: Tablet Computers), PMPs (Portable Media Players: Portable Multimedia Players), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 4 The collision warning device for the electric driverless vehicle shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0049] Figure 4The collision warning device of the electric unmanned vehicle shown may include a processing device 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. Various programs and data required for the operation of the collision warning device of the electric unmanned vehicle are also stored in RAM1004. The processing device 1001, ROM1002, and RAM1004 are connected to each other via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 1003 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 1009. The communication device 1009 can allow the collision warning device of the electric unmanned vehicle to communicate with other devices wirelessly or by wire to exchange data. Although the collision warning device of the electric unmanned vehicle with various systems is shown in the figure, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems may be implemented or have alternatively.

[0050] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0051] The collision warning device for an electric unmanned vehicle provided by the present application adopts the collision warning method for an electric unmanned vehicle in the above-mentioned embodiment, which can solve the technical problem that the traditional collision warning method cannot meet the accuracy and real-time requirements of the collision warning. Compared with the prior art, the beneficial effects of the collision warning device for an electric unmanned vehicle provided by the present application are the same as the beneficial effects of the collision warning method for an electric unmanned vehicle provided by the above-mentioned embodiment, and the other technical features of the collision warning device for an electric unmanned vehicle are the same as the features disclosed in the method of the previous embodiment, which will not be described in detail here.

[0052] The various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0053] The present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the collision warning method for an electric unmanned vehicle as described above.

[0054] The computer program product provided by the present application can solve the technical problem that the traditional collision warning method cannot meet the accuracy and real-time requirements of the collision warning. Compared with the prior art, the beneficial effects of the computer program product provided by the present application are the same as the beneficial effects of the collision warning method for an electric unmanned vehicle provided by the above embodiment, and will not be elaborated here.

[0055] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A collision warning method for an electric unmanned vehicle, characterized in that: The method comprises the following steps: Step S10: Acquire environmental information data around the vehicle in real time through sensors; Step S20: pre-processing the collected environmental information data to obtain valid data that can be used for subsequent analysis; Step S30: Calculate the predicted driving trajectory of the vehicle according to the obtained valid data, and track the motion trajectory of the target object; Step S40: Calculate the probability of collision using a collision probability model according to the calculation result of step S30 and the tracked motion trajectory of the target object; Step S50: When the calculated collision probability exceeds a preset threshold, a warning signal is sent to the vehicle control system; Step S60: After receiving the warning signal, the vehicle control system takes corresponding emergency braking measures to avoid collision; Among them, the sensors refer to a variety of sensors installed on electric unmanned vehicles, including cameras, lidars, millimeter-wave radars and ultrasonic sensors; the environmental information data around the vehicle includes the location, speed and shape of obstacles.

2. The collision warning method for an electric unmanned vehicle according to claim 1, characterized in that: The data preprocessing step in step S20 includes noise removal, missing value filling and feature extraction.

3. The collision warning method for an electric unmanned vehicle according to claim 1, characterized in that: In step S30, the predicted driving trajectory of the vehicle is calculated by formula (1): （1） in, x ( t ) is the time at which the vehicle t The position at x 0 is the initial position, v x is the initial velocity, a x is the acceleration.

4. The collision warning method for an electric unmanned vehicle according to claim 1, characterized in that: The process of tracking the motion trajectory of the target object in step S30 includes: Target tracking: Use multi-target tracking algorithms to track surrounding dynamic objects; Target correction: Update the speed and direction of the target object through the Kalman filter algorithm.

5. The collision warning method for an electric unmanned vehicle according to claim 4, characterized in that: The update formula of the Kalman filter algorithm is shown in formula (2): （2） in, is the current estimated state, is the Kalman gain, is the measured value, H is the measurement matrix, is the estimated state at the previous moment.

6. The collision warning method for an electric unmanned vehicle according to claim 1, characterized in that: The collision probability model in step S40 adopts a Bayesian network model, which integrates information data of the vehicle's surrounding environment collected by various sensors through the Bayesian network and evaluates the probability of collision. The steps of constructing the collision probability model include: Define variables: Determine the variables that can affect the probability of a collision, including the vehicle's surrounding environment information, vehicle status, and obstacle information; Formulate hypotheses: Make assumptions about the relationships between variables based on the impact of the vehicle's speed and the location of obstacles on the vehicle's collision; Determine the conditional probability table: Based on the selected variables and assumptions, construct a conditional probability table to describe the dependencies between variables; Build Bayesian network structures: Use Bayesian network graphs to visualize variables and their dependencies; Parameter learning: Using existing data, the parameters of the constructed Bayesian network are estimated through maximum likelihood estimation or Bayesian Learning; Reasoning and prediction: Based on the existing data and learned parameters, reasoning and prediction are performed, and the probability of a collision at a certain point in the future is predicted based on the information data of the vehicle's surrounding environment collected by sensors; Update and Optimize: Update the parameters in the Bayesian network.

7. The collision warning method for an electric unmanned vehicle according to claim 4, characterized in that: When updating the parameters in the Bayesian network, the update formula used is as shown in formula (3): （3） in, P ( Collision | Data ) represents the collision probability under the given vehicle surrounding environment information data collected by the sensor, P ( Data | Collision ) represents the conditional probability of the vehicle surrounding environment information data collected by the sensor when a collision occurs, P ( Collision ) is the prior probability of collision, P ( Data ) represents the prior probability of the information data of the vehicle's surrounding environment collected by the sensor.

8. A collision warning device for an electric unmanned vehicle, characterized in that: Data acquisition module: used to obtain environmental information data around the vehicle in real time through sensors; Data preprocessing module: used to preprocess the collected environmental information data to obtain valid data that can be used for subsequent analysis; Trajectory prediction and tracking module: used to calculate the predicted trajectory of the vehicle based on the valid data obtained, and track the motion trajectory of the target object; Collision probability calculation module: used to calculate the probability of collision using the collision probability model based on the calculation results and the tracked target object motion trajectory; Warning module: used to send a warning signal to the vehicle control system when the calculated collision probability exceeds a preset threshold; Warning response module: After receiving the warning signal, the vehicle control system takes corresponding emergency braking measures to avoid collision.

9. A collision warning device for an electric unmanned vehicle, characterized in that: The collision warning device for an electric unmanned vehicle comprises: A memory, a processor, and a collision warning program for an electric unmanned vehicle stored in the memory and executable on the processor, wherein when the collision warning program for the electric unmanned vehicle is executed by the processor, a collision warning method for an electric unmanned vehicle as described in any one of claims 1 to 7 is implemented.

10. A computer program product, characterized in that The computer program product includes a collision warning program for an electric unmanned vehicle, and when the collision warning program for an electric unmanned vehicle is executed by a processor, a collision warning method for an electric unmanned vehicle as described in any one of claims 1 to 7 is implemented.

Images