Vehicle lane change prediction method, device and electronic equipment

Abstract

The application is suitable for the field of vehicle lane change prediction technology, and provides a vehicle lane change prediction method, device and electronic equipment, which comprises the following steps: acquiring a motion vector of a target vehicle, the motion vector comprising a tangential velocity and / or a displacement; calculating a vector entropy of the motion vector, the vector entropy being used to describe a stability degree of a motion trend of the target vehicle; acquiring an RSSI of the target vehicle; inputting the motion vector, the vector entropy and the RSSI into a trained decision tree model to obtain a prediction result output by the trained decision tree model, the prediction result being used to indicate whether the target vehicle changes lanes. Through the above method, the accuracy of the prediction result can be improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of vehicle lane change prediction, and particularly relates to a vehicle lane change prediction method and device, an electronic device, and a computer readable storage medium. BACKGROUND

[0002] During vehicle driving, if a vehicle suddenly changes lanes, it will pose a certain safety threat to other vehicles. In order to improve the safety of vehicle driving, it is necessary to predict lane changes of the vehicle.

[0003] The existing vehicle lane change prediction method, after obtaining the corresponding point cloud graph by using a laser radar, performs background filtering, clustering and other processing on the point cloud graph, and then predicts lane changes of vehicles in transit. Since the performance of the laser radar is poor in extreme weather such as rain, snow and fog, the accuracy of the obtained lane change prediction result is low. SUMMARY

[0004] The vehicle lane change prediction method, device and electronic device provided by the embodiments of the present application can solve the problem of low accuracy of the lane change prediction result.

[0005] In a first aspect, the embodiments of the present application provide a vehicle lane change prediction method, comprising:

[0006] obtaining a motion vector of a target vehicle, the motion vector comprising a tangential velocity and / or a displacement;

[0007] calculating a vector entropy of the motion vector, the vector entropy being used to describe a stability degree of a motion trend of the target vehicle;

[0008] obtaining an RSSI of the target vehicle;

[0009] inputting the motion vector, the vector entropy and the RSSI into a trained decision tree model to obtain a prediction result output by the trained decision tree model, the prediction result being used to indicate whether the target vehicle changes lanes, wherein the trained decision tree model is obtained by training according to a data set corresponding to a group of vehicles with known lane change or not, and the data set comprises motion vector samples, vector entropy samples and RSSI samples of the vehicles with known lane change or not.

[0010] In a second aspect, the embodiments of the present application provide a vehicle lane change prediction device, comprising:

[0011] a motion vector obtaining module, configured to obtain a motion vector of a target vehicle, the motion vector comprising a tangential velocity and / or a displacement;

[0012] a vector entropy calculating module, configured to calculate a vector entropy of the motion vector, the vector entropy being used to describe a stability degree of a motion trend of the target vehicle;

[0013] an RSSI acquisition module, configured to acquire an RSSI of the target vehicle;

[0014] a prediction result output module, configured to input the motion vector, the vector entropy and the RSSI into a trained decision tree model to obtain a prediction result output by the trained decision tree model, the prediction result being used to indicate whether the target vehicle changes lane, wherein the trained decision tree model is trained according to a data set corresponding to a group of vehicles with known lane changing or not, and the data set includes motion vector samples, vector entropy samples and RSSI samples of the vehicles with known lane changing or not.

[0015] In a third aspect, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the method according to the first aspect when executing the computer program.

[0016] In a fourth aspect, a computer readable storage medium is provided, and the computer readable storage medium stores a computer program, and the computer program is executable on a processor to implement the method according to the first aspect.

[0017] In a fifth aspect, a computer program product is provided, and when the computer program product is executed on an electronic device, the electronic device executes the method according to the first aspect.

[0018] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0019] In the embodiments of the present application, since the received signal strength indication (RSSI) can reflect the distance between the measuring signal point (i.e. the target vehicle) and the receiving point, the RSSI of the target vehicle can reflect the relative position of the target vehicle and the measuring signal point, i.e. the position of the target vehicle. At the same time, since the motion vector such as the cutting speed and / or displacement of the target vehicle can reflect the motion trend of the target vehicle, the vector entropy corresponding to the motion vector can reflect the stability degree of the motion trend of the target vehicle, and when the vehicle changes lane, its motion trend will inevitably be different from the original motion trend, and the motion trend will continue to be stable for a period of time. Therefore, after inputting the RSSI, the motion vector and the corresponding vector entropy of the target vehicle into the trained decision tree model, an accurate prediction result can be obtained, i.e. the method can accurately predict whether the target vehicle changes lane. BRIEF DESCRIPTION OF DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiments or the prior art description will be briefly introduced below.

[0021] Figure 1 is a flowchart of a vehicle lane changing prediction method provided by an embodiment of the present application.

[0022] Figure 2 is an installation schematic diagram of a radar, RSU and roadside equipment provided by an embodiment of the present application.

[0023] Figure 3 is a structural schematic diagram of a decision tree model provided by an embodiment of the present application.

[0024] Figure 4 is a flowchart of whether a lane changing behavior is abnormal provided by an embodiment of the present application.

[0025] Figure 5 is a structural schematic diagram of a vehicle lane changing prediction device provided by an embodiment of the present application.

[0026] Figure 6 is a structural schematic diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0027] In the following description, for the purpose of explanation and not limitation, specific details are set forth, such as particular system configurations, techniques, etc., in order to provide a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application can be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.

[0028] It should be understood that the term "comprises" when used in this specification and the appended claims indicates the presence of the described features, integers, steps, operations, elements, and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0029] It should also be understood that the term "and / or" when used in this specification and the appended claims indicates that the associated listed items can be present one or more of the associated listed items, and that the combinations of the associated listed items are also included.

[0030] In addition, in the description of the present application and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish descriptions, and cannot be understood as indicating or implying relative importance.

[0031] Reference to "one embodiment" or "some embodiments" or "one implementation" or "some implementations" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment. The appearances of the phrase "in one embodiment" or "in some embodiments" or "in other embodiments" or "in still other embodiments" or other similar phrases in the specification are not necessarily all referring to the same embodiment.

[0032] Example One

[0033] In the process of lane change prediction of a vehicle, if only laser radar is used for prediction, it is difficult to obtain an accurate prediction result.

[0034] In order to improve the accuracy of the obtained prediction result of the lane change of the vehicle, an embodiment of the present application provides a vehicle lane change prediction method. In the method, whether a target vehicle will change lanes is predicted through the motion vector of the target vehicle, the vector entropy of the motion vector, and the received signal strength indication (RSSI) of the target vehicle.

[0035] Since the motion vector such as the tangential velocity and / or displacement of the target vehicle can reflect the motion trend of the target vehicle, the vector entropy corresponding to the motion vector can reflect the stability degree of the motion trend of the target vehicle, and the RSSI of the target vehicle can reflect the position of the target vehicle, therefore, according to the above information, whether the target vehicle will change lanes can be accurately judged.

[0036] The vehicle lane change prediction method provided by the embodiment of the present application is described below in combination with the drawings.

[0037] Figure 1 A flowchart of a vehicle lane change prediction method provided by an embodiment of the present application is shown, and the details are as follows:

[0038] In step S11, the motion vector of the target vehicle is obtained, and the motion vector includes the tangential velocity and / or displacement.

[0039] The tangential velocity refers to the velocity of the target vehicle in the direction of the line connecting the target vehicle and the receiving point. When radar is used for detection, the tangential velocity is the velocity of the target vehicle in the direction of the line connecting the target vehicle and the radar, which can be directly obtained by radar detection, that is, the data on the radar data frame obtained by radar detection can be processed to obtain the tangential velocity.

[0040] The displacement can be determined according to the coordinates of the target vehicle at two different positions. Specifically, the coordinates of the target vehicle at different positions are obtained through data in different radar data frames, and the corresponding displacement is determined according to the coordinates. For example, it is assumed that at time t1, the target vehicle is at position 1 (x1, y1), and at the next time (assuming time t2), the target vehicle is at position 2 (x2, y2), and the corresponding displacement of the target vehicle is (x2-x1, y2-y1). The displacement can be determined by subtracting the coordinates at position 2 and position 1.

[0041] In some embodiments, the radar described above is a millimeter wave radar. Compared with laser radar, the measurement range of the millimeter wave radar is larger, the cost is lower, and the millimeter wave radar is not easily affected by the weather environment, that is, the detection result of the millimeter wave radar will not be affected by cloudy, rainy, snowy and foggy weather. Therefore, when the millimeter wave radar is used to obtain the tangential velocity and other information, the accuracy of the tangential velocity and other information obtained by the millimeter wave radar can be ensured, thereby improving the accuracy of the subsequent prediction results.

[0042] In step S12, the vector entropy of the motion vector is calculated. The vector entropy is used to describe the stability of the motion trend of the target vehicle.

[0043] Specifically, when the motion vector includes the tangential velocity, the vector entropy of the tangential velocity needs to be calculated; when the motion vector includes the displacement, the vector entropy of the displacement needs to be calculated.

[0044] In this embodiment, the vector entropy of the motion vector is determined by using a preset calculation method that can reflect the stability of the motion trend.

[0045] In step S13, the RSSI of the target vehicle is obtained.

[0046] In this embodiment, when the target vehicle enters a binding area corresponding to a dedicated short range communication (DSRC), a road side unit (RSU) will establish communication with an on board unit (OBU) of the target vehicle. Specifically, the RSU reads the electronic tag of the OBU, obtains the vehicle type, license plate information and other information of the vehicle (i.e. the target vehicle) where the OBU is located, and obtains the RSSI of the target vehicle.

[0047] The RSU is a device installed on the roadside, which uses the DSRC technology to communicate with the OBU, realizes vehicle identity recognition and electronic demerit.

[0048] In some embodiments, the RSU and the radar can be managed and configured information is issued by the roadside device, that is, the roadside device as the upper machine of the RSU and the radar. The roadside device, the radar and the RSU are usually installed on the gantry or side pole at a certain distance on the road. Figure 2 A schematic diagram of the installation of a radar, an RSU and a roadside device is shown, in Figure 2 In the above embodiment, only one roadside device is shown, but in actual situations, multiple roadside devices can be deployed as needed.

[0049] In step S14, the motion vector, the vector entropy and the RSSI are input into the trained decision tree model to obtain a prediction result output by the trained decision tree model, the prediction result being used to indicate whether the target vehicle changes lanes, wherein the trained decision tree model is trained according to a data set corresponding to a group of vehicles whose lane changing is known, and the data set includes motion vector samples, vector entropy samples and RSSI samples of the vehicles whose lane changing is known.

[0050] The motion vector sample refers to a sample with a motion vector, the vector entropy sample refers to a sample with a vector entropy, and the RSSI sample refers to a sample with an RSSI.

[0051] Suppose the motion vector sample includes a tangential velocity and a displacement, the samples in the data set can be as shown in Table 1.

[0052] Table 1:

[0053]

[0054] In the above embodiment, since the decision tree model is trained according to a data set including motion vector samples, vector entropy samples and RSSI samples of vehicles whose lane changing is known, when the motion vector, the vector entropy and the RSSI of the target vehicle are input into the trained decision tree model, a prediction result corresponding to the target vehicle can be obtained.

[0055] In the above embodiment, since the RSSI can reflect the distance between the measurement signal point (i.e. the target vehicle) and the receiving point, the RSSI of the target vehicle can reflect the relative position of the target vehicle and the measurement signal point, i.e. the position of the target vehicle. Meanwhile, since the motion vector such as the tangential velocity and / or the displacement of the target vehicle can reflect the motion trend of the target vehicle, the vector entropy corresponding to the motion vector can reflect the stability of the motion trend of the target vehicle, and when the vehicle changes lanes, its motion trend will inevitably be different from the original motion trend, and the motion trend will continue to be stable for a period of time. Therefore, when the RSSI, the motion vector and the corresponding vector entropy of the target vehicle are input into the trained decision tree model, an accurate prediction result can be obtained, i.e. the target vehicle can be accurately predicted to change lanes or not by the above method.

[0056] It should be noted that the vehicle lane change prediction method of the embodiments of the present application can be applied in a radar, can also be applied in an RSU, and can also be applied in a roadside device.

[0057] In some embodiments, the vehicle lane change prediction method described above further includes: constructing a decision tree model to be trained, and training the decision tree model to be trained according to a data set corresponding to a group of vehicles whose lane changes are known, to obtain a trained decision tree model. Specifically, after obtaining a data set corresponding to a group of vehicles whose lane changes are known, part of the data in the data set (for example, 70% of the data) can be used to construct a decision tree model to be trained, and the remaining data in the data set (for example, 30% of the data) can be used to train the constructed decision tree model to be trained.

[0058] The decision tree model to be trained can be constructed in the following manner:

[0059] A1, respectively calculating the lane change probability and the no-lane change probability of the motion vector sample, the vector entropy sample and the RSSI sample.

[0060] The lane change probability of the motion vector sample refers to the probability of the target vehicle changing lanes under the condition of the data corresponding to the motion vector sample. The no-lane change probability of the motion vector sample refers to the probability of the target vehicle not changing lanes under the condition of the data corresponding to the motion vector sample.

[0061] A2, determining the class entropy of the motion vector sample according to the lane change probability and the no-lane change probability of the motion vector sample, wherein the class entropy is a value obtained by modifying the information entropy.

[0062] In this embodiment, it is assumed that N i is the confidence degree corresponding to the i-th frame of radar data, which is used to describe the reliability of the i-th frame of radar data, N max is the maximum value in the confidence degrees corresponding to the plurality of or all frames of radar data, then the class entropy is calculated according to the following formula: -(p*log2p+q*log2q)*N i / N maxwherein, “-(p*log2p+q*log2q)” represents information entropy. When calculating the class entropy, it is agreed that if p=0, then p*log2p=0, and p and q represent the variable lane probability and the invariable lane probability of the corresponding attribute respectively. For example, when calculating the class entropy of the motion vector sample, the above p represents the variable lane probability of the motion vector sample, and q represents the invariable lane probability of the motion vector sample. When calculating the class entropy of the vector entropy sample, the above p represents the variable lane probability of the vector entropy sample, and q represents the invariable lane probability of the vector entropy sample. The calculation of the class entropy of the subsequent vector entropy sample and the class entropy of the RSSI sample is similar, and will not be described here.

[0063] In the embodiments of the present application, the class entropy is determined according to the confidence of the normalized data frame (such as the radar data frame), and the confidence is used to describe the reliability of the data frame. Therefore, the class entropy determined by the above method can improve the accuracy of the variable lane prediction. In addition, since normalization can ensure that the class entropy and the information entropy are kept in the same order of magnitude, the class entropy retains the characteristics of entropy (i.e., the smaller the more stable).

[0064] Assuming that the motion vector includes the tangential velocity and the displacement, the class entropy corresponding to different attributes can be as shown in Table 2.

[0065] Table 2:

[0066]

[0067] A3, determining the class entropy of the vector entropy sample according to the variable lane probability and the invariable lane probability of the vector entropy sample.

[0068] A4, determining the class entropy of the RSSI sample according to the variable lane probability and the invariable lane probability of the RSSI sample.

[0069] A5, determining the target attribute as the root node and other attributes as the child nodes according to the sizes of the class entropy of the motion vector sample, the class entropy of the vector entropy sample, and the class entropy of the RSSI sample, wherein the target attribute belongs to a preset attribute set, the preset attribute set includes the motion vector attribute, the vector entropy attribute, and the RSSI attribute, the target attribute is the attribute corresponding to the smallest class entropy, and the other attributes are the attributes in the preset attribute set other than the target attribute.

[0070] For example, assuming that the class entropy of the motion vector sample is the smallest, the motion vector is taken as the root node, and the vector entropy and the RSSI are taken as the child nodes.

[0071] A6, generating a decision tree model to be trained according to the target attribute as the root node and the other attributes as the child nodes.

[0072] In the steps A1-A6, since the class entropy can measure the purity of the sample set, the attribute with the highest class entropy (i.e., the attribute corresponding to the smallest class entropy) is taken as the root node of the decision tree model, and the attributes with other numerical class entropies are taken as the child nodes, so that the importance of the attribute with the highest purity can be highlighted.

[0073] In some embodiments, the step A5 includes:

[0074] A51, determining the target attribute as the root node according to the class entropy of the motion vector sample, the class entropy of the vector entropy sample, and the class entropy of the RSSI sample.

[0075] A52, calculating the lane change probability and the lane non-change probability of other attributes under the condition of the target attribute.

[0076] In the present embodiment, considering that the target vehicle does not need to be judged whether to change lanes according to other attributes when the target attribute is determined to change lanes, the lane change probability and the lane non-change probability of other attributes can be directly calculated under the condition that the target attribute does not change lanes.

[0077] The calculation formula of the conditional probability is as follows: That is, the probability of event B occurring under the condition that the probability of event A is P(A). In the present embodiment, event A is the event that the target attribute does not change lanes, and event B is the lane change event or the lane non-change event of other attributes. For example, assuming that the target attribute is the cutting speed and the other attributes include displacement, the probability of displacement changing lanes under the condition that the cutting speed does not change lanes needs to be calculated, and the probability of displacement not changing lanes under the condition that the cutting speed does not change lanes needs to be calculated.

[0078] A53, determining the class entropy corresponding to the other attributes according to the lane change probability and the lane non-change probability of the other attributes.

[0079] A54, determining the attribute corresponding to the smallest class entropy in the class entropy corresponding to the other attributes as the attribute of the next node, and if the preset attribute set still has attributes that are not taken as nodes, setting the attribute of the next node as a new target attribute, and returning to the step of calculating the lane change probability and the lane non-change probability of other attributes under the condition of the target attribute and the subsequent steps until all attributes in the preset attribute set are taken as nodes.

[0080] The process of generating the decision tree model will be described below with a specific example.

[0081] Assuming that the preset attribute set includes five attributes: cutting speed, displacement, vector entropy of cutting speed, vector entropy of displacement, and RSSI.

[0082] Assuming the target attribute is tangential velocity, and tangential velocity is taken as the root node, the entropy class of displacement, the entropy class of tangential velocity vector entropy, the entropy class of displacement vector entropy, and the entropy class of RSSI are calculated subsequently, assuming no lane change occurs at tangential velocity. If the entropy class of displacement is determined to be the minimum, then displacement is taken as the next node after the root node.

[0083] Assuming that there is no change in tangential velocity and no change in displacement, among the vector entropy of tangential velocity, the vector entropy of displacement, and the entropy of RSSI, the entropy of RSSI is the smallest. Therefore, RSSI is the next node after the next node of the root node.

[0084] At this point, since only the vector entropy of tangential velocity and displacement remains in the preset attribute set, it is necessary to continue calculating the class entropy until all five attributes in the preset attribute set are used as nodes in the decision tree model. Assuming that in subsequent calculations the class entropy of tangential velocity is less than the class entropy of displacement, the resulting decision tree model is as follows. Figure 3 As shown.

[0085] When making predictions based on this decision tree model, predictions are made first based on the parent node and then based on the child nodes. In the decision tree model generated by steps A51 to A54, the class entropy purity of the attribute corresponding to each child node is lower than that of its corresponding parent node. Moreover, the higher the class entropy purity, the lower the uncertainty. Therefore, the decision tree constructed in this way can make predictions first from the attributes with low uncertainty and then from the attributes with high uncertainty, which is conducive to obtaining more accurate prediction results.

[0086] In some embodiments, step S12 includes:

[0087] B1. If the above motion vector includes tangential velocity, then calculate the vector entropy of the tangential velocity based on the lane change probability and lane non-change probability of the tangential velocity.

[0088] Specifically, the probability p of the aforementioned shear velocity being greater than a preset velocity threshold is determined. v1 The above p v1 The probability of lane change is defined as the aforementioned shear speed, and the probability p of determining that the aforementioned shear speed is not greater than the aforementioned preset speed threshold is also defined. v2 The above p v2 As the invariant path probability of the aforementioned tangent velocity, according to the aforementioned p v1 and p v2 Calculate the vector entropy of the aforementioned tangential velocity.

[0089] Wherein, the vector entropy S of the tangential velocity Vy S can be determined using the following formula: Vy =-(p v1*log2p v1 +p v2 *log2p v2 In the formula, p v1 It is the probability, p, that the cutting speed exceeds a preset speed threshold (let's say V_threshold) in a series of consecutive frames of data. v2 This represents the probability that the cut-off velocity is not greater than V_shreshold. For example, after acquiring a radar data frame at the current moment, we can calculate the probabilities of cut-off velocities greater than V_shreshold and the probabilities of cut-off velocities not greater than V_shreshold in the radar data frame acquired at the current moment and in multiple historical radar data frames prior to the current moment.

[0090] B2. If the above motion vector includes displacement, then calculate the vector entropy of the above displacement based on the probability of changing lanes and the probability of not changing lanes.

[0091] Specifically, the probability p of the aforementioned displacement being greater than a preset distance threshold is determined. s1 And, the probability p of determining that the above displacement is not greater than the above preset distance threshold. s2 The above p s1 As the probability of lane change for displacement, the above p s2 As the invariant path probability of the displacement, and based on the above p s1 and p s2 Calculate the vector entropy of the above displacements.

[0092] The above displacement entropy S x =-(p s1 *log2p s1 +p s1 *log2p s1 ), where p s1 It is the probability, p, that the displacement exceeds a preset distance threshold (let's say X_threshold) in multiple consecutive radar data frames. s1 It is the probability that the displacement is less than X_shreshold.

[0093] In some embodiments, in order to obtain information about target vehicles that may change lanes in advance, after step S14, the method further includes:

[0094] Output a list of vehicles to be changed, which includes the license plate of the target vehicle, the current lane number of the target vehicle, the lane number to which the target vehicle is to go, and its location information.

[0095] In this embodiment, for each radar data frame, if there is a target vehicle that may change lane in the radar data frame, a pre-lane change list is generated according to the target vehicle. Of course, the pre-lane change list can also include the time of obtaining the lane number (or other information) that the target vehicle intends to go to, which will not be described here.

[0096] Specifically, when the road is monitored by the radar, corresponding radar data will be obtained, which includes target ID, target position information (i.e. position information in a Cartesian coordinate system with the radar as the origin, the lane as the x-axis, and the lane as the y-axis), target speed (such as the tangential speed of the target vehicle), and can also include signal-to-noise ratio (SNR) and the like. At the same time, after the RSU communicates with the OBU, the RSU will obtain corresponding RSU data, which includes vehicle type, license plate, RSSI (or RSSI position information), and the like. After obtaining the radar data and the RSU data, the radar data and the RSU data are aligned, and the license plate and the position information belonging to the pre-lane change list can be obtained, and the current lane number and the lane number that the target vehicle intends to go to can be obtained by calculation.

[0097] In this embodiment, after the position information of the target vehicle is determined, the current lane number of the target vehicle can be determined according to the following method:

[0098]

[0099] The first lane edge line position information and the lane width can be pre-set, such as the lane width can be pre-set to 3.75 meters. Of course, the first lane edge line position information and the lane width can also be determined by recognizing the lane line, which is not limited here.

[0100] The obtained pre-lane change vehicle relative lane position information is matched in intervals, and different intervals correspond to different lane numbers, which will be described in detail in the following table 3.

[0101] Table 3:

[0102] Relative lane position information (0，1.0] (1.0，2.0] (2.0，3.0] (3.0，4.0] Lane number 1 2 3 4

[0103] In this embodiment, after the tangential speed of the target vehicle is determined, the lane number that the target vehicle intends to go to can be determined according to the direction of the tangential speed and the current lane number of the target vehicle.

[0104] In some embodiments, after the pre-lane change list is generated, the target vehicle in the pre-lane change list can be tracked to determine whether the target vehicle has completed lane changing, i.e. the vehicle lane change prediction method of the present application also includes:

[0105] C1, determining a target license plate and a current lane number corresponding to the target license plate from a first pre-lane-changing list, the target license plate being a license plate corresponding to a target vehicle which needs to be determined whether the lane changing is completed, and the first pre-lane-changing list being one of the pre-lane-changing lists.

[0106] In the embodiment, it is assumed that the motion vector of the target vehicle is obtained from the radar data frame, and one of the pre-lane-changing lists is generated for each radar data frame.

[0107] C2, if the target license plate exists in the second pre-lane-changing list, obtaining the current lane number of the target license plate in the second pre-lane-changing list, if the current lane number of the target license plate obtained from the first pre-lane-changing list is the same as the current lane number of the target license plate obtained from the second pre-lane-changing list, it is determined that the vehicle corresponding to the target license plate has not completed the lane changing, if the current lane number of the target license plate obtained from the first pre-lane-changing list is not the same as the current lane number of the target license plate obtained from the second pre-lane-changing list, it is determined that the vehicle corresponding to the target license plate has completed the lane changing, or if the current lane number of the target license plate obtained from the first pre-lane-changing list is not the same as the current lane number of the target license plate obtained from the second pre-lane-changing list, and the current lane number of the target license plate obtained from the second pre-lane-changing list is the same as the target lane number of the target license plate obtained from the first pre-lane-changing list, it is determined that the vehicle corresponding to the target license plate has completed the lane changing, wherein the second pre-lane-changing list is one of the pre-lane-changing lists adjacent to and after the first pre-lane-changing list.

[0108] Since each pre-lane-changing list includes the license plate of the target vehicle, the current lane number of the target vehicle, the target lane number and the position information, it can be determined whether the target vehicle has completed the lane changing by comparing the lane numbers in the adjacent pre-lane-changing lists.

[0109] For example, if it is needed to determine whether the vehicle corresponding to the license plate (supposedly license plate 1) in the first pre-lane-changing list has completed lane changing, the pre-lane-changing list generated after the first pre-lane-changing list (supposedly the second pre-lane-changing list) is obtained first, and then it is determined from the second pre-lane-changing list whether the license plate 1 exists. If the license plate 1 exists, the current lane number of the license plate 1 obtained from the second pre-lane-changing list is obtained, and the current lane number of the license plate 1 obtained from the second pre-lane-changing list is compared with the current lane number of the license plate 1 obtained from the first pre-lane-changing list. If the two current lane numbers are the same, it is determined that the lane changing has not been completed. Then, the pre-lane-changing list generated after the second pre-lane-changing list (supposedly the third pre-lane-changing list) is continuously obtained, and the current lane number of the license plate 1 obtained from the third pre-lane-changing list is compared with the current lane number of the license plate 1 obtained from the second pre-lane-changing list (or the current lane number of the license plate 1 obtained from the first pre-lane-changing list). The specific comparison process is similar to the above comparison process, which will not be described here.

[0110] In the embodiment, if the current lane number of the license plate 1 obtained from the second pre-lane-changing list is different from the current lane number of the license plate 1 obtained from the first pre-lane-changing list, it is determined that the vehicle corresponding to the license plate 1 has completed lane changing. Further, considering that there may be some errors in the position information corresponding to the radar data, that is, there may be errors in the current lane number calculated according to the position information of the target vehicle. At this time, after it is determined that the current lane number of the license plate 1 obtained from the second pre-lane-changing list is different from the current lane number of the license plate 1 obtained from the first pre-lane-changing list, it is further determined whether the current lane number of the license plate 1 obtained from the second pre-lane-changing list is the same as the intended lane number of the license plate 1 obtained from the first pre-lane-changing list. If the current lane number and the intended lane number are the same, it is determined that the vehicle corresponding to the license plate 1 has completed lane changing. Since double determination is performed, the accuracy of the determination result of whether the vehicle has completed lane changing is improved.

[0111] In some embodiments, for the target vehicle determined to have completed lane changing in the same pre-lane-changing list, a corresponding lane-changing completion list is generated. The lane-changing completion list includes the license plate, the current lane, the previous lane, the position information, etc. Of course, the lane-changing completion list can also include SNR, time, etc. information, which is not limited here.

[0112] In some embodiments, after it is determined that the target vehicle has completed lane changing, it can be analyzed whether the lane changing behavior is abnormal to improve the safety of vehicle driving, that is, the vehicle lane changing prediction method of the embodiment of the application further includes:

[0113] D1, determining the lane changing time and the lane changing offset angle of the target vehicle determined to have completed lane changing.

[0114] Specifically, after determining that the target vehicle has completed its lane change, the time and location information of the target vehicle's lane change completion are recorded (let's say s1). Then, the time (e.g., obtained from the pre-lane change list) and location information of the lane number that the target vehicle intends to go to are obtained (let's say s2).

[0115] Subtracting the two time points gives the lane change time of the target vehicle, and the lane change offset angle can be obtained from the two positional information. For example, assuming the coordinates of s1 are (x1, y1) and the coordinates of s2 are (x2, y2), then the lane change offset angle = arctan[|s1-s2| / |x1-x2|].

[0116] D2. Determine whether the lane change behavior is abnormal based on the lane change time and lane change offset angle mentioned above.

[0117] Specifically, lