Battery car anti-collision early warning method and system based on intelligent algorithm

Abstract

The invention relates to a battery car anti-collision early warning method and system based on an intelligent algorithm. The method comprises the steps that a dynamic image and a static image of an obstacle around a battery car and the distance between the battery car and the obstacle are obtained through image obtaining equipment and an ultrasonic sensor; and taking the environmental data as the input of the FusionNet multilayer perceptron model to call the FusionNet multilayer perceptron model to learn and analyze the environmental data so as to identify the obstacle and predict the moving path of the obstacle. And calling a dynamic risk assessment algorithm based on Bayesian statistics to calculate a risk assessment result of collision between the battery car and the obstacle according to the moving path of the obstacle, and adjusting a collision early warning level and obtaining an anti-collision measure according to the risk assessment result. And when the collision early warning level exceeds the preset level, an early warning signal is sent based on the anti-collision measures, and a reaction mechanism of the battery car is started, so that the battery car is prevented from colliding with the obstacle.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent transportation and battery vehicle anti-collision warning, and in particular to an intelligent algorithm-based battery vehicle anti-collision warning method and system. Background Art

[0002] With the increase of urban traffic density and the improvement of road safety awareness, the safety of electric vehicles as a common means of transportation has received increasing attention. At present, the intelligent anti-collision technologies for electric vehicles on the market mainly include camera-based visual recognition systems, radar and ultrasonic distance sensing systems, and intelligent driving assistance systems integrating multi-sensor data fusion. These systems can improve driving safety and reduce accidents to a certain extent.

[0003] However, despite the increasing maturity of battery vehicle anti-collision warning technology, the anti-collision warning system of battery vehicles still faces many challenges. For example, the existing system often behaves unstable in complex traffic environments (such as rainy days, foggy days and at night), and false alarms and missed alarms occur from time to time. In addition, these systems often need to rely on high-power, high-cost hardware devices, such as high-resolution cameras and high-precision radars, which to a certain extent increase the economic burden and maintenance cost of the anti-collision warning system. In addition, with the rapid development of artificial intelligence technology, the future anti-collision system of battery vehicles is expected to be more intelligent and precise. Through deep learning and big data analysis technology, the anti-collision warning system can learn and predict potential dangers more effectively, so as to respond in advance. At the same time, with the application of Internet of Things technology, the anti-collision system of battery vehicles may have closer data interaction with the urban traffic management system, and obtain traffic signals and road conditions information in real time to further improve the accuracy and response speed of anti-collision warning.

[0004] At present, the existing anti-collision warning methods for electric vehicles widely use ultrasonic sensors, radars, cameras and other devices to detect and track obstacles around the electric vehicles. Ultrasonic sensors can provide accurate distance measurement, but their detection range and accuracy are limited when driving at high speeds or in complex environments. Although radar sensors provide good range and speed data, they are expensive and increase the complexity of the anti-collision warning system. Cameras provide rich visual information and are suitable for obstacle identification and classification in complex environments, but they are highly dependent on lighting conditions and the algorithms for processing video data are relatively complex, which in turn increases the demand for computing resources. In addition, the existing anti-collision warning methods for electric vehicles usually use image processing and machine learning algorithms to analyze sensor data. Although they perform well in static images or low-dynamic environments, the accuracy of their anti-collision warnings in high-dynamic or more complex traffic environments needs to be further improved, and their intelligence level is low.

[0005] In summary, the existing electric vehicle collision avoidance warning method has high operation and maintenance costs due to its complex internal algorithm and high hardware and computing resource requirements, and the accuracy and intelligence of the collision avoidance warning in highly dynamic or complex traffic environments need to be further improved. Summary of the invention

[0006] Based on this, it is necessary to provide an electric vehicle anti-collision warning method and system based on an intelligent algorithm with low operation and maintenance cost and high accuracy and intelligence of anti-collision warning in dynamic or complex traffic environments to address the above technical problems.

[0007] The present invention provides a battery vehicle anti-collision warning method based on an intelligent algorithm, the method comprising:

[0008] Acquire environmental data around the battery vehicle through an image acquisition device and an ultrasonic sensor, wherein the environmental data includes dynamic images and static images of obstacles around the vehicle body and the distance between the battery vehicle and the obstacles;

[0009] Using the environmental data as input of a FusionNet multi-layer perceptron model, so as to call the FusionNet multi-layer perceptron model to learn and analyze the environmental data, so as to identify the obstacle and predict the movement path of the obstacle;

[0010] Calling a dynamic risk assessment algorithm based on Bayesian statistics to calculate a risk assessment result of a collision between the electric vehicle and the obstacle according to the moving path of the obstacle, and adjusting the collision warning level and obtaining anti-collision measures according to the risk assessment result;

[0011] When the collision warning level exceeds the preset level, a warning signal is issued based on the anti-collision measure and a reaction mechanism of the battery vehicle is activated to prevent the battery vehicle from colliding with the obstacle;

[0012] Among them, the risk assessment result includes the risk probability of the electric vehicle colliding with the obstacle, the collision warning level is obtained by dividing according to the risk probability, and the higher the risk probability, the higher the collision warning level, the anti-collision measures include the warning signal and the reaction mechanism, and the reaction mechanism includes controlling the electric vehicle to decelerate and cutting off the throttle of the electric vehicle.

[0013] In one embodiment, the step of acquiring environmental data around the battery vehicle by using an image acquisition device and an ultrasonic sensor includes:

[0014] Acquire dynamic images and static images of obstacles in a first space outside the battery vehicle through the image acquisition device, and acquire the real-time distance between the battery vehicle and each obstacle in the first space through the ultrasonic sensor;

[0015] Preprocessing the static image, dynamic image and real-time distance so that the environmental data meets the input requirements of the FusionNet multi-layer perceptron model;

[0016] Among them, the image acquisition device is a camera, and the camera and the ultrasonic sensor are fixedly arranged on the body of the electric vehicle, the first space is the monitoring range of the image acquisition device and the ultrasonic sensor, and the preprocessing includes cleaning, deduplication, denoising and blank value filling.

[0017] In one of the embodiments, the environmental data also includes the spatial position coordinates of the obstacle in three-dimensional space;

[0018] The method of acquiring environmental data around the battery vehicle by using an image acquisition device and an ultrasonic sensor also includes:

[0019] Acquire the pixel coordinates of the obstacle in the two-dimensional space based on the static image and the dynamic image, and calculate the spatial position coordinates of the obstacle in the three-dimensional space by combining the real-time distance and the calibration information of the camera through geometric transformation;

[0020] The calibration information includes the focal length of the camera and the relative position between the camera and the obstacle.

[0021] In one embodiment, the using the environmental data as an input of a FusionNet multi-layer perceptron model to call the FusionNet multi-layer perceptron model to learn and analyze the environmental data to identify the obstacle and predict the movement path of the obstacle includes:

[0022] Inputting the environmental data vector corresponding to the preprocessed environmental data into the FusionNet multi-layer perceptron model, and calling the activation function in the FusionNet multi-layer perceptron model to process the environmental data vector in combination with network parameters and deviations to output obstacle types and movement trajectories;

[0023] The FusionNet multi-layer perceptron model is trained by a back propagation algorithm, so as to call the back propagation algorithm to update the network parameters and deviations by calculating the error between the predicted output of the model and the actual output;

[0024] The activation function includes a ReLU function and a Sigmoid function, and the network parameter is a weight matrix, which is used to convert input data into useful output.

[0025] In one embodiment, the calling of a dynamic risk assessment algorithm based on Bayesian statistics to calculate a risk assessment result of a collision between the electric vehicle and the obstacle according to the moving path of the obstacle, and adjusting the collision warning level and obtaining anti-collision measures according to the risk assessment result include:

[0026] Calling a Bayesian statistical algorithm to calculate the collision risk probability of the current environmental data based on the collision probability statistics of the historical environmental data, and updating the collision risk probability to adjust the risk assessment result;

[0027] Dividing the collision risk probability into a plurality of intervals according to the numerical value of the collision risk probability, so as to adjust the collision warning level, and obtain a plurality of different collision warning levels corresponding to different collision risk probability intervals;

[0028] Each interval corresponds to one of the collision warning levels, and each of the collision warning levels corresponds to one of the collision prevention measures.

[0029] In one embodiment, the collision warning level includes at least a first warning level, a second warning level and a third warning level, and the collision risk probability interval corresponding to the first warning level is greater than the collision risk probability interval corresponding to the second warning level, and the collision risk probability interval corresponding to the second warning level is greater than the collision risk probability interval corresponding to the third warning level;

[0030] When the collision warning level exceeds the preset level, a warning signal is issued based on the anti-collision measure and a reaction mechanism of the battery vehicle is activated to prevent the battery vehicle from colliding with the obstacle, including:

[0031] When the collision warning level is at the first warning level, a first warning signal is issued and a first control instruction is received to notify the driver and control the battery vehicle to cut off the throttle of the battery vehicle in response to the first control instruction;

[0032] When the collision warning level is at the second warning level, a second warning signal is issued and a second control instruction is received to notify the driver and control the battery vehicle to decelerate in response to the second control instruction;

[0033] When the collision warning level is at the third warning level, a third warning signal is issued to inform the driver to avoid collision.

[0034] In one embodiment, the battery vehicle is provided with an indicator light, and the indicator light is used to remind the driver to pay attention to collision avoidance, and the method further includes:

[0035] When the collision warning level is at the first warning level, in response to the first warning signal, controlling the flashing frequency of the indicator light to be a first frequency;

[0036] When the collision warning level is at the second warning level, in response to the second warning signal, controlling the flashing frequency of the indicator light to be a second frequency;

[0037] When the collision warning level is at the third warning level, in response to the third warning signal, controlling the flashing frequency of the indicator light to be a third frequency;

[0038] Among them, the first frequency is greater than the second frequency, the second frequency is greater than the third frequency, and the greater the flashing frequency of the indicator light, the closer the electric vehicle is to the obstacle, and the higher the probability of collision risk.

[0039] The present invention also provides a battery vehicle anti-collision warning system based on an intelligent algorithm, the system comprising:

[0040] A data acquisition module, used to acquire environmental data around the battery vehicle through an image acquisition device and an ultrasonic sensor, wherein the environmental data includes dynamic images and static images of obstacles around the vehicle body and the distance between the battery vehicle and the obstacles;

[0041] An obstacle recognition module, used to use the environmental data as an input of a FusionNet multi-layer perceptron model, so as to call the FusionNet multi-layer perceptron model to learn and analyze the environmental data, so as to identify the obstacle and predict the movement path of the obstacle;

[0042] A collision risk assessment module, which is used to call a dynamic risk assessment algorithm based on Bayesian statistics to calculate a risk assessment result of a collision between the electric vehicle and the obstacle according to the moving path of the obstacle, and to adjust the collision warning level and obtain anti-collision measures according to the risk assessment result;

[0043] An anti-collision warning module is used to send out a warning signal based on the anti-collision measures and start the reaction mechanism of the battery vehicle to prevent the battery vehicle from colliding with the obstacle when the collision warning level exceeds a preset level;

[0044] Among them, the risk assessment result includes the risk probability of the electric vehicle colliding with the obstacle, the collision warning level is obtained by dividing according to the risk probability, and the higher the risk probability, the higher the collision warning level, the anti-collision measures include the warning signal and the reaction mechanism, and the reaction mechanism includes controlling the electric vehicle to decelerate and cutting off the throttle of the electric vehicle.

[0045] The present invention also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, it implements any of the above-mentioned battery vehicle anti-collision warning methods based on intelligent algorithms.

[0046] The present invention also provides a computer storage medium storing a computer program, wherein when the computer program is executed by a processor, the battery vehicle anti-collision warning method based on the intelligent algorithm as described above is implemented.

[0047] The above-mentioned battery vehicle anti-collision warning method and system based on intelligent algorithm obtains dynamic images and static images of obstacles around the battery vehicle and the distance between the battery vehicle and the obstacles through image acquisition equipment and ultrasonic sensors, and then uses the obtained environmental data as the input of the FusionNet multi-layer perceptron model to call the FusionNet multi-layer perceptron model to learn and analyze the environmental data to identify obstacles and predict the movement path of obstacles. Subsequently, the dynamic risk assessment algorithm based on Bayesian statistics is called to calculate the risk assessment result of the battery vehicle colliding with the obstacle according to the movement path of the obstacle, and the collision warning level is adjusted according to the risk assessment result and the anti-collision measures are obtained. When the collision warning level exceeds the preset level, a warning signal is issued based on the anti-collision measures and the reaction mechanism of the battery vehicle is activated to control the battery vehicle to decelerate and cut off the throttle of the battery vehicle to prevent the battery vehicle from colliding with the obstacle. This method uses the FusionNet multi-layer perceptron model to analyze and identify multi-angle environmental data around the electric vehicle, and combines the Bayesian statistical dynamic risk assessment algorithm to evaluate the collision risk of the electric vehicle, so that the driver can perform anti-collision warning for the electric vehicle in different environments. The algorithm is relatively simple, which reduces the operation and maintenance costs to a certain extent. The anti-collision warning is accurate and intelligent in dynamic or complex traffic environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are 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.

[0049] Figure 1 One of the flow charts of the battery vehicle anti-collision warning method based on intelligent algorithm provided by the present invention;

[0050] Figure 2 A schematic diagram of the overall process of the battery vehicle anti-collision warning method for the battery vehicle based on the intelligent algorithm in the specific embodiment provided by the present invention;

[0051] Figure 3 A schematic diagram of the overall architecture of an electric vehicle anti-collision warning method for an electric vehicle based on an intelligent algorithm in a specific embodiment of the present invention;

[0052] Figure 4 The second flowchart of the battery vehicle anti-collision warning method based on the intelligent algorithm provided by the present invention;

[0053] Figure 5 The third flow chart of the battery vehicle anti-collision warning method based on the intelligent algorithm provided by the present invention;

[0054] Figure 6 The fourth flowchart of the battery vehicle anti-collision warning method based on the intelligent algorithm provided by the present invention;

[0055] Figure 7 The fifth flow chart of the battery vehicle anti-collision warning method based on the intelligent algorithm provided by the present invention;

[0056] Figure 8 The sixth flow chart of the battery vehicle anti-collision warning method based on the intelligent algorithm provided by the present invention;

[0057] Fig. 9 A schematic diagram of the structure of the battery vehicle anti-collision warning system based on the intelligent algorithm provided by the present invention;

[0058] Fig.10 This is a diagram of the internal structure of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are 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.

