CN115716476A - Lane changing method, device, electronic device and storage medium for automatic driving - Google Patents

Abstract

The invention relates to the technical field of automatic driving, and discloses an automatic driving lane changing method, an automatic driving lane changing device, electronic equipment and a storage medium. The method comprises the following steps: acquiring real-time road condition information and real-time vehicle condition information; acquiring electroencephalogram information of a driver to obtain real-time electroencephalogram information, wherein the electroencephalogram information represents driving intention of the driver; predicting a driving route of the self-vehicle running on the current road condition by using a trained lane change decision model based on the real-time road condition information and the real-time vehicle condition information to obtain a route prediction result; and comparing the route prediction result with the driving intention of the driver represented by the electroencephalogram information, and controlling the self-vehicle to execute lane-changing driving action along the route prediction result when the route prediction result is consistent with the driving intention of the driver. The embodiment of the invention can improve the synchronism and safety during automatic driving lane change.

Description

Lane changing method, device, electronic device and storage medium for automatic driving

technical field

The present invention relates to the technical field of automatic driving, in particular to an automatic driving lane changing method, device, electronic equipment and storage medium.

Background technique

At present, road traffic injuries have become one of the important causes of death among people. About 95% of all kinds of motor vehicle accidents are caused to a certain extent by improper operation of drivers, while traffic accidents caused entirely by inappropriate driving behaviors of drivers Accidents accounted for three quarters of the total number of accidents. Therefore, it is very necessary to analyze the driver's behavior and driver's state.

At present, autonomous driving is mostly used in specific scenarios, and it is still unable to match the decision-making and control capabilities of human drivers under complex working conditions, making it difficult to achieve full-working-condition applications.

Contents of the invention

The object of the present invention is to provide a method, device, electronic equipment and storage medium for automatic driving lane change, aiming at improving the synchronization and safety of automatic driving lane change.

In the first aspect, a method for changing lanes in automatic driving is provided, including:

Obtain real-time road condition information and real-time vehicle condition information. The road condition information includes the number of lanes, the lane where the vehicle is located, the driving vehicle, and the distance between obstacles representing the relative distance between the vehicle and the vehicle or obstacles. The vehicle condition information includes the speed of the vehicle and ego positioning;

Collect the EEG of the driver to obtain real-time EEG information, which represents the driver's driving intention;

Based on real-time road condition information and real-time vehicle condition information, use the trained lane change decision-making model to predict the driving route of the own vehicle in the current road condition, and obtain the route prediction result;

The route prediction result is compared with the driver's driving intention represented by EEG information, and when the route prediction result is consistent with the driver's driving intention, the ego vehicle is controlled to change lanes along the route prediction result.

In some embodiments, before the acquisition of road condition information and vehicle condition information, it also includes:

Obtain offline training information, where the offline training information includes offline road condition information and offline vehicle condition information;

Construct a decision table using offline training information;

Discretize the offline training information in the decision table, use the preset neural network model to classify the discretized offline training information, and obtain discrete classification results;

Perform attribute reduction on the decision table according to the discrete classification results, and determine the minimum conditional attribute set for expressing the decision;

Integrate the minimum conditional attribute set that expresses the same decision result, and calculate the coverage and confidence level of the integrated minimum conditional attribute set to determine the offline decision rule;

The lane-changing decision-making model is trained using offline decision-making rules, and the route prediction results of the lane-changing decision-making model are compared with offline EEG information.

In some embodiments, performing attribute reduction on the decision table according to the discrete classification, and determining the minimum conditional attribute set for expressing the decision result include:

Construct a training data set according to the offline training information, and randomly encode each type of offline training information in the training data set;

Calculate the dependence of decision attributes on offline training information required for decision-making;

Determine whether the current degree of dependence is equal to the minimum relative dependence value;

If yes, output the minimum conditional attribute set when the training data set with optimal fitness for several consecutive generations is no longer improved;

If not, randomly select, crossover and mutate the training data set iteratively, copy the training data set with the best fitness to the next generation training data set, until the training data set with the best fitness for several consecutive generations is no longer Raise and output a minimal set of conditional attributes.

In some embodiments, the EEG collection of the driver to obtain the EEG information representing the driver's driving intention includes:

Collect the brain wave signal generated by the driver while driving to obtain the initial brain wave information;

Preprocessing the initial EEG information, using a preset EEG analog template to compare the preprocessed initial EEG information, to obtain an EEG analog result;

According to the EEG analogy results, the error potentials in the preprocessed initial EEG information are determined and eliminated, and the EEG information representing the driver's driving intention is obtained.

In some embodiments, the controlling the ego vehicle to perform lane-changing actions along the predicted route includes:

Find several preview points according to the distance between the vehicle and the obstacle, and use the Bezier curve constructed by the preview points to draw the lane change route;

Boundary expansion processing is performed on obstacles, and the expansion radius of obstacles is set according to the current lane change route and the environment information of the vehicle;

Determine whether the expansion radius of the obstacle is greater than the width of the obstacle;

If so, control the vehicle to travel along the current lane change route;

If not, reset the position of the preview point, use the Bezier curve constructed by the reset preview point to redraw the road change route, and the steering angle of the vehicle corresponding to the redrawn road change route is greater than the last drawn The steering angle of the ego vehicle corresponding to the road line returns to the step of performing boundary expansion processing on obstacles.

In some embodiments, the controlling the ego vehicle to travel along the current lane change route includes:

When there is no continuous obstacle within the detection range, control the ego vehicle after changing lanes to drive past the obstacle and then change to the lane of the initial driving route; Change to the lane of the original driving route after obstacles;

The continuous obstacle is an obstacle whose distance from the previous obstacle in the same lane is smaller than a preset obstacle distance threshold.

