High beam and low beam self-adaptive switching device based on fuzzy neural network and working method of high beam and low beam self-adaptive switching device

Abstract

The invention discloses a high beam and low beam self-adaptive switching device based on a fuzzy neural network and a working method of the high beam and low beam self-adaptive switching device. The high beam and low beam adaptive switching device based on the fuzzy neural network comprises a high-precision map module, a vehicle sensor, a video data processing module, a light sensor, a driving behavior prediction module, a vehicle lamp controller, a T-BOX and a cloud parameter optimization calculation unit. The high-precision map module is used for acquiring information of roads in front of the current vehicle; the vehicle sensor is used for acquiring current vehicle driving information in real time; the video data processing module is used for acquiring road vehicle information; the light sensor is used for detecting the illumination intensity of the surrounding environment of the current vehicle; the driving behavior prediction module predicts the driving behavior; and the vehicle lamp controller makes a high beam and low beam switching strategy. The method has the advantages of accurately and automatically switching the vehicle lamps and improving the driving comfort.

Description

Technical Field

[0001] The present invention belongs to the technical field of vehicle lamp control, specifically relates to the adaptive switching of high and low beam headlights, and particularly relates to an adaptive switching device for high and low beam headlights based on a fuzzy neural network and its working method. Background Art

[0002] Traditional switching of high and low beam headlights requires manual operation by the driver and cannot be intelligently adjusted according to road and traffic conditions. If the driver operates improperly, it will lead to limited visibility when using low beam headlights in low-light conditions at night; or cause glare to other road users when using high beam headlights in sufficient light conditions at night, increasing the risk of traffic accidents. In recent years, with the development of intelligent driving and vehicle networking technologies, automatic switching of high and low beam headlights can be achieved. However, the existing technologies do not consider the subjective driving decisions of different drivers, resulting in incorrect switching problems and affecting the driving experience. To solve such problems, an adaptive switching device for high and low beam headlights based on a fuzzy neural network and its working method are proposed. Summary of the Invention

[0003] The present invention aims to solve at least one of the technical problems existing in the prior art.

[0004] Therefore, the present invention provides an adaptive switching device for high and low beam headlights based on a fuzzy neural network and its working method. The adaptive switching device for high and low beam headlights based on a fuzzy neural network has the advantages of accurately and automatically switching vehicle lamps and improving driving comfort.

[0005] An adaptive high and low beam switching device based on a fuzzy neural network according to an embodiment of the present invention includes: a high-precision map module, a vehicle sensor, a video data processing module, a light sensor, a driving behavior prediction module, a vehicle lamp controller, a T-BOX, and a cloud parameter optimization calculation unit; the high-precision map module is used to obtain road information in front of the current vehicle; the vehicle sensor is used to obtain real-time driving information of the current vehicle; the video data processing module is connected to a multi-functional camera, and the video data processing module is used to obtain road vehicle information in combination with the current vehicle driving information; the light sensor is used to detect the illumination intensity of the surrounding environment of the current vehicle; the driving behavior prediction module predicts driving behavior based on the long short-term memory network algorithm according to the road information in front of the current vehicle and the road vehicle information; the vehicle lamp controller makes a high and low beam switching strategy based on the road information, road vehicle information, driving behavior prediction result, and illumination intensity, based on the fuzzy neural network algorithm whose parameters have reached the optimal solution; the vehicle lamp controller is connected to a high beam driving module and a low beam driving module, the high beam driving module is connected to a high beam lamp, the low beam driving module is connected to a low beam lamp, and the high beam driving module and the low beam driving module are used to control the high beam lamp and the low beam lamp to execute the high and low beam switching strategy; the T-BOX is used for data interaction between the vehicle electronic control unit and the cloud server; the cloud parameter optimization calculation unit is used to combine the road information in front of the current vehicle, the road vehicle information, and the current vehicle driving information, perform parameter training on different data, realize customized optimization of the long short-term memory network algorithm parameters, and update the algorithm parameters to the driving behavior prediction module through the T-BOX.

[0006] A working method of an adaptive high and low beam switching device based on a fuzzy neural network, using the adaptive high and low beam switching device based on a fuzzy neural network described in any one of the above, includes the following steps:

[0007] S1. Obtain road information;

[0008] S2. Obtain road vehicle information;

[0009] S3. Obtain illumination intensity;

[0010] S4. Predict driving behavior prediction based on the long short-term memory network;

[0011] S5. Perform a high and low beam switching strategy based on the fuzzy neural network;

[0012] S6. Control the high beam and low beam lamps.

[0013] According to an embodiment of the present invention, the optimization method of the long short-term memory network parameters is:

[0014] S41. Obtain road information;

[0015] S42. Obtain road vehicle information;

[0016] S43. Obtain vehicle driving information; (vehicle speed, steering information)

[0017] S44. Forward the above information to the T-BOX;

[0018] S45. The T-BOX sends the above information to the cloud server;

[0019] S46. The cloud parameter optimization calculation unit performs parameter optimization;

[0020] S47. The cloud server sends the optimized parameters to the T-BOX;

[0021] S48. The T-BOX forwards the received optimized parameters to the driving behavior prediction module;

[0022] S49. The driving behavior prediction module updates the parameters.

[0023] According to an embodiment of the present invention, based on the constructed data set and the high and low beam switching instruction, the backpropagation algorithm is selected as the parameter learning algorithm of the fuzzy neural network, and the fuzzy neural network is trained offline for parameters.

[0024] According to an embodiment of the present invention, the road information in front of the current vehicle includes: lane type, number of lanes, and traffic direction.

[0025] According to an embodiment of the present invention, the road vehicle information includes: vehicle type, relative speed of the vehicle to the current vehicle, and vehicle position.

[0026] According to an embodiment of the present invention, the backpropagation algorithm includes the following steps:

[0027] S51, Parameter initialization setting;

[0028] S52, Information forward propagation, calculate the network error under the current parameters;

[0029] S53, Weight update, membership function center update, membership function width vector update;

[0030] S54, Update the parameters according to the network error, check whether the accuracy requirement is met; (reach the iteration number or obtain the optimal solution)

[0031] S55, Result analysis and output.

[0032] According to an embodiment of the present invention, the fuzzy neural network includes an input layer, a membership layer, a rule layer, a normalization layer, and an output layer; the inputs of the input layer are respectively: driving behavior prediction (x1), light intensity (x2), road information (x3), and road vehicle information (x4, x5, x6, x 12 ); the output of the output layer is: the switching instruction of the high and low beam lights (y).

[0033] According to an embodiment of the present invention, the calculation method of the network error is as follows:

[0034] Membership function:

[0035] where u ij is the node output, x i is the input data, c ij is the center point of the membership function, b ij is the width vector of the membership function; m is the number of fuzzy classifications of the input;

[0036] Then the data is passed to the rule layer, and the layer function is: a j =Πu ij ;

[0037] Next, the output of the rule layer is normalized in the normalization layer, and the normalization function is:

[0038]

[0039] Finally, weighted processing is performed in the output layer to obtain the final output, and the weighting function is: where, w j is the connection weight value of the output layer, the expected output value is y r , then the network error is:

[0040] The beneficial effects of the present invention are as follows: The present invention uses a driving behavior prediction module to predict the driving behavior of the driver based on the long short-term memory network by obtaining road information and road vehicle information, and then controls the high and low beam lights based on the light intensity, driving behavior prediction, road information, and road vehicle information through the fuzzy neural network algorithm, which solves the problem of frequent switching or inaccurate switching and improves the accuracy of light switching; by obtaining daily driving data, uploading it to the cloud through the T-BOX, the cloud computing unit optimizes the parameters of the long short-term memory network, and updates it to the driving prediction module through the T-BOX, so as to realize the personalized customization of the switching strategy and improve driving comfort.

[0041] Other features and advantages of the present invention will be described in the subsequent specification, and some of them will become obvious from the specification or be understood by implementing the present invention.

[0042] In order to make the above objects, features, and advantages of the present invention more obvious and understandable, the following provides preferred embodiments in conjunction with the accompanying drawings and describes them in detail as follows. Description of the Drawings

[0043] The above and / or additional aspects and advantages of the present invention will become obvious and easy to understand from the description of the embodiments in conjunction with the following drawings, where:

[0044] Figure 1 is a schematic diagram of the overall architecture of the present invention;

[0045] Figure 2 is a schematic diagram of the working method steps of the high and low beam adaptive switching device of the present invention;

[0046] Figure 3 is a schematic diagram of the long short-term memory network structure of the present invention;

[0047] Figure 4 is a schematic diagram of the fuzzy neural network structure of the present invention;

[0048] Figure 5 is a schematic diagram of the backpropagation algorithm flow of the present invention;

[0049] Figure 6 is a schematic diagram of the road condition of the present invention;

[0050] Figure 7 is a schematic diagram of the road vehicle information of the present invention;

[0051] Figure 8 is a schematic diagram of the driving behavior prediction information of the present invention;

[0052] Figure 9 is a schematic diagram of the long short-term memory network parameter optimization process of the present invention; Detailed Embodiments

[0053] The following details the embodiments of the present invention. The examples of the embodiments are shown in the drawings, where the same or similar reference numerals represent the same or similar elements or elements with the same or similar functions from beginning to end. The embodiments described below by referring to the drawings are exemplary and are only used to explain the present invention and should not be construed as a limitation of the present invention.

[0054] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by the terms "center", "longitudinal", "transverse", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention. In addition, the features defined as "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present invention, unless otherwise specified, the meaning of "a plurality" is two or more.

[0055] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "mounted", "connected" and "coupled" shall be construed broadly. For example, it may be a fixed connection, a detachable connection, or an integral connection; it may be a mechanical connection or an electrical connection; it may be a direct connection or an indirect connection through an intermediate medium, and it may be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention may be understood according to specific circumstances.

[0056] The following specifically describes a high and low beam adaptive switching device based on a fuzzy neural network and its working method according to an embodiment of the present invention with reference to the drawings.

[0057] As Figures 1-9As shown in the figure, the high and low beam adaptive switching device based on a fuzzy neural network according to an embodiment of the present invention includes: a high-precision map module, a vehicle sensor, a video data processing module, a light sensor, a driving behavior prediction module, a vehicle lamp controller, a T-BOX, and a cloud parameter optimization calculation unit; the high-precision map module is used to obtain the road information in front of the current vehicle, including but not limited to lane types: straight lanes, curved lanes, the number of lanes: single lane, double lanes, three lanes and above, traffic directions: one-way, two-way; the vehicle sensor is used to obtain the current vehicle driving information in real time, including but not limited to the vehicle speed and the vehicle steering information; the video data processing module is connected to a multi-functional camera, and the video data processing module is used to obtain road vehicle information in combination with the current vehicle driving information, including but not limited to vehicle types: large vehicles, small vehicles, relative vehicle speed to the vehicle: low speed, same speed, high speed, vehicle positions: directly in front of the same lane, left front of the same lane, right front of the same lane, left side of the same lane, right side of the same lane, directly behind the same lane, left rear of the same lane, right rear of the same lane, oncoming lane behind, oncoming lane in front; the light sensor is used to detect the light intensity of the surrounding environment of the current vehicle, and the detected light intensity value is used as the judgment input condition for the system to control the low beam and high beam; the driving behavior prediction module predicts the driving behavior based on the long short-term memory network algorithm according to the road information in front of the current vehicle and the road vehicle information, and the prediction results include but not limited to lane changing, accelerating, decelerating, maintaining the current driving state; the vehicle lamp controller makes a high and low beam switching strategy based on the road information, the road vehicle information, the driving behavior prediction results, and the light intensity, based on the fuzzy neural network algorithm with the parameters having reached the optimal solution; the vehicle lamp controller is connected to a high beam driving module and a low beam driving module, the high beam driving module is connected to the high beam, the low beam driving module is connected to the low beam, and the high beam driving module and the low beam driving module are used to control the high beam and the low beam to execute the high and low beam switching strategy; the T-BOX is used for data interaction between the vehicle electronic control unit (including but not limited to: high-precision map module, video data processing module, driving behavior prediction module, vehicle sensor) and the cloud server; the cloud parameter optimization calculation unit is used to combine the road information in front of the current vehicle, the road vehicle information, and the current vehicle driving information, perform parameter training on different data, realize the customized optimization of the long short-term memory network algorithm parameters, and update the algorithm parameters to the driving behavior prediction module through the T-BOX.

[0058] In this embodiment, the driving behavior prediction module obtains road information and road vehicle information, predicts the driver's driving behavior based on the long short-term memory network, and then controls the high and low beam lights based on the light intensity, driving behavior prediction, road information, and road vehicle information using the fuzzy neural network algorithm, solving the problem of frequent or inaccurate switching and improving the accuracy of light switching. By obtaining daily driving data, uploading it to the cloud through T-BOX, optimizing the parameters of the long short-term memory network by the cloud computing unit, and updating it to the driving prediction module through T-BOX, personalized customization of the switching strategy is achieved, improving driving comfort.

[0059] Specifically, as Figure 3 shown, the structure of the long short-term memory network algorithm includes: a forgetting gate, an input gate, and an output gate. By analyzing the changing trend of the influence of road information and road vehicle information on driving behavior in the time dimension, the prediction of the driver's driving behavior is realized. The input is: road information (x3), road vehicle information (x4~x 12 ); the output is: driving behavior prediction (x1).

[0060] As Figure 2 shown, a working method of an adaptive high and low beam light switching device based on a fuzzy neural network uses the above-mentioned adaptive high and low beam light switching device based on a fuzzy neural network, and includes the following steps:

[0061] S1. Obtain the road information in front of the current vehicle; as Figure 6 shown, define the road information as x3. According to the road information, after data integration and processing, the corresponding road condition table is obtained; the actual application conditions include but are not limited to the content described in this table, the traffic direction. In a two-way road with a median strip, it is considered that the vehicle's lights will not affect oncoming vehicles, so the traffic direction is determined to be one-way.

[0062] For example: when the front condition is: straight road, four lanes in both directions, no median strip, x3 = 5; when the front condition is: curve, two lanes in both directions, with a median strip, x3 = 7.

[0063] S2. Obtain the road vehicle information; as Figure 7 shown, according to the road vehicle information, after data integration and processing, the corresponding road vehicle information status is obtained. The subsequent algorithm uses the road vehicle information number to represent the road vehicle information; define: the status of the vehicle directly in front in the same lane (x4), the status of the vehicle diagonally in front on the left in the same lane (x5), the status of the vehicle diagonally in front on the right in the same lane (x6), the status of the vehicle on the left side in the same lane (x7), the status of the vehicle on the right side in the same lane (x8), the status of the vehicle directly behind in the same lane (x9), the status of the vehicle diagonally behind on the left in the same lane (x 10) Status of the vehicle behind on the right side of the same-direction lane (x 11 ) Status of the vehicle in the oncoming lane ahead (x 12 ).

[0064] For example: If it is determined that there is a large vehicle with the same speed directly in front of the vehicle in the same lane, then x4 = 35; if it is determined that there is no vehicle in front in the left lane of the vehicle, then x5 = 1.

[0065] If the distance between the vehicle and the vehicle ahead exceeds the switching distance of the high and low beam lights, it is considered that there is no vehicle driving at this position.

[0066] S3. Obtain the light intensity;

[0067] S4. Predict driving behavior based on the long short-term memory network. The prediction results include but are not limited to lane change driving, accelerating, decelerating, and maintaining the current driving state; as Figure 8 shown, define x t as the input vector at the current moment; define h t-1 as the output of the hidden state at the previous moment; define C t-1 as the cell state at the previous moment; define h t as the output of the hidden state at the current moment; define C t as the updated cell state at the current moment. The output formula of the forget gate is: f t = σ(W f *[h t-1 , x t +b f );

[0068] Among them, W f is the weight matrix of the forget gate; b f is the bias term of the forget gate; σ is the Sigmoid activation function, and the output value is between 0 and 1;

[0069] The output formula of the input gate is: i t = σ(W i / [h t-1 , x t +b i );

[0070] The formula for the candidate memory cell is:

[0071] Among them, W i is the weight matrix of the input gate; b i is the bias term of the input gate; W c is the weight matrix of the candidate memory cell; b C is the bias term of the candidate memory cell; tanh is the activation function, which compresses the candidate memory to the range of [-1, 1];

[0072] The output formula of the output gate is: o t = σ(W o * [h t-1 , x t + b o );

[0073] Among them, W o is the weight matrix of the output gate; b o is the bias term of the output gate;

[0074] The updated cell state is:

[0075] The hidden state at the current moment is: h t = o t tanh(C t ).

[0076] S5. Based on the fuzzy neural network, perform the high and low beam switching strategy;

[0077] S6. Control the high beam and low beam.

[0078] As Figure 9 shown, the optimization method of the long short-term memory network parameters is:

[0079] S41. Obtain road information;

[0080] S42. Obtain road vehicle information;

[0081] S43. Obtain vehicle driving information; (vehicle speed, steering information)

[0082] S44. Forward the above information to the T-BOX;

[0083] S45. The T-BOX sends the above information to the cloud server;

[0084] S46. The cloud parameter optimization calculation unit performs parameter optimization;

[0085] S47. The cloud server sends the optimized parameters to the T-BOX;

[0086] S48. The T-BOX forwards the received optimized parameters to the driving behavior prediction module;

[0087] S49. The driving behavior prediction module updates the parameters.

[0088] For example: In working condition 1, when the light intensity x2 is less than 50 lux, the road information x3 = 6, the road vehicle information x4 = 1, x5 = 1, x6 = 1, x 12= 1; that is, the light intensity meets the condition for the vehicle to turn on the high beam. The road ahead is a six-lane two-way road, and there are no vehicles moving ahead. The headlight controller makes a decision to turn on the high beam according to the input.

[0089] Operating condition 2: When the light intensity x2 is less than 50 lux, the road information x3 = 6, and the vehicle information on the road x4 = 27, x5 = 1, x6 = 29, x7 = 1, x8 = 1, x9 = 1, x 10 = 1, x 11 = 1, x 12 = 1; that is, the road ahead is a six-lane two-way road, and there is a large vehicle moving slowly directly ahead in the same lane and a large vehicle moving slowly in the right front of the same lane. The driving behavior prediction module makes a driving behavior prediction based on the parameters obtained from training with historical data according to the above operating condition. In a time series, assume the predicted driving behavior: the vehicle changes lanes (x1 = 1), accelerates (x1 = 2), maintains the current driving state (x1 = 4); that is, the vehicle switches to the left lane and overtakes. The headlight controller makes a high and low beam control decision according to the prediction result and other input information: the light is switched to the low beam. After changing lanes, two high and low beam switching controls are performed. After overtaking is completed, the high beam is turned on and the low beam is turned off.

[0090] If the above operating condition 2 makes a driving prediction behavior according to different optimization parameters: the vehicle changes lanes (x1 = 1), maintains the current driving state (x1 = 4); that is, the vehicle switches to the left lane to drive. Then the headlight controller makes a high and low beam control decision according to the prediction result and other input information: the light is switched to the low beam.

[0091] As Figure 5 shown, the backpropagation algorithm includes the following steps:

[0092] S51, parameter initialization setting;

[0093] S52, forward propagation of information, calculating the network error under the current parameters;

[0094] S53, weight update, membership function center update, membership function width vector update;

[0095] Weight update formula:

[0096] Membership function center update formula:

[0097] Membership function width vector update formula: where μ is the learning rate.

[0098] S54, updating the parameters according to the network error, checking whether the accuracy requirement is met; reaching the iteration number or obtaining the optimal solution;

[0099] S55, Result analysis and output.

[0100] As Figure 4 shown, the fuzzy neural network includes an input layer, a membership layer, a rule layer, a normalization layer, and an output layer. The inputs of the input layer are respectively: driving behavior prediction (x1), light intensity (x2), road information (x3), and road vehicle information (x4, x5, x6, x 12 ); the output of the output layer is: the switching instruction of the high and low beam lights (y).

[0101] The calculation method of the network error is:

[0102] Membership function:

[0103] where u ij is the node output, x i is the input data, c ij is the center point of the membership function, b ij is the width vector of the membership function; m is the number of fuzzy classifications of the input;

[0104] Then the data is passed to the rule layer, and the layer function is: a j = Πu ij ;

[0105] Next, the output of the rule layer is normalized in the normalization layer, and the normalization function is:

[0106] Finally, weighted processing is performed in the output layer to obtain the final output, and the weighting function is:

[0107] where, w j is the connection weight value of the output layer, the expected output value is y t , then the network error is

[0108]

[0109] Based on the constructed data set and the switching instruction of the high and low beam lights, the backpropagation algorithm is selected as the parameter learning algorithm of the fuzzy neural network, and the fuzzy neural network is trained offline for parameters.

[0110] In the description of this specification, the descriptions referring to terms such as "one embodiment", "some embodiments", "schematic embodiments", "examples", "specific examples", or "some examples", etc. mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0111] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the claims and their equivalents.

Claims

1. A high and low beam adaptive switching device based on fuzzy neural network, characterized in that: include: A high-precision map module, which is used to obtain the road information ahead of the current vehicle; A vehicle sensor, wherein the vehicle sensor is used to obtain current vehicle driving information in real time; A video data processing module, the video data processing module is connected to a multifunctional camera, and the video data processing module is used to obtain road vehicle information in combination with current vehicle driving information; A light sensor, the light sensor is used to detect the light intensity of the current vehicle surrounding environment; A driving behavior prediction module, which predicts driving behavior based on a long short-term memory network algorithm according to the current vehicle front road information and road vehicle information; A headlight controller, wherein the headlight controller makes a high-beam and low-beam switching strategy based on the fuzzy neural network algorithm whose parameters have reached an optimal solution according to the road information, road vehicle information, driving behavior prediction results and light intensity; the headlight controller is connected to a high-beam driving module and a low-beam driving module, the high-beam driving module is connected to the high-beam, the low-beam driving module is connected to the low-beam, and the high-beam driving module and the low-beam driving module are used to control the high-beam and the low-beam to execute the high-beam and low-beam switching strategy; T-BOX, which is used for data exchange between the vehicle electronic control unit and the cloud server; The cloud-based parameter optimization calculation unit is used to combine the current vehicle front road information, road vehicle information and current vehicle driving information to perform parameter training on different data, realize customized optimization of long short-term memory network algorithm parameters, and update the algorithm parameters to the driving behavior prediction module through T-BOX.

2. A working method of a high and low beam adaptive switching device based on a fuzzy neural network, characterized in that: The high and low beam adaptive switching device based on fuzzy neural network as claimed in claim 1 comprises the following steps: S1. Obtain road information; S2. Obtain road vehicle information; S3. Obtaining light intensity; S4. Driving behavior prediction based on long short-term memory network; S5. High and low beam switching strategy based on fuzzy neural network; S6. High beam and low beam control.

3. The working method of the high and low beam adaptive switching device based on fuzzy neural network according to claim 2 is characterized in that: The optimization method of the long short-term memory network parameters is: S41. Obtain road information; S42. Obtain road vehicle information; S43. Obtain vehicle driving information; S44. Forward the above information to T-BOX; S45.T-BOX sends the above information to the cloud server; S46. The cloud parameter optimization calculation unit performs parameter optimization; S47. The cloud server sends the optimized parameters to T-BOX; S48.T-BOX forwards the received optimized parameters to the driving behavior prediction module; S49. The driving behavior prediction module updates parameters.

4. The working method of the high and low beam adaptive switching device based on fuzzy neural network according to claim 2 is characterized in that: Based on the constructed data set and the high and low beam switching instructions, the back propagation algorithm is selected as the parameter learning algorithm of the fuzzy neural network, and the fuzzy neural network is trained offline.

5. The working method of the high and low beam adaptive switching device based on fuzzy neural network according to claim 2 is characterized in that: The current road information ahead of the vehicle includes: lane type, lane number and travel direction.

6. The working method of the high and low beam adaptive switching device based on fuzzy neural network according to claim 2 is characterized in that: Road vehicle information includes: vehicle type, speed relative to the current vehicle, and vehicle position.

7. The working method of the high and low beam adaptive switching device based on fuzzy neural network according to claim 4 is characterized in that: The back-propagation algorithm consists of the following steps: S51, parameter initialization setting; S52, information forward propagation, calculating the network error under the current parameters; S53, weight update, membership function center update, membership function width vector update; S54, updating parameters according to the network error and checking whether the accuracy requirements are met; S55, result analysis and output.

8. The working method of the high and low beam adaptive switching device based on fuzzy neural network according to claim 7 is characterized in that: The fuzzy neural network includes an input layer, a membership layer, a rule layer, a normalization layer, and an output layer; the input layer inputs are: driving behavior prediction, light intensity, road information, and road vehicle information; the output layer outputs: high and low beam switching instructions.

9. The working method of the high and low beam adaptive switching device based on fuzzy neural network according to claim 8 is characterized in that: The network error is calculated as: Membership function: where u ij is the node output, x i is the input data, c ij is the center point of the membership function, b ij is the width vector of the membership function; m is the number of fuzzy classifications of the input; The data is then passed to the rule layer, and the layer function is: a j =Πu ij ; Then the output of the regular layer is normalized in the normalization layer, and the normalization function is: Finally, weighted processing is performed in the output layer to obtain the final output. The weighted function is: Among them, w j is the connection weight of the output layer, and the expected output value is y r , then the network error is

Images