[0060] Combine the following Figures 1 to 10 The present invention describes an electric vehicle anti-collision warning method and system based on an intelligent algorithm.

[0061] like Figure 1 As shown, in one embodiment, a battery vehicle anti-collision warning method based on an intelligent algorithm includes the following steps:

[0062] Step S110, obtaining environmental data around the electric vehicle through an image acquisition device and an ultrasonic sensor, the environmental data including dynamic images and static images of obstacles around the vehicle body and the distance between the electric vehicle and the obstacles.

[0063] Specifically, the battery vehicle driving system obtains dynamic images and static images of obstacles around the battery vehicle body and the distance between the battery vehicle and the obstacles through an image acquisition device and an ultrasonic sensor.

[0064] Combination Figure 2 and Figure 3 As shown, in a specific embodiment, the battery vehicle anti-collision warning method based on intelligent algorithm provided by the present invention uses a camera and an ultrasonic sensor to fully capture the environmental information around the battery vehicle. Among them, the camera is responsible for recording visual information, such as dynamic and static images of pedestrians, vehicles and other obstacles, and the ultrasonic sensor is used to measure the distance between the battery vehicle and surrounding objects to provide accurate spatial positioning. The data of these sensors are collected in real time, providing a basis for subsequent data processing and risk assessment. Through this advanced perception technology, the battery vehicle driving system can understand the surrounding environment in detail, thereby improving the accuracy of obstacle recognition and the response speed of anti-collision warning.

[0065] Safety is a core issue in the design of electric vehicles, especially in complex traffic environments. In order to improve the safety of electric vehicles, this example designs a multi-dimensional perception data acquisition unit, which is mainly composed of cameras and ultrasonic sensors. These sensors are distributed on the front, side and rear of the electric vehicle, and can capture environmental information around the electric vehicle in all directions. Among them, the main function of the camera is to capture visual information, including vehicles, pedestrians, traffic signs and other potential obstacles. The camera can provide rich environmental visual data through real-time video streams. These visual data are essential for identifying the type and dynamic characteristics of obstacles. Ultrasonic sensors are used to measure the precise distance between the electric vehicle and surrounding objects. These sensors determine the distance by emitting ultrasonic waves and receiving their echoes, which is more effective for avoiding collisions and performing precise spatial positioning.

[0066] In the multi-dimensional perception data acquisition unit, the working principle of the ultrasonic sensor is based on the time ranging method of sound waves, and the expression is: D=v*t.

[0067] Where D is the distance from the sensor to the obstacle (unit: meter), v is the speed of sound waves in the air, usually taken as 343m / s (under standard atmospheric conditions), and t is the time it takes for the sound wave to be transmitted and received (unit: second).

[0068] This expression is the basis of ultrasonic ranging technology, which can provide real-time and accurate distance data. These distance data are crucial for electric vehicle driving systems because they can be used to determine the safe distance of the surrounding environment and make corresponding driving decisions.

[0069] In this embodiment, in order to achieve more accurate obstacle positioning and tracking, the visual data captured by the camera can be combined with the distance data of the ultrasonic sensor to perform positioning calculation in three-dimensional space. First, the two-dimensional position (pixel coordinates) of the obstacle is determined by the image captured by the camera, and then the distance data D provided by the ultrasonic sensor is used in combination with the calibration information of the camera (such as focal length, camera position, etc.) to calculate the position of the obstacle in three-dimensional space through geometric transformation.

[0070] In practical applications, the above-mentioned multi-dimensional perception data acquisition unit can significantly improve the safe driving capabilities of electric vehicles. For example, when driving on busy urban streets, the camera can identify and classify various dynamic and static obstacles (such as pedestrians, bicycles, parked vehicles, etc.), while the ultrasonic sensor provides the precise distance measurement required to avoid collisions. When the electric vehicle driving system detects that the obstacle is too close, it can automatically trigger the brakes or issue a warning through the control system to prevent or mitigate the occurrence of a collision. This integrated perception system is an indispensable part of the intelligent safety system of modern electric vehicles and is of great significance for achieving a highly automated and intelligent driving environment.

[0071] Step S120, using the environmental data as the input of the FusionNet multi-layer perceptron model to call the FusionNet multi-layer perceptron model to learn and analyze the environmental data, so as to identify obstacles and predict the movement path of the obstacles.

[0072] Specifically, the electric vehicle driving system uses the environmental data collected in step S110 as the input of the FusionNet multi-layer perceptron model to call the FusionNet multi-layer perceptron model to learn and analyze the environmental data, thereby identifying obstacles around the electric vehicle and predicting possible movement paths of the identified obstacles.

[0073] Combination Figure 2 and Figure 3 As shown, in a specific embodiment, the electric vehicle anti-collision warning method based on intelligent algorithm provided by the present invention, "FusionNet" is a multi-layer perceptron (MLP) model specially designed for the electric vehicle driving system, and the model includes an input layer, multiple hidden layers and an output layer. This structure enables the model to process and learn to extract features from complex input data, thereby making more accurate predictions.

[0074] In this embodiment, "FusionNet" performs forward propagation and learning and adjustment processes through the following steps and formulas:

[0075] Forward propagation process: the expression is y=f(W*x+b).

[0076] Where x is the data vector passed in from the multi-dimensional perception data acquisition unit, which contains key information such as distance, shape and speed collected from various sensors (such as cameras and ultrasonic sensors). W is the network parameter, that is, the weight matrix, which is used to convert input data into useful output. b is the deviation, which is used for bias adjustment in the network to help the model better fit the data. f is an activation function, such as ReLU (Rectified Linear Unit) or Sigmoid, which is used to add nonlinear factors to enable the network to learn and simulate complex functions. y is the output of the network, which includes the predicted obstacle type and movement trajectory.

[0077] It should be noted that the choice of activation function has a significant impact on the performance of the network model. For example, the ReLU function (f(z) = max(0,z)) can prevent the gradient vanishing problem when training deep networks due to its non-saturated form. The Sigmoid function (f(z) = 1 / (1+e^(-z))) is often used in the output layer of the model, especially when output probability predictions are required.

[0078] Backpropagation and weight adjustment: The network is trained using a backpropagation algorithm, which updates the weights and biases by calculating the error between the predicted output and the actual output and backpropagating this error. The error is usually calculated using mean squared error (MSE) or cross entropy loss.

[0079] Among them, the expression for weight update is: W = W-eta*dW.

[0080] Where W is the weight before update, eta is the learning rate, which determines the step size of weight update in each iteration, and dW is the gradient of weight, which is the partial derivative of loss function with respect to weight.

[0081] In practical applications, the efficient data processing capabilities of "FusionNet" enable electric vehicles to identify and predict obstacles in their driving paths in real time. For example, in urban traffic environments, the network can accurately identify static obstacles (such as parked vehicles) and dynamic obstacles (such as pedestrians and other vehicles) and predict their movement trajectories, which is crucial for decision-making systems such as automatic braking or warning drivers to take evasive measures. "FusionNet" can dynamically adapt to different driving conditions by continuously learning changes in the traffic environment. The training data of the network model comes from actual driving scenarios, including various weather conditions and traffic density, so that the model can maintain efficient and accurate collision avoidance warnings in various environments. In short, "FusionNet" enables the intelligent collision avoidance system of electric vehicles to perform better in processing multi-source data, performing complex calculations and making real-time responses.

[0082] Step S130, calling a dynamic risk assessment algorithm based on Bayesian statistics to calculate the risk assessment result of the battery vehicle colliding with the obstacle according to the moving path of the obstacle, and adjusting the collision warning level and obtaining anti-collision measures according to the risk assessment result.

[0083] Among them, the risk assessment results include the risk probability of a collision between an electric vehicle and an obstacle. The collision warning level is obtained according to the risk probability. The higher the risk probability, the higher the collision warning level. The anti-collision measures include warning signals and response mechanisms. The response mechanism includes controlling the electric vehicle to decelerate and cutting off the throttle of the electric vehicle.

[0084] Specifically, the electric vehicle driving system calls a dynamic risk assessment algorithm based on Bayesian statistics to calculate the risk assessment results of the electric vehicle colliding with an obstacle based on historical environmental data and the movement path of the obstacle, and adjusts the collision warning level according to the risk assessment results, while obtaining anti-collision measures corresponding to different collision warning levels.

[0085] Combination Figure 2 and Figure 3 As shown, in a specific embodiment, the present invention provides an electric vehicle anti-collision warning method based on an intelligent algorithm. In the electric vehicle intelligent anti-collision, the dynamic risk assessment algorithm is one of the cores, which is used to evaluate the risk of collision in real time and initiate corresponding response measures accordingly. Among them, the design of the algorithm is based on the Bayesian statistical method, combined with the real-time environmental data from the multi-dimensional perception data acquisition unit and the electric vehicle driving data, the risk assessment expression is: R = P (collision|v, D, t).

[0086] In the formula, R represents the assessment value of the collision risk, which is a value from 0 to 1, where 0 means no collision risk and 1 means that a collision is almost certain to occur. P(collision|v,D,t) is the probability of a collision under the conditions of a given speed v (current speed of the electric vehicle), distance D (distance to the nearest obstacle), and time t (estimated time to reach the obstacle). This probability is obtained through statistical analysis and historical data modeling.

[0087] It should be noted that the Bayesian method updates the estimate of the collision probability through prior knowledge (collision statistics based on historical data) and newly acquired evidence (real-time sensor data). By calculating the new posterior probability, the electric vehicle collision avoidance warning system can dynamically adjust the risk assessment of the current driving situation.

[0088] In this embodiment, based on the risk assessment results, the immediate response mechanism can take prompt action to prevent an impending collision or at least significantly reduce the severity of the collision.

[0089] The expression for deceleration control of electric vehicle is: a=-k*v.

[0090] Where a is the acceleration to be applied (actually deceleration, hence the minus sign), in meters per second squared. k is the reaction coefficient dynamically adjusted based on the risk assessment value R, a dimensionless coefficient used to adjust the intensity of the reaction. v is the current speed of the vehicle, in meters per second.

[0091] It should be noted that when the R value increases, it indicates that the risk of collision increases, the k value also increases accordingly, and the absolute value of a (deceleration) increases, so that the vehicle speed can be reduced more quickly to avoid collision or reduce the impact of collision. If the R value is very low, it indicates that the possibility of collision is small, and the k value will decrease, thereby reducing the deceleration measures of the battery vehicle and maintaining the smooth driving of the battery vehicle.

[0092] During the driving process of the electric vehicle, the dynamic risk assessment and immediate response algorithm unit continuously receives data from sensors, such as the speed and distance of the vehicle in front, the status of traffic signals, and the movement of pedestrians. The electric vehicle driving system uses this data to calculate the collision risk R in real time, and dynamically adjusts the vehicle's driving speed based on the risk assessment value. For example, when the electric vehicle quickly approaches a pedestrian crossing the street, the system will calculate a higher R value, and the immediate response mechanism will automatically increase the k value, thereby increasing the deceleration and quickly decelerating to avoid or mitigate possible collisions. This intelligent deceleration measure greatly improves the safety of electric vehicles in complex traffic environments. It not only improves the driving safety of electric vehicles, but also reduces traffic congestion by reducing the possibility of accidents and improves overall traffic efficiency.

[0093] Step S140, when the collision warning level exceeds the preset level, a warning signal is issued based on anti-collision measures and a reaction mechanism of the battery vehicle is activated to prevent the battery vehicle from colliding with an obstacle.

[0094] Specifically, when the collision warning level exceeds the preset level, it means that the probability of collision risk is high at this time. The electric vehicle driving system will issue a warning signal based on the preset anti-collision measures and activate the corresponding response mechanism of the electric vehicle to prevent the electric vehicle from colliding with obstacles.

[0095] Combination Figure 2 and Figure 3As shown, in a specific embodiment, the anti-collision warning method for electric vehicle based on intelligent algorithm provided by the present invention includes a speed sensor, a distance sensor, a control unit, a throttle control and a brake system. Among them, the function of the speed sensor is to monitor the speed of the electric vehicle in real time. This information is crucial for calculating the collision probability between the electric vehicle and the obstacle in front. The speed data not only helps to determine whether the vehicle can stop in time before the collision occurs, but also provides data support for evaluating the response measures required in emergency situations. The distance sensor is usually an ultrasonic sensor or a laser radar, which is used to measure the real-time distance between the electric vehicle and the obstacle in front. This data is the key to calculating the collision time and determining the timing of the warning. The accuracy of the distance sensor determines whether the warning system can accurately assess the collision risk. The control unit is the brain of the anti-collision warning system. It receives data from the speed sensor and the distance sensor, and calculates the collision probability of the obstacle in front based on this information. The control unit uses an algorithm to evaluate the upcoming situation and issues instructions based on the evaluation results. If the control unit determines that the collision risk is high, it will send corresponding control signals to the throttle control and brake systems.

[0096] The throttle control receives instructions from the control unit and decides whether to cut off the throttle according to the level of collision risk. In high-risk situations, the throttle control will disconnect the throttle switch and quickly reduce the vehicle speed to avoid a collision. Once it is safe, the control unit will instruct the throttle control to reactivate the throttle and resume vehicle driving. The brake system is activated when the control unit determines that emergency braking is required. The system can automatically perform deceleration or complete parking operations to ensure that the damage caused by the collision is minimized when a collision is inevitable. The response speed and efficiency of the brake system are crucial to collision avoidance safety. When the control unit determines that the collision risk reaches a certain level, the multi-channel warning interaction unit will warn the driver through visual, auditory and tactile channels. This multi-sensory warning method ensures that it can still attract attention when the driver is distracted or the environment is noisy, thereby improving the effectiveness of collision avoidance warnings.