In some embodiments, a simulated driving scenario is constructed based on real-time road condition information and real-time vehicle condition information, and the simulated driving scenario is displayed. The simulated driving scenario converts real-time road condition information into simulated road condition information and converts the real-time vehicle condition information Convert to simulated vehicle condition information.

In a second aspect, an automatic driving lane changing device is provided, the device comprising:

The acquisition module is used to obtain real-time road condition information and real-time vehicle condition information. The road condition information includes the number of lanes, the lane where the vehicle is located, the driving vehicle, and the obstacle distance representing the relative distance between the vehicle and the vehicle or obstacle. The vehicle condition Information includes ego speed and ego location;

The acquisition module is used to collect the EEG of the driver to obtain real-time EEG information, and the EEG information represents the driving intention of the driver;

The prediction module is used to predict the driving route of the own vehicle in the current road condition by using the trained lane change decision model based on the real-time road condition information and the real-time vehicle condition information, and obtain the route prediction result;

The control module is used to compare the route prediction result with the driver's driving intention represented by the EEG information, and when the route prediction result is consistent with the driver's driving intention, control the own vehicle to perform a lane-changing action along the route prediction result.

In a third aspect, an electronic device is provided, the electronic device includes a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the method for changing lanes for automatic driving according to the first aspect is realized .

In a fourth aspect, a computer-readable storage medium is provided, the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the lane-changing method for automatic driving described in the first aspect is implemented.

Beneficial effects of the present invention: based on the method of deep learning, the lane change decision model is trained based on offline road condition information, vehicle condition information and EEG information representing the driver's driving intention, which promotes the realization of human-like automatic driving. The driver's intention recognition is objective and accurate, improving the synchronization and safety when changing lanes in automatic driving.

Description of drawings

Fig. 1 is a schematic flowchart of a lane changing method for automatic driving shown in the first embodiment.

Fig. 2 is a schematic flowchart of a lane changing method for automatic driving shown in the second embodiment.

FIG. 3 is a schematic flow chart of the first type of step S204 in FIG. 2 .

FIG. 4 is a second schematic flowchart of step S102 in FIG. 1 .

FIG. 5 is a third schematic flow chart of step S104 in FIG. 1 .

Fig. 6 is a schematic structural diagram of an automatic driving lane changing device provided by an embodiment of the present application.

FIG. 7 is a schematic diagram of a hardware structure of an electronic device provided by an embodiment of the present application.

Fig. 8 is a schematic diagram of drawing a lane changing route shown in the first embodiment.

Fig. 9 is a schematic diagram of drawing a lane changing route shown in the second embodiment.

Fig. 10 is a schematic diagram of drawing a lane changing route shown in the third embodiment.

Detailed ways

In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the present invention will be further described below in conjunction with the embodiments and the accompanying drawings.

Reference will now be made in detail to the exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, the same numerals in different drawings refer to the same or similar elements unless otherwise indicated. The implementations described in the following exemplary examples do not represent all implementations consistent with the present disclosure. Rather, they are merely examples of apparatuses and methods consistent with aspects of the present disclosure as recited in the appended claims.

In order to make the purpose, technical solution and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present application, not to limit the present application.

It should be noted that although the functional modules are divided in the schematic diagram of the device, and the logical sequence is shown in the flowchart, in some cases, it can be executed in a different order than the module division in the device or the flowchart in the flowchart. steps shown or described. The terms "first", "second" and the like in the specification and claims and the above drawings are used to distinguish similar objects, and not necessarily used to describe a specific sequence or sequence.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application, and are not intended to limit the present application.

First, analyze some nouns involved in this application:

1) Artificial Intelligence (AI)

Artificial intelligence is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technique of computer science that attempts to understand the nature of intelligence and produce a new kind of intelligent machine that can respond in a similar way to human intelligence. Artificial intelligence is to study the design principles and implementation methods of various intelligent machines, so that the machines have the functions of perception, reasoning and decision-making.

Artificial intelligence technology is a comprehensive subject that involves a wide range of fields, including both hardware-level technology and software-level technology. Artificial intelligence basic technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation/interaction systems, and mechatronics. Artificial intelligence software technology mainly includes several major directions such as computer vision technology, speech processing technology, natural language processing technology, and machine learning/deep learning.

2) Machine Learning (Machine Learning, ML)

Machine learning is a multi-field interdisciplinary subject, involving probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory and other disciplines. Specializes in the study of how computers simulate or implement human learning behaviors to acquire new knowledge or skills, and reorganize existing knowledge structures to continuously improve their performance. Machine learning is the core of artificial intelligence and the fundamental way to make computers intelligent, and its application pervades all fields of artificial intelligence. Machine learning and deep learning usually include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and teaching and learning.

In related technologies, the behavior decision-making algorithms mainly include rule-based behavior decision-making algorithms and learning-based behavior decision-making algorithms; the rule-based decision-making method has advantages in the breadth of scene traversal, strong logical interpretability, and is easy to design in modules according to the scene. There are many examples of decision-making systems using finite state machines at home and abroad. However, its system structure determines that it has a certain bottleneck in the depth of scene traversal and the accuracy of decision-making, and it is difficult to handle complex working conditions; the advantage of the behavior decision system based on learning algorithms is that it has the advantage of scene traversal depth. , it is easier to cover all working conditions through the big data system, but the learning algorithm does not have the advantage of scene traversal breadth, and the learning models required for different scenarios may be completely different; machine learning requires a large amount of experimental data as learning samples; decision-making effects depend on data quality, Insufficient samples, poor data quality, unreasonable network structure, etc. will lead to problems such as over-learning and under-learning. Under complex working conditions, it is still unable to match the decision-making and control capabilities of human drivers, and it is difficult to realize the application of all working conditions.

Based on this, embodiments of the present application provide a method, device, electronic device, and storage medium for lane changing in automatic driving, aiming at improving synchronization and safety when changing lanes in automatic driving.

The automatic driving lane changing method, device, electronic device, and storage medium provided in the embodiments of the present application are specifically described through the following embodiments. First, the recommended method in the embodiments of the present application is described.

The embodiments of the present application may acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. .

Artificial intelligence basic technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation/interaction systems, and mechatronics. Artificial intelligence software technology mainly includes computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning/deep learning.

The lane changing method for automatic driving provided in the embodiment of the present application relates to the technical field of artificial intelligence. The lane changing method for automatic driving provided in the embodiment of the present application can be applied to a terminal, can also be applied to a server, and can also be software running on the terminal or the server. In some embodiments, the terminal can be a smart phone, a tablet computer, a notebook computer, a desktop computer, etc.; the server end can be configured as an independent physical server, or can be configured as a server cluster or a distributed system composed of multiple physical servers, or It can be configured as a cloud that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server; the software may be an application to realize the method for extracting table information, but is not limited to the above forms.

The application can be used in numerous general purpose or special purpose computer system environments or configurations. Examples: personal computers, server computers, handheld or portable devices, tablet-type devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, including A distributed computing environment for any of the above systems or devices, etc. This application may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The application may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including storage devices.

It should be noted that, in each specific implementation of the present application, when it comes to related processing based on user information, user behavior data, user history data, user location information and other data related to user identity or characteristics, all will first Obtain the user's permission or consent, and the collection, use and processing of these data will comply with the relevant laws, regulations and standards of the relevant countries and regions. In addition, when the embodiment of this application needs to obtain the user's sensitive personal information, the user's separate permission or separate consent will be obtained through a pop-up window or jump to a confirmation page, etc. After the user's separate permission or separate consent is clearly obtained, the Obtain necessary user-related data for the normal operation of this embodiment of the application.

Fig. 1 is a schematic flowchart of a method for changing lanes for automatic driving shown in the first embodiment. The method in Fig. 1 may include but not limited to steps S101 to S104.

Step S101, acquiring real-time road condition information and real-time vehicle condition information.

The road condition information includes the number of lanes, the lane where the own vehicle is located, the driving vehicle, and the distance between obstacles representing the relative distance between the own vehicle and the driving vehicle or obstacles. The vehicle condition information includes the speed of the own vehicle and the location of the own vehicle.

In a possible implementation manner, the road condition information and vehicle condition information are collected by a device with a perception module including a laser radar, a camera, a navigation system, and a vehicle-machine system.

In a possible implementation, road condition information may refer to all data related to vehicle driving safety in the environment where the vehicle is located that is collected by sensing modules such as lidar or cameras. The driving state of the vehicle may also be the driving environment in the environment where the vehicle is located (such as lane line data, obstacle data, real-time positioning data, and map data), which is not particularly limited in this embodiment of the present application.

In a possible implementation, the vehicle condition information may refer to all the data collected by the vehicle-machine system and other sensing modules related to the vehicle's own driving state during driving, such as vehicle state information, actual vehicle speed data, steering wheel, brake pedal, etc. , accelerator pedal, and sensor data on the clutch, which are not specifically limited in this embodiment of the present application.

Step S102 , collecting brain electricity from the driver to obtain real-time brain electricity information, which represents the driver's driving intention.

It is understandable that when the brain nerve cells react, they will generate corresponding electric waves. These electric waves are called brain waves. Collecting the brain waves and expressing them in the form of signals is the EEG information. By analyzing the EEG information The waveform can determine the driver's driving intentions.

In a possible implementation, the EEG information can be the EEG information of the driver collected by a device with a brain-computer interface, the brain-computer interface has at least two electrodes, and the brain-computer interface During the signal collection process of the driver, the two electrodes are located in different areas of the driver's head, so as to collect the EEG information generated in different areas of the driver.

More specifically, a cosine similarity matrix is constructed according to the calculated cosine similarity, and the type of the cosine similarity matrix is predicted through the trained convolutional neural network to obtain the type of driving intention to which the brain wave signal belongs. For example, if the eigenvectors of the EEG information are all 128, then 128 cosine similarities can be calculated, then at this time, the cosine similarity matrix can be constructed based on the 128 cosine similarities, and then through the convolution after training The neural network predicts the type of the cosine similarity matrix to obtain the driving intention type of the EEG information.

Step S103, based on the real-time road condition information and real-time vehicle condition information, use the trained lane change decision model to predict the driving route of the own vehicle under the current road condition, and obtain the route prediction result.

The road condition information and vehicle condition information obtained in real time are input into the trained lane change decision model, and the lane change decision model is used to predict the driving route. The route prediction results may include lane keeping mode, lane changing driving mode or intersection turning mode. It can be understood that the lane keeping mode, lane changing driving mode and intersection steering mode are all preset driving modes, and the route prediction result is based on the road condition information, vehicle condition information and EEG information to judge the current road condition of the vehicle and the driver's For driving intentions, when the driving space in the real-time driving area meets the preset conditions of the corresponding driving mode, the driver's driving intentions are executed on the basis of ensuring safety, so as to obtain the route prediction result.

In this embodiment, the lane keeping mode can be understood as the driving mode in which the own vehicle maintains driving along the current lane, the lane-changing driving mode can be understood as the driving mode in which the own vehicle switches from the current driving lane to other lanes, and the intersection turning can be understood as When the ego vehicle arrives at the intersection, it turns according to the initial driving route or the real-time driving route.

Step S104, comparing the route prediction result with the driver's driving intention represented by the EEG information, and controlling the vehicle to change lanes along the route prediction result when the route prediction result is consistent with the driver's driving intention.

In a possible implementation, the route prediction result is compared with the driver's driving intention represented by the EEG information. When the route prediction result is consistent with the driver's driving intention, the route will The prediction result is sent to the control system of the vehicle, and the vehicle control system converts the driving mode indicated by the received route prediction result into a control command for the vehicle, and controls the vehicle to drive according to the driving mode indicated by the route prediction result. When the prediction result is a lane change, the ego vehicle is controlled to perform a lane change driving action; otherwise, the driving action corresponding to the driver's driving intention is performed.

Fig. 2 is a schematic flowchart of a method for changing lanes for automatic driving shown in the second embodiment. On the basis of the embodiment in Fig. 1, the method in Fig. 2 may include but not limited to steps S201 to S206.

Step S201, acquiring offline training information.

The offline training information includes offline road condition information and offline vehicle condition information.

It can be understood that the offline training information is historical information saved, that is, historical data road condition information and historical vehicle condition information, and the historical data road condition information and historical vehicle condition information are matched according to the generation time to form a training set.

Step S202, using offline training information to construct a decision table.

Among them, the decision table, also known as the judgment table, is a graphical tool in the form of a table, which is suitable for describing the situation where there are many judgment conditions, the conditions are combined with each other, and there are multiple decision-making options. A way of describing complex logic with precision and succinctness, mapping multiple conditions to actions to be performed if those conditions are met. However, unlike the control statements in traditional programming languages, the decision table can clearly express the direct connection between multiple independent conditions and multiple actions.

In a possible implementation, the offline training information is preprocessed so that the offline road condition information and the offline vehicle condition information are matched according to time. For incomplete matching results, the rough neural network algorithm is applied to eliminate the incomplete offline training information to form a decision table.

Step S203, discretize the offline training information in the decision table, classify the discretized offline training information by using a preset neural network model, and obtain a discrete classification result.

In a possible implementation, the discretized offline training information is input into a preset neural network model, and the preset neural network model is preset with a probability parameter of a classification distribution, which can be a classification distribution that obeys the preset probability parameter , the specific probability value of the preset probability parameter can be initially set in advance, and each group of data in the offline training information is numbered in turn. After obtaining the preset probability parameter, obtain the data with the same label as the distribution selection parameter, and select Classify the data into the corresponding categories to obtain discrete classification results.

Step S204, perform attribute reduction on the decision table according to the discrete classification result, and determine the minimum conditional attribute set for expressing the decision.

It can be understood that attribute reduction is to delete the offline training information that has little decision-making effect in the decision table and reduce the data volume of the decision table. This is because the discretized offline training information has large data samples and diverse data. Due to the high quality of the data samples, the deleted offline training information will not affect the expression of the object, and these redundant data are deleted to simplify the conditional attributes of the decision.

Step S205 , integrating the minimum conditional attribute sets that express the same decision result, and calculating the coverage and confidence level of the integrated minimum conditional attribute sets to determine offline decision rules.

In a possible implementation, the minimum set of conditional attributes that express the same decision result is summarized, so as to extract the decision rules. Specifically, the rule extraction is a process of summarizing the previous work. The decision table is divided into The attribute reduction situation is re-simplified, and the discretized and reduced conditional attribute intervals and the rules with the same decision attribute are combined together, and the coverage of the rule is calculated. The coverage is the rule that accounts for the same decision attribute. percentage of all rules.

Step S206, using the offline decision rule to train the lane change decision model, and comparing the route prediction result of the lane change decision model with the offline EEG information.

In a possible implementation, the parameters of the lane change decision model are iteratively adjusted during the training process, and the training ends when the matching degree between the route prediction result of the lane change decision model and the offline EEG information reaches a preset accuracy rate.

Please refer to FIG. 3 , in some embodiments, step S204 may include but not limited to include steps S301 to S305.

Step S301, constructing a training data set according to offline training information, and randomly encoding each type of offline training information in the training data set.

In a possible implementation, encoding each type of offline training information in the training data set may be represented by a fixed-length binary string and a random binary number, where each bit Indicates an attribute, for example, 1 means that the decision attribute is selected as the required data, and 0 means that the decision attribute is not selected as the required data.

Exemplarily, for typical lane-changing scenarios, the training data set constructed from offline training information is {xi _, y _i , v _i , xi _environment , y _{i environment} , v _{i environment} }, where v _i is the vehicle speed , (x _i , y _i ) is the horizontal and vertical coordinates of the GPS position of the self-vehicle, (xi _environment , y _{i environment} ) is the position coordinates of the surrounding vehicles or obstacles relative to the self-vehicle, and the v _i environment is the position coordinates of the surrounding vehicles relative to the self-vehicle The speed of the training data set is randomly encoded to obtain a binary string {0,0,1,1,0,0}.

Step S302, calculating the degree of dependence of the decision attribute on the offline training information required for the decision.

In a possible implementation, first define a fitness value function, and the fitness value function is:

Use the dependence degree formula to calculate the dependence degree of the decision attribute on the offline training information required for decision-making. The dependence degree formula is:

Among them, f(x) is the fitness function, h(x) is the number of 1s in a binary string, which is the number of selected offline training information data, and n is the total number of characters in this binary string , that is, the number of all offline training information data, m is the dependence degree of the decision attribute on the condition attribute represented by 1 in this binary string, the larger the value of m, the stronger the dependence of the decision attribute on the condition attribute, the goal is to ensure Keep h(x) as small as possible while the dependence of the decision-making attribute on the whole remains unchanged, V _C (D) represents the dependence degree of the decision-making attribute D, c is the data of offline training information, and U is the domain of discourse.

Step S303, judging whether the current degree of dependence is equal to the minimum relative dependence value. If yes, execute step S304; if no, execute step S305.

Step S304, when the training data set with optimal fitness for several consecutive generations is no longer improved, output the minimum conditional attribute set.

Step S305, iteratively perform random selection, crossover and mutation processing on the training data set, copy the training data set with the best fitness to the next generation training data set, until the training data set with the best fitness for several consecutive generations is no longer Raise and output a minimal set of conditional attributes.

Please refer to FIG. 4 , in some embodiments, step S102 may include but not limited to include steps S401 to S403.

Step S401, collecting brain wave signals generated by the driver while driving to obtain initial brain wave information.

Step S402, preprocessing the initial EEG information, using a preset EEG analog template to compare the preprocessed initial EEG information to obtain an EEG analog result.

Step S403, according to the EEG analogy results, determine the error potentials in the preprocessed initial EEG information and perform elimination processing to obtain the EEG information representing the driving intention of the driver.

In a possible implementation, based on the Biopac actiCHAmp 64-channel EEG acquisition and analysis system, the EEG signals of 32 EEG channels FP1-O2 are collected; according to the position of the vehicle, it is divided into left front, front front, The front right, rear left, front rear, and rear right six areas collect the position and speed information of the own vehicle and the position and speed information of other vehicles in the six areas. EEG signals using FP1, F3, F7, FC5, FC1, FP2, F7, F4, F8, FC6, FC2 channels are spatially filtered using Common Average Reference (CAR) to eliminate noise; use 1-10hz Band-pass filtering is used for band-pass filtering; canonical correlation analysis (CCA) is used to enhance the characteristics of EEG signals, and finally, template matching is performed based on analog templates to achieve classification, and if false correlation potentials are generated, they are eliminated.

Please refer to FIG. 5 , in some embodiments, step S104 may include but not limited to include steps S501 to S505.

Step S501, find several preview points according to the distance between the ego vehicle and the obstacle, and use the Bezier curve constructed by the preview points to draw a lane change route.

A Bézier curve, also known as a Bézier curve or a Bézier curve, is a mathematical curve applied to two-dimensional graphics applications. The Bezier curve creates and edits graphics by controlling four points on the curve (start point, end point, and two mutually separated intermediate points), among which the control line in the center of the curve plays an important role. The line is virtual, intersected with a Bezier curve in the middle, and controlled endpoints at both ends. When the endpoints at both ends are moved, the Bezier curve changes the curvature (the degree of curvature); when the middle point is moved (that is, the virtual control line is moved), the Bezier curve moves evenly with the start point and end point locked .

Bezier curves have the advantages of continuous curvature, easy tracking and satisfying vehicle dynamic constraints. Therefore, the embodiment of the present invention chooses to use the fifth-order Bezier curve to draw the road change route.

The fifth-order Bezier curve can be expressed as:

可以表示为：

It can be expressed as:

Among them, B(t) represents the Bezier curve, P ₀ , P ₁ , P ₂ , P ₃ , P ₄ and P ₅ are preview points, t∈[0,1], n∈[0,5] , i∈[0,5].

Finding several preview points according to the distance between the vehicle and the obstacle can be selected in the real-time driving area, and the selected preview points are connected to obtain a Bezier curve as a lane change route. As shown in Figure 8, the coordinates P _i = (X _i , Y _i ) of the six preview points in the fifth-order Bezier curve are determined according to the distance between the selected self-driving vehicle and the obstacle, specifically:

First determine the coordinates of the preview points P ₀ and P ₅ :

P ₀ =(X ₀ ,Y ₀ );

P ₅ =(X ₅ ,Y ₅ )=(L,w _r );

Divide the distance between the ego vehicle and the obstacle into 4 parts, and obtain the coordinates of P ₁ , P ₂ , P ₃ and P ₄ :

P ₁ =(X ₁ ,Y ₁ )=(X ₅ /4,Y ₀ );

P ₂ =(X ₂ ,Y ₂ )=(X ₅ /2,Y ₀ );

P ₃ =(X ₃ ,Y ₃ )=(X ₅ /2,Y ₅ );

P ₄ =(X ₄ ,Y ₄ )=(3X ₅ /4,Y ₅ );

Among them, L is the distance between the vehicle and the obstacle in the frenet coordinate system detected by the perception system; w _r is the lane width information provided by the lane line detection in the perception system.

Step S502, performing boundary expansion processing on the obstacle, and setting the expansion radius of the obstacle according to the current lane change route and the environment information of the self-vehicle.

When using the Bezier curve to draw the road-changing route, since the vehicle is regarded as a mass point and the width of the vehicle itself is ignored, there is a possibility that the self-vehicle may collide with obstacles when driving along the road-changing route, and the boundary expansion of obstacles is processed , compare the distance between the ego vehicle and the obstacle when driving along the road-changing route and the collision radius of the obstacle processed by boundary expansion, and perform collision evaluation. The formula for collision assessment can be expressed as:

为车辆自身宽度，d_min为期望的障碍物与自车的最小安全距离，y_ob为障碍物中心在frenet坐标系下与车道线的投影距离，w_c1为感知系统检测的障碍物宽度。Among them, r is the expansion radius of the obstacle,

is the width of the vehicle itself, d _min is the minimum safe distance between the expected obstacle and the own vehicle, y _ob is the projected distance between the center of the obstacle and the lane line in the frenet coordinate system, and w _c1 is the width of the obstacle detected by the perception system.

Step S503, judging whether the expansion radius of the obstacle is greater than the width of the obstacle. If yes, execute step S504; if not, execute step S505.

When judging whether the expansion radius of the obstacle is greater than the width of the obstacle, that is:

Step S404, controlling the ego vehicle to travel along the current lane changing route.

Step S505, reset the position of the preview point, use the Bezier curve constructed by the reset preview point to redraw the road change route, and the steering angle of the vehicle corresponding to the redrawn road change route is greater than the previous drawing The steering angle of the ego vehicle corresponding to the road line. Return to step S502.

Reselect the preview point in the real-time driving area, first re-determine the coordinates of the preview point _P5 , and increase the steering angle of the own vehicle by shortening the distance between _P0 and _P5 , so that the redrawn lane change line corresponds to The steering angle of the own vehicle is greater than the steering angle of the own vehicle corresponding to the lane change route drawn last time, which reduces the risk of collision between the own vehicle and obstacles. In this embodiment, the coordinates of the preview point _P5 are changed with the following formula:

P ₅ =(X ₅ ,Y ₅ )=(4L/5,w _r ).

As shown in Figure 9, after the self-vehicle travels along the current lane-changing route and bypasses obstacles, control the ego-vehicle to return to the lane of the initial driving route, which can be drawn using a mirrored Bezier curve to return to the lane of the initial driving route to return to the lane of the initial travel route after going around the obstacle.

In some embodiments, as shown in FIG. 10 , controlling the ego vehicle to travel along the current lane-changing route may be that when no continuous obstacles appear within the detection range, the ego vehicle after the lane-changing is controlled to drive past the obstacle and then change to the initial state. The lane of the driving route; when continuous obstacles appear within the detection range, the ego vehicle after the lane change is controlled to pass through the continuous obstacles and then change to the lane of the initial driving route. Wherein, the continuous obstacle is an obstacle whose distance from the previous obstacle in the same lane is smaller than a preset obstacle distance threshold.

In the above embodiments, the control of the ego vehicle includes lateral control and longitudinal control.

Lateral control adopts LQR algorithm, and the steps of LQR algorithm in discrete cases are as follows:

1. Determine the iteration range N;

2. Set the initial value of iteration, let P _N =Q _f , where Q _f =Q;

3. Loop iteration, from back to front t=N,...,1:

P _t-1 ＝Q+A ^T P _t AA ^T P _t B(R+B ^T P _t B) ^-1 B ^T P _t A;

4. Calculate the feedback coefficient cyclically from t=0,...,N-1:

K _t = (R+B ^T P _t+1 B) ^-1 B ^T P _t+1 A;

5. Finally, the optimized control amount is obtained:

The longitudinal control is controlled based on the PID algorithm, specifically:

Construct the control quantity output model based on the target speed, the control quantity output model is:

Among them, U _t represents the control output, e(t)=v _{d -va} , v _d represents the target speed, v _a represents the current speed, K _P represents the proportional amplification factor, K _I represents the integral time constant, K _D represents the differential time constant;

Discretization is performed on the control quantity output model to obtain a discrete control quantity output model, and the discrete control quantity output model is:

Among them, T is the command cycle of the controller, e _j represents the speed deviation at time j, e _k represents the speed deviation at time k, and e _k-1 represents the speed deviation at time k-1;

The discrete control output model is used to calculate the control output, and the assisted driving control of the vehicle is performed through the calculated control output.

In some embodiments, a simulated driving scenario is constructed based on real-time road condition information and real-time vehicle condition information, and the simulated driving scenario is displayed. The simulated driving scenario converts real-time road condition information into simulated road condition information and converts the real-time vehicle condition information Convert to simulated vehicle condition information.

As can be seen from the above, in the automatic driving lane change method provided in the embodiment of the present application, the lane change decision model is trained based on the deep learning method based on offline road condition information, vehicle condition information and EEG information representing the driver's driving intention. , to promote the realization of human-like automatic driving, the driver's intention recognition based on EEG is objective and accurate, and improve the synchronization and safety of automatic driving when changing lanes.

In order to better implement the above method, an embodiment of the present invention further provides a driving assistance device, and the driving assistance device may specifically be integrated into an electronic device such as a server or a terminal.

Please refer to Fig. 6, the embodiment of the present application also provides an automatic driving lane changing device, which can implement the automatic driving lane changing method mentioned in the above embodiment, the device includes:

The acquisition module 601 is used to acquire real-time road condition information and real-time vehicle condition information. The road condition information includes the number of lanes, the lane where the vehicle is located, the driving vehicle, and the obstacle distance representing the relative distance between the vehicle and the vehicle or obstacles. Vehicle condition information includes vehicle speed and vehicle location;

The acquisition module 602 is used to collect the EEG of the driver to obtain real-time EEG information, which represents the driver's driving intention;

Prediction module 603, for based on real-time road condition information and real-time vehicle condition information, utilizes the well-trained lane-changing decision-making model to predict the traveling route of self-vehicle in current road condition, obtains route prediction result;

The control module 604 is used to compare the route prediction result with the driver's driving intention represented by the EEG information, and when the route prediction result is consistent with the driver's driving intention, control the self-vehicle to perform a lane-changing action along the route prediction result.

The specific implementation of the automatic driving lane changing device is basically the same as the specific embodiment of the above automatic driving lane changing method, and will not be repeated here.

An embodiment of the present application also provides an electronic device, the electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the above-mentioned lane changing method for automatic driving when executing the computer program. The electronic device may be any intelligent terminal including a tablet computer, a vehicle-mounted computer, and the like.

Please refer to FIG. 7. FIG. 7 illustrates a hardware structure of an electronic device in another embodiment. The electronic device includes:

The processor 701 may be implemented by a general-purpose CPU (Central Processing Unit, central processing unit), a microprocessor, an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits, and is used to execute related programs to realize The technical solutions provided by the embodiments of the present application;

The memory 702 may be implemented in the form of a read-only memory (ReadOnlyMemory, ROM), a static storage device, a dynamic storage device, or a random access memory (RandomAccessMemory, RAM). The memory 702 can store operating systems and other application programs. When implementing the technical solutions provided by the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 702 and called by the processor 701 to execute the implementation of this application. Example of automatic driving lane change method;

The input/output interface 703 is used to realize information input and output;

The communication interface 704 is used to realize the communication interaction between the device and other devices, and the communication can be realized through a wired method (such as USB, network cable, etc.), or can be realized through a wireless method (such as a mobile network, WIFI, Bluetooth, etc.);

A bus 705, which transmits information between various components of the device (such as a processor 701, a memory 702, an input/output interface 703, and a communication interface 704);

The processor 701 , the memory 702 , the input/output interface 703 and the communication interface 704 are connected to each other within the device through the bus 705 .

The embodiment of the present application also provides a computer-readable storage medium, where a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, the above-mentioned lane changing method for automatic driving is realized.

As a non-transitory computer-readable storage medium, memory can be used to store non-transitory software programs and non-transitory computer-executable programs. In addition, the memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid-state storage devices. In some embodiments, the memory optionally includes memory located remotely from the processor, and these remote memories may be connected to the processor via a network. Examples of the aforementioned networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

The automatic driving lane change method, device, electronic equipment, and storage medium provided in the embodiments of the present application are based on deep learning methods, and based on offline road condition information, vehicle condition information, and EEG information representing the driver's driving intention, the lane change decision model is implemented. Training to promote the realization of human-like automatic driving. The driver's intention recognition based on EEG is objective and accurate, and the synchronization and safety of automatic driving when changing lanes are improved.

The embodiments described in the embodiments of the present application are to illustrate the technical solutions of the embodiments of the present application more clearly, and do not constitute a limitation to the technical solutions provided by the embodiments of the present application. Those skilled in the art know that with the evolution of technology and new For the emergence of application scenarios, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems.

Those skilled in the art can understand that the technical solution shown in the figure does not constitute a limitation to the embodiment of the present application, and may include more or less steps than those shown in the figure, or combine some steps, or different steps.

The device embodiments described above are only illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place, or may be distributed to multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, the functional modules/units in the system, and the device can be implemented as software, firmware, hardware, and an appropriate combination thereof.

The terms "first", "second", "third", "fourth", etc. (if any) in the description of the present application and the above drawings are used to distinguish similar objects and not necessarily to describe specific sequence or sequence. It is to be understood that the data so used are interchangeable under appropriate circumstances such that the embodiments of the application described herein can be practiced in sequences other than those illustrated or described herein. Furthermore, the terms "comprising" and "having", as well as any variations thereof, are intended to cover a non-exclusive inclusion, for example, a process, method, system, product or device comprising a sequence of steps or elements is not necessarily limited to the expressly listed instead, may include other steps or elements not explicitly listed or inherent to the process, method, product or apparatus.

It should be understood that in this application, "at least one (item)" means one or more, and "multiple" means two or more. "And/or" is used to describe the association relationship of associated objects, indicating that there can be three types of relationships, for example, "A and/or B" can mean: only A exists, only B exists, and A and B exist at the same time , where A and B can be singular or plural. The character "/" generally indicates that the contextual objects are an "or" relationship. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one item (piece) of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c ", where a, b, c can be single or multiple.

In the several embodiments provided in this application, it should be understood that the disclosed devices and methods may be implemented in other ways. For example, the device embodiments described above are only illustrative. For example, the division of the above units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or can be Integrate into another system, or some features may be ignored, or not implemented. In another point, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces, and the indirect coupling or communication connection of devices or units may be in electrical, mechanical or other forms.

The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, each unit may exist separately physically, or two or more units may be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

If the integrated unit is realized in the form of a software function unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or part of the contribution to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium , including multiple instructions to make a computer device (which may be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the method in each embodiment of the present application. The aforementioned storage media include: U disk, mobile hard disk, read-only memory (Read-Only Memory, ROM for short), random access memory (Random Access Memory, RAM for short), magnetic disk or optical disk, etc., which can store programs. medium.

The preferred embodiments of the embodiments of the present application have been described above with reference to the accompanying drawings, which does not limit the scope of rights of the embodiments of the present application. Any modifications, equivalent replacements and improvements made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall fall within the scope of rights of the embodiments of the present application.

Claims (10)

1. An automatic driving lane change method, comprising:

acquiring real-time road condition information and real-time vehicle condition information, wherein the road condition information comprises the number of lanes, lanes where self vehicles are located, running vehicles and barrier distances representing relative distances between the self vehicles and the running vehicles or barriers, and the vehicle condition information comprises self vehicle speed and self vehicle positioning;

performing electroencephalogram acquisition on a driver to obtain real-time electroencephalogram information, wherein the electroencephalogram information represents the driving intention of the driver;

predicting a driving route of the self-vehicle running on the current road condition by using a trained lane change decision model based on the real-time road condition information and the real-time vehicle condition information to obtain a route prediction result;

and comparing the route prediction result with the driving intention of the driver represented by the electroencephalogram information, and controlling the self-vehicle to execute lane-changing driving action along the route prediction result when the route prediction result is consistent with the driving intention of the driver.

2. The automatic driving lane changing method according to claim 1, further comprising, before the obtaining the traffic information and the vehicle condition information:

acquiring offline training information, wherein the offline training information comprises offline road condition information and offline vehicle condition information;

constructing a decision table by using the offline training information;

discretizing the off-line training information in the decision table, and classifying the discretized off-line training information by using a preset neural network model to obtain a discrete classification result;

performing attribute reduction on the decision table according to the discrete classification result, and determining a minimum condition attribute set for expressing a decision;

integrating minimum condition attribute sets with the same expression decision result, and calculating the coverage and confidence level of the integrated minimum condition attribute sets to determine an offline decision rule;

and training the lane change decision model by using an offline decision rule, and comparing a route prediction result of the lane change decision model with offline electroencephalogram information.

3. The automatic driving lane-changing method according to claim 2, wherein the attribute reduction of the decision table according to the discrete classification, and the determination of the minimum condition attribute set for expressing the decision result comprises:

constructing a training data set according to the off-line training information, and randomly coding each type of off-line training information in the training data set;

calculating the degree of dependence of the decision attributes on off-line training information required by decision;

judging whether the current dependency degree is equal to the minimum relative dependency value;

if yes, outputting a minimum condition attribute set when the training data sets with the optimal fitness of a plurality of continuous generations are not improved any more;

if not, the training data set is subjected to random selection, crossover and mutation treatment iteratively, and the training data set with the optimal fitness is copied to the training data set of the next generation until the training data sets with the optimal fitness of a plurality of continuous generations are not improved any more and a minimum condition attribute set is output.

4. The automatic driving lane-changing method according to claim 1, wherein the electroencephalogram acquisition of the driver to obtain electroencephalogram information representing driving intention of the driver comprises:

collecting brain wave signals generated when a driver drives to obtain initial brain wave information;

preprocessing initial electroencephalogram information, and comparing the preprocessed initial electroencephalogram information by using a preset electroencephalogram analog template to obtain an electroencephalogram analog result;

and determining error potentials in the preprocessed initial electroencephalogram information according to the electroencephalogram analog result, and removing to obtain electroencephalogram information representing the driving intention of the driver.

5. The automated driving lane-change method according to claim 1, wherein the controlling of the host vehicle to perform the lane-change travel action along the route prediction result includes:

finding a plurality of pre-aiming points according to the distance between the self-vehicle and the barrier, and drawing a lane changing route by using a Bezier curve constructed by the pre-aiming points;

performing boundary expansion processing on the barrier, and setting the expansion radius of the barrier according to the current lane change route and the self-vehicle environment information;

judging whether the expansion radius of the barrier is larger than the width of the barrier;

if so, controlling the self vehicle to run along the current lane changing route;

if not, the position of the preview point is reset, a lane change line is redrawn by using a Bezier curve constructed by the reset preview point, the steering angle of the vehicle corresponding to the redrawn lane change line is larger than the steering angle of the vehicle corresponding to the last drawn lane change line, and the step of performing boundary expansion processing on the obstacle is returned.

6. The automated driving lane-change method according to claim 5, wherein the controlling the host vehicle to travel along the current lane-change route comprises:

when no continuous obstacles appear in the detection range, controlling the vehicle after lane changing to drive the obstacles and then changing to the lane of the initial driving route; when continuous obstacles appear in the detection range, controlling the self-vehicle after lane changing to drive the continuous obstacles and then changing to the lane of the initial driving route;

the continuous obstacles are obstacles, the distance between which and the previous obstacle in the same lane is less than a preset obstacle distance threshold value.

7. The automated driving lane change method of claim 1, further comprising:

and constructing a simulated driving scene according to the real-time road condition information and the real-time vehicle condition information, and displaying the simulated driving scene, wherein the simulated driving scene converts the real-time road condition information into simulated road condition information and converts the real-time vehicle condition information into simulated vehicle condition information.

8. An automatic driving lane-changing device, characterized in that the device comprises:

the system comprises an acquisition module, a processing module and a display module, wherein the acquisition module is used for acquiring real-time road condition information and real-time vehicle condition information, the road condition information comprises the number of lanes, the lane where a vehicle is located, the running vehicle and the obstacle distance representing the relative distance between the vehicle and the running vehicle or an obstacle, and the vehicle condition information comprises the speed and the positioning of the vehicle;

the acquisition module is used for carrying out electroencephalogram acquisition on a driver to obtain real-time electroencephalogram information, and the electroencephalogram information represents the driving intention of the driver;

the prediction module is used for predicting a driving route of the self-vehicle running on the current road condition by utilizing the trained lane-changing decision model based on the real-time road condition information and the real-time vehicle condition information to obtain a route prediction result;

and the control module is used for comparing the route prediction result with the driver driving intention represented by the electroencephalogram information and controlling the self-vehicle to execute lane-changing driving action along the route prediction result when the route prediction result is consistent with the driver driving intention.

9. An electronic device, comprising a memory storing a computer program and a processor that implements the autopilot lane-change method of any of claims 1-7 when executing the computer program.

10. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the automated driving lane-change method according to any one of claims 1 to 7.

Images