[0097] The above-mentioned battery vehicle anti-collision warning method based on intelligent algorithm obtains dynamic images and static images of obstacles around the battery vehicle and the distance between the battery vehicle and the obstacles through image acquisition equipment and ultrasonic sensors, and then uses the obtained environmental data as the input of the FusionNet multi-layer perceptron model to call the FusionNet multi-layer perceptron model to learn and analyze the environmental data to identify obstacles and predict the movement path of obstacles. Subsequently, the dynamic risk assessment algorithm based on Bayesian statistics is called to calculate the risk assessment result of the battery vehicle colliding with the obstacle according to the movement path of the obstacle, and the collision warning level is adjusted according to the risk assessment result and the anti-collision measures are obtained. When the collision warning level exceeds the preset level, a warning signal is issued based on the anti-collision measures and the reaction mechanism of the battery vehicle is activated to control the battery vehicle to decelerate and cut off the throttle of the battery vehicle to prevent the battery vehicle from colliding with the obstacle. This method uses the FusionNet multi-layer perceptron model to analyze and identify multi-angle environmental data around the electric vehicle, and combines the Bayesian statistical dynamic risk assessment algorithm to evaluate the collision risk of the electric vehicle, so that the driver can perform anti-collision warning for the electric vehicle in different environments. The algorithm is relatively simple, which reduces the operation and maintenance costs to a certain extent. The anti-collision warning is accurate and intelligent in dynamic or complex traffic environments.

[0098] like Figure 4 As shown, in one embodiment, the battery vehicle anti-collision warning method based on the intelligent algorithm provided by the present invention obtains environmental data around the battery vehicle through an image acquisition device and an ultrasonic sensor, and specifically includes the following steps:

[0099] Step S112, obtaining dynamic images and static images of obstacles in a first space outside the battery vehicle body by an image acquisition device, and obtaining a real-time distance between the battery vehicle and each obstacle in the first space by an ultrasonic sensor.

[0100] Step S114, pre-processing the static image, dynamic image and real-time distance so that the environmental data meets the input requirements of the FusionNet multi-layer perceptron model.

[0101] Among them, the image acquisition device is a camera, and the camera and the ultrasonic sensor are fixedly installed on the body of the electric vehicle. The first space is the monitoring range of the image acquisition device and the ultrasonic sensor. The preprocessing includes cleaning, deduplication, denoising and blank value filling.

[0102] Step S116, obtaining pixel coordinates of the obstacle in the two-dimensional space based on the static image and the dynamic image, and calculating the spatial position coordinates of the obstacle in the three-dimensional space by combining the real-time distance and the calibration information of the camera through geometric transformation.

[0103] The calibration information includes the focal length of the camera and the relative position between the camera and the obstacle.

[0104] like Figure 5 As shown, in one embodiment, the battery vehicle anti-collision warning method based on the intelligent algorithm provided by the present invention uses environmental data as the input of the FusionNet multi-layer perceptron model to call the FusionNet multi-layer perceptron model to learn and analyze the environmental data to identify obstacles and predict the movement path of obstacles, specifically including the following steps:

[0105] Step S122, input the environmental data vector corresponding to the preprocessed environmental data into the FusionNet multi-layer perceptron model, and call the activation function in the FusionNet multi-layer perceptron model to process the environmental data vector in combination with network parameters and deviations to output the obstacle type and movement trajectory.

[0106] Step S124, training the FusionNet multi-layer perceptron model through a back propagation algorithm, so as to call the back propagation algorithm to update the network parameters and deviations by calculating the error between the model predicted output and the actual output.

[0107] Among them, the activation functions include ReLU function and Sigmoid function, and the network parameters are weight matrices, which are used to convert input data into useful output.

[0108] like Figure 6 As shown, in one embodiment, the battery vehicle anti-collision warning method based on the intelligent algorithm provided by the present invention calls a dynamic risk assessment algorithm based on Bayesian statistics to calculate the risk assessment result of the battery vehicle colliding with an obstacle according to the moving path of the obstacle, and adjusts the collision warning level and obtains anti-collision measures according to the risk assessment result, specifically including the following steps:

[0109] Step S132, calling the Bayesian statistical algorithm to calculate the collision risk probability of the current environmental data according to the collision probability statistics of the historical environmental data, and updating the collision risk probability to adjust the risk assessment result.

[0110] Step S134 , dividing the collision risk probability into a plurality of intervals according to the numerical value of the collision risk probability, so as to adjust the collision warning level and obtain a plurality of different collision warning levels corresponding to different collision risk probability intervals.

[0111] Each interval corresponds to a collision warning level, and each collision warning level corresponds to an anti-collision measure.

[0112] like Figure 7As shown, in one embodiment, the battery vehicle anti-collision warning method based on the intelligent algorithm provided by the present invention, when the collision warning level exceeds the preset level, sends out a warning signal based on the anti-collision measures and starts the reaction mechanism of the battery vehicle to prevent the battery vehicle from colliding with the obstacle, specifically comprising the following steps:

[0113] It should be noted that the collision warning level includes at least the first warning level, the second warning level and the third warning level, and the collision risk probability interval corresponding to the first warning level is greater than the collision risk probability interval corresponding to the second warning level, and the collision risk probability interval corresponding to the second warning level is greater than the collision risk probability interval corresponding to the third warning level.

[0114] Step S142, when the collision warning level is at the first warning level, a first warning signal is issued and a first control instruction is received to notify the driver and control the battery vehicle to cut off the throttle of the battery vehicle in response to the first control instruction.

[0115] Step S144, when the collision warning level is at the second warning level, a second warning signal is issued and a second control instruction is received to notify the driver and control the electric vehicle to decelerate in response to the second control instruction.

[0116] Step S146: when the collision warning level is at the third warning level, a third warning signal is issued to inform the driver to prevent collision.

[0117] like Figure 8 As shown, in one embodiment, the battery vehicle anti-collision warning method based on the intelligent algorithm provided by the present invention further includes the following steps:

[0118] Step S810: when the collision warning level is at the first warning level, in response to the first warning signal, the flashing frequency of the indicator light is controlled to be a first frequency.

[0119] Step S820: When the collision warning level is at the second warning level, in response to the second warning signal, the flashing frequency of the indicator light is controlled to be a second frequency.

[0120] Step S830: when the collision warning level is at the third warning level, in response to the third warning signal, the flashing frequency of the indicator light is controlled to be a third frequency.

[0121] Among them, the first frequency is greater than the second frequency, the second frequency is greater than the third frequency, and the greater the flashing frequency of the indicator light, the closer the electric vehicle is to the obstacle, and the higher the probability of collision risk.

[0122] Combination Figure 2 and Figure 3As shown, in a specific embodiment, the battery vehicle anti-collision warning method based on the intelligent algorithm provided by the present invention, the multi-channel warning interaction unit is a more critical part of the intelligent anti-collision, which is responsible for sending warning signals to the driver through three sensory channels of vision, hearing and touch according to the collision risk assessment value R provided by the dynamic risk assessment and immediate response algorithm unit. The design of this unit can ensure that the driver receives warnings in a timely and effective manner in a complex traffic environment, so as to take appropriate anti-collision measures. Among them, the visual signal is mainly realized by the flashing frequency of the LED indicator light, and its expression is: L = f(R).

[0123] Where L represents the flashing frequency of the LED light, in units of times / second. R is the collision risk assessment value obtained from the risk assessment algorithm unit, ranging from 0 (no risk) to 1 (high risk). f(R) is a function that converts the risk assessment value into the flashing frequency of the LED light. Usually, this function is linear or nonlinear, for example, f(R) = a + bR^c, where a, b, c are parameters for adjusting the sensitivity of the frequency response.

[0124] It should be noted that when the R value is low (i.e. the collision risk is low), the LED light flashes at a low frequency or does not flash to avoid unnecessary interference to the driver. When the R value increases, the flashing frequency increases to attract the driver's attention and remind him to pay attention to the potential collision risk.

[0125] In this embodiment, the auditory signal is mainly adjusted according to the R value through the volume and frequency of the alarm sound. For example, the volume of the alarm sound can be set to V = h(R), where h(R) is a function that depends on R and can be linear or nonlinear. An increase in volume and / or alarm tone helps to attract the driver's attention in higher-risk situations. The tactile signal is mainly converted into vibration intensity through the vibration intensity of the steering wheel or seat, T = g(R), which is also a function. The vibration signal is very effective in providing direct physical feedback to the driver, especially in noisy environments or when the driver's eyes are not on the road.

[0126] In actual driving situations, such as driving in fog or at night, visual signals may not be as obvious as during the day, and the importance of auditory and tactile signals becomes particularly prominent. Similarly, when driving at high speeds, the rapidly increasing LED flashing frequency and emergency alarm sound can effectively remind the driver that there are high-risk obstacles ahead and that he needs to slow down or take evasive measures immediately. In addition, it is particularly effective in congested urban traffic environments, and can accurately warn drivers to pay attention to pedestrians or bicycles that may suddenly rush out from all sides. By using a combination of visual, auditory and tactile signals, it can ensure that warnings can be noticed by drivers in various environments, thereby greatly improving driving safety. Therefore, this multi-channel warning method not only enhances the effectiveness of the warning, but also improves the driver's response speed to emergency situations.

[0127] The battery vehicle anti-collision warning system based on intelligent algorithm provided by the present invention is described below. The battery vehicle anti-collision warning system based on intelligent algorithm described below and the battery vehicle anti-collision warning method based on intelligent algorithm described above can be referenced to each other.

[0128] like Fig. 9 As shown, in one embodiment, a battery vehicle anti-collision warning system based on an intelligent algorithm includes a data acquisition module 910 , an obstacle recognition module 920 , a collision risk assessment module 930 and an anti-collision warning module 940 .

[0129] The data acquisition module 910 is used to acquire environmental data around the battery vehicle through an image acquisition device and an ultrasonic sensor. The environmental data includes dynamic images and static images of obstacles around the vehicle body and the distance between the battery vehicle and the obstacles.

[0130] The obstacle recognition module 920 is used to use the environmental data as the input of the FusionNet multi-layer perceptron model to call the FusionNet multi-layer perceptron model to learn and analyze the environmental data to identify obstacles and predict the movement path of the obstacles.

[0131] The collision risk assessment module 930 is used to call a dynamic risk assessment algorithm based on Bayesian statistics to calculate the risk assessment result of the collision between the electric vehicle and the obstacle according to the moving path of the obstacle, and adjust the collision warning level and obtain anti-collision measures according to the risk assessment result.

[0132] The anti-collision warning module 940 is used to send out a warning signal based on anti-collision measures and activate the reaction mechanism of the battery vehicle when the collision warning level exceeds a preset level to prevent the battery vehicle from colliding with an obstacle.

[0133] Among them, the risk assessment results include the risk probability of a collision between an electric vehicle and an obstacle. The collision warning level is obtained according to the risk probability. The higher the risk probability, the higher the collision warning level. The anti-collision measures include warning signals and response mechanisms. The response mechanism includes controlling the electric vehicle to decelerate and cutting off the throttle of the electric vehicle.

[0134] In this embodiment, the battery vehicle anti-collision warning system based on intelligent algorithm provided by the present invention, the data acquisition module 910 is specifically used for:

[0135] The dynamic image and static image of the obstacles in the first space outside the battery vehicle body are obtained by the image acquisition device, and the real-time distance between the battery vehicle and each obstacle in the first space is obtained by the ultrasonic sensor.

[0136] Static images, dynamic images, and real-time distances are preprocessed so that the environmental data meets the input requirements of the FusionNet multi-layer perceptron model.

[0137] Among them, the image acquisition device is a camera, and the camera and the ultrasonic sensor are fixedly installed on the body of the electric vehicle. The first space is the monitoring range of the image acquisition device and the ultrasonic sensor. The preprocessing includes cleaning, deduplication, denoising and blank value filling.

[0138] In this embodiment, the battery vehicle anti-collision warning system based on intelligent algorithm provided by the present invention, the data acquisition module 910 is specifically used for:

[0139] The pixel coordinates of obstacles in two-dimensional space are obtained based on static images and dynamic images, and the spatial position coordinates of obstacles in three-dimensional space are calculated through geometric transformation combined with real-time distance and camera calibration information.

[0140] The calibration information includes the focal length of the camera and the relative position between the camera and the obstacle.

[0141] In this embodiment, the battery vehicle anti-collision warning system based on the intelligent algorithm provided by the present invention, the obstacle recognition module 920 is specifically used for:

[0142] The environmental data vector corresponding to the preprocessed environmental data is input into the FusionNet multi-layer perceptron model, and the activation function in the FusionNet multi-layer perceptron model is called to process the environmental data vector in combination with network parameters and deviations to output the obstacle type and movement trajectory.

[0143] The FusionNet multi-layer perceptron model is trained through the back-propagation algorithm, which calls the back-propagation algorithm to update the network parameters and deviations by calculating the error between the model's predicted output and the actual output.

[0144] Among them, the activation functions include ReLU function and Sigmoid function, and the network parameters are weight matrices, which are used to convert input data into useful output.

[0145] In this embodiment, the battery vehicle anti-collision warning system based on the intelligent algorithm provided by the present invention, the collision risk assessment module 930 is specifically used for:

[0146] The Bayesian statistical algorithm is called to calculate the collision risk probability of the current environmental data based on the collision probability statistics of the historical environmental data, and the collision risk probability is updated to adjust the risk assessment result.

[0147] The collision risk probability is divided into multiple intervals according to the numerical value of the collision risk probability, so as to adjust the collision warning level and obtain multiple different collision warning levels corresponding to different collision risk probability intervals.

[0148] Each interval corresponds to a collision warning level, and each collision warning level corresponds to an anti-collision measure.

[0149] In this embodiment, the electric vehicle anti-collision warning system based on the intelligent algorithm provided by the present invention, the collision warning level includes at least a first warning level, a second warning level and a third warning level, and the collision risk probability interval corresponding to the first warning level is greater than the collision risk probability interval corresponding to the second warning level, and the collision risk probability interval corresponding to the second warning level is greater than the collision risk probability interval corresponding to the third warning level.

[0150] The anti-collision warning module 940 is specifically used for:

[0151] When the collision warning level is at the first warning level, a first warning signal is issued and a first control instruction is received to notify the driver and control the battery vehicle to cut off the throttle of the battery vehicle in response to the first control instruction.

[0152] When the collision warning level is at the second warning level, a second warning signal is issued and a second control instruction is received to notify the driver and control the battery vehicle to decelerate in response to the second control instruction.

[0153] When the collision warning level is at the third warning level, a third warning signal is issued to inform the driver to prevent collision.

[0154] In this embodiment, the battery vehicle anti-collision warning system based on the intelligent algorithm provided by the present invention further includes an indicator light control module for:

[0155] When the collision warning level is at the first warning level, the flashing frequency of the indicator light is controlled to be the first frequency in response to the first warning signal.

[0156] When the collision warning level is at the second warning level, the flashing frequency of the indicator light is controlled to be a second frequency in response to the second warning signal.

[0157] When the collision warning level is at the third warning level, the flashing frequency of the indicator light is controlled to be a third frequency in response to the third warning signal.

[0158] Among them, the first frequency is greater than the second frequency, the second frequency is greater than the third frequency, and the greater the flashing frequency of the indicator light, the closer the electric vehicle is to the obstacle, and the higher the probability of collision risk.

[0159] Fig.10 An example of a physical structure diagram of an electronic device is shown. The electronic device may be a smart terminal, and its internal structure diagram may be as follows: Fig.10 As shown. The electronic device includes a processor, an internal memory and a network interface connected via a system bus. The processor of the electronic device is used to provide computing and control capabilities. The memory of the electronic device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the electronic device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a battery vehicle anti-collision warning method based on an intelligent algorithm is implemented, and the method includes:

[0160] The environmental data around the battery vehicle is obtained through the image acquisition device and the ultrasonic sensor, and the environmental data includes dynamic images and static images of obstacles around the vehicle body and the distance between the battery vehicle and the obstacles;

[0161] The environmental data is used as the input of the FusionNet multi-layer perceptron model to call the FusionNet multi-layer perceptron model to learn and analyze the environmental data, so as to identify obstacles and predict the movement path of obstacles;

[0162] The dynamic risk assessment algorithm based on Bayesian statistics is called to calculate the risk assessment result of the collision between the electric vehicle and the obstacle according to the moving path of the obstacle, and the collision warning level is adjusted and the anti-collision measures are obtained according to the risk assessment result;

[0163] When the collision warning level exceeds the preset level, a warning signal is issued based on the anti-collision measures and the reaction mechanism of the battery vehicle is activated to prevent the battery vehicle from colliding with obstacles;

[0164] Among them, the risk assessment results include the risk probability of a collision between an electric vehicle and an obstacle. The collision warning level is obtained according to the risk probability. The higher the risk probability, the higher the collision warning level. The anti-collision measures include warning signals and response mechanisms. The response mechanism includes controlling the electric vehicle to decelerate and cutting off the throttle of the electric vehicle.

[0165] Those skilled in the art will understand that Fig.10 The structure shown in the figure is only a block diagram of a part of the structure related to the scheme of the present invention, and does not constitute a limitation on the electronic device to which the scheme of the present invention is applied. The specific electronic device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0166] On the other hand, the present invention also provides a computer storage medium storing a computer program, which, when executed by a processor, implements a battery vehicle anti-collision warning method based on an intelligent algorithm, the method comprising:

[0167] The environmental data around the battery vehicle is obtained through the image acquisition device and the ultrasonic sensor, and the environmental data includes dynamic images and static images of obstacles around the vehicle body and the distance between the battery vehicle and the obstacles;

[0168] The environmental data is used as the input of the FusionNet multi-layer perceptron model to call the FusionNet multi-layer perceptron model to learn and analyze the environmental data, so as to identify obstacles and predict the movement path of obstacles;

[0169] The dynamic risk assessment algorithm based on Bayesian statistics is called to calculate the risk assessment result of the collision between the electric vehicle and the obstacle according to the moving path of the obstacle, and the collision warning level is adjusted and the anti-collision measures are obtained according to the risk assessment result;

[0170] When the collision warning level exceeds the preset level, a warning signal is issued based on the anti-collision measures and the reaction mechanism of the battery vehicle is activated to prevent the battery vehicle from colliding with obstacles;

[0171] Among them, the risk assessment results include the risk probability of a collision between an electric vehicle and an obstacle. The collision warning level is obtained according to the risk probability. The higher the risk probability, the higher the collision warning level. The anti-collision measures include warning signals and response mechanisms. The response mechanism includes controlling the electric vehicle to decelerate and cutting off the throttle of the electric vehicle.

[0172] In another aspect, a computer program product or a computer program is provided, the computer program product or the computer program comprising computer instructions, the computer instructions being stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium, and when the processor executes the computer instructions, a battery vehicle anti-collision warning method based on an intelligent algorithm is implemented, the method comprising:

[0173] The environmental data around the battery vehicle is obtained through the image acquisition device and the ultrasonic sensor, and the environmental data includes dynamic images and static images of obstacles around the vehicle body and the distance between the battery vehicle and the obstacles;

[0174] The environmental data is used as the input of the FusionNet multi-layer perceptron model to call the FusionNet multi-layer perceptron model to learn and analyze the environmental data, so as to identify obstacles and predict the movement path of obstacles;

[0175] The dynamic risk assessment algorithm based on Bayesian statistics is called to calculate the risk assessment result of the collision between the electric vehicle and the obstacle according to the moving path of the obstacle, and the collision warning level is adjusted and the anti-collision measures are obtained according to the risk assessment result;

[0176] When the collision warning level exceeds the preset level, a warning signal is issued based on the anti-collision measures and the reaction mechanism of the battery vehicle is activated to prevent the battery vehicle from colliding with obstacles;

[0177] Among them, the risk assessment results include the risk probability of a collision between an electric vehicle and an obstacle. The collision warning level is obtained according to the risk probability. The higher the risk probability, the higher the collision warning level. The anti-collision measures include warning signals and response mechanisms. The response mechanism includes controlling the electric vehicle to decelerate and cutting off the throttle of the electric vehicle.

[0178] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing related hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory.

[0179] By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0180] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0181] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A battery vehicle anti-collision warning method based on intelligent algorithm, characterized in that: The method comprises: Acquire environmental data around the battery vehicle through an image acquisition device and an ultrasonic sensor, wherein the environmental data includes dynamic images and static images of obstacles around the vehicle body and the distance between the battery vehicle and the obstacles; Using the environmental data as input of a FusionNet multi-layer perceptron model, so as to call the FusionNet multi-layer perceptron model to learn and analyze the environmental data, so as to identify the obstacle and predict the movement path of the obstacle; Calling a dynamic risk assessment algorithm based on Bayesian statistics to calculate a risk assessment result of a collision between the electric vehicle and the obstacle according to the moving path of the obstacle, and adjusting the collision warning level and obtaining anti-collision measures according to the risk assessment result; When the collision warning level exceeds the preset level, a warning signal is issued based on the anti-collision measure and a reaction mechanism of the battery vehicle is activated to prevent the battery vehicle from colliding with the obstacle; Among them, the risk assessment result includes the risk probability of the electric vehicle colliding with the obstacle, the collision warning level is obtained by dividing according to the risk probability, and the higher the risk probability, the higher the collision warning level, the anti-collision measures include the warning signal and the reaction mechanism, and the reaction mechanism includes controlling the electric vehicle to decelerate and cutting off the throttle of the electric vehicle.

2. The battery vehicle anti-collision warning method based on intelligent algorithm according to claim 1 is characterized in that: The method of obtaining environmental data around the battery vehicle by using an image acquisition device and an ultrasonic sensor includes: Acquire dynamic images and static images of obstacles in a first space outside the battery vehicle through the image acquisition device, and acquire the real-time distance between the battery vehicle and each obstacle in the first space through the ultrasonic sensor; Preprocessing the static image, dynamic image and real-time distance so that the environmental data meets the input requirements of the FusionNet multi-layer perceptron model; Among them, the image acquisition device is a camera, and the camera and the ultrasonic sensor are fixedly arranged on the body of the electric vehicle, the first space is the monitoring range of the image acquisition device and the ultrasonic sensor, and the preprocessing includes cleaning, deduplication, denoising and blank value filling.

3. The battery vehicle anti-collision warning method based on intelligent algorithm according to claim 2 is characterized in that: The environmental data also includes the spatial position coordinates of the obstacle in the three-dimensional space; The method of acquiring environmental data around the battery vehicle by using an image acquisition device and an ultrasonic sensor also includes: Acquire the pixel coordinates of the obstacle in the two-dimensional space based on the static image and the dynamic image, and calculate the spatial position coordinates of the obstacle in the three-dimensional space by combining the real-time distance and the calibration information of the camera through geometric transformation; The calibration information includes the focal length of the camera and the relative position between the camera and the obstacle.

4. The battery vehicle anti-collision warning method based on intelligent algorithm according to claim 3 is characterized in that: The step of using the environmental data as an input of a FusionNet multi-layer perceptron model to call the FusionNet multi-layer perceptron model to learn and analyze the environmental data to identify the obstacle and predict the movement path of the obstacle includes: Inputting the environmental data vector corresponding to the preprocessed environmental data into the FusionNet multi-layer perceptron model, and calling the activation function in the FusionNet multi-layer perceptron model to process the environmental data vector in combination with network parameters and deviations to output obstacle types and movement trajectories; The FusionNet multi-layer perceptron model is trained by a back propagation algorithm, so as to call the back propagation algorithm to update the network parameters and deviations by calculating the error between the predicted output of the model and the actual output; The activation function includes a ReLU function and a Sigmoid function, and the network parameter is a weight matrix, which is used to convert input data into useful output.

5. The battery vehicle anti-collision warning method based on intelligent algorithm according to claim 4 is characterized in that: The calling of the dynamic risk assessment algorithm based on Bayesian statistics to calculate the risk assessment result of the battery vehicle colliding with the obstacle according to the moving path of the obstacle, and adjusting the collision warning level and obtaining anti-collision measures according to the risk assessment result, includes: Calling a Bayesian statistical algorithm to calculate the collision risk probability of the current environmental data based on the collision probability statistics of the historical environmental data, and updating the collision risk probability to adjust the risk assessment result; Dividing the collision risk probability into a plurality of intervals according to the numerical value of the collision risk probability, so as to adjust the collision warning level, and obtain a plurality of different collision warning levels corresponding to different collision risk probability intervals; Each interval corresponds to one of the collision warning levels, and each of the collision warning levels corresponds to one of the collision prevention measures.

6. The battery vehicle anti-collision warning method based on intelligent algorithm according to claim 5 is characterized in that: The collision warning level includes at least a first warning level, a second warning level and a third warning level, and the collision risk probability interval corresponding to the first warning level is greater than the collision risk probability interval corresponding to the second warning level, and the collision risk probability interval corresponding to the second warning level is greater than the collision risk probability interval corresponding to the third warning level; When the collision warning level exceeds the preset level, a warning signal is issued based on the anti-collision measure and a reaction mechanism of the battery vehicle is activated to prevent the battery vehicle from colliding with the obstacle, including: When the collision warning level is at the first warning level, a first warning signal is issued and a first control instruction is received to notify the driver and control the battery vehicle to cut off the throttle of the battery vehicle in response to the first control instruction; When the collision warning level is at the second warning level, a second warning signal is issued and a second control instruction is received to notify the driver and control the battery vehicle to decelerate in response to the second control instruction; When the collision warning level is at the third warning level, a third warning signal is issued to inform the driver to avoid collision.

7. The battery vehicle anti-collision warning method based on intelligent algorithm according to claim 6 is characterized in that: The battery vehicle is provided with an indicator light, which is used to remind the driver to pay attention to collision avoidance. The method further includes: When the collision warning level is at the first warning level, in response to the first warning signal, controlling the flashing frequency of the indicator light to be a first frequency; When the collision warning level is at the second warning level, in response to the second warning signal, controlling the flashing frequency of the indicator light to be a second frequency; When the collision warning level is at the third warning level, in response to the third warning signal, controlling the flashing frequency of the indicator light to be a third frequency; Among them, the first frequency is greater than the second frequency, the second frequency is greater than the third frequency, and the greater the flashing frequency of the indicator light, the closer the electric vehicle is to the obstacle, and the higher the probability of collision risk.

8. A battery vehicle anti-collision warning system based on intelligent algorithm, characterized in that: The system comprises: A data acquisition module, used to acquire environmental data around the battery vehicle through an image acquisition device and an ultrasonic sensor, wherein the environmental data includes dynamic images and static images of obstacles around the vehicle body and the distance between the battery vehicle and the obstacles; An obstacle recognition module, used to use the environmental data as an input of a FusionNet multi-layer perceptron model, so as to call the FusionNet multi-layer perceptron model to learn and analyze the environmental data, so as to identify the obstacle and predict the movement path of the obstacle; A collision risk assessment module, which is used to call a dynamic risk assessment algorithm based on Bayesian statistics to calculate a risk assessment result of a collision between the electric vehicle and the obstacle according to the moving path of the obstacle, and to adjust the collision warning level and obtain anti-collision measures according to the risk assessment result; An anti-collision warning module is used to send out a warning signal based on the anti-collision measures and start the reaction mechanism of the battery vehicle to prevent the battery vehicle from colliding with the obstacle when the collision warning level exceeds a preset level; Among them, the risk assessment result includes the risk probability of the electric vehicle colliding with the obstacle, the collision warning level is obtained by dividing according to the risk probability, and the higher the risk probability, the higher the collision warning level, the anti-collision measures include the warning signal and the reaction mechanism, and the reaction mechanism includes controlling the electric vehicle to decelerate and cutting off the throttle of the electric vehicle.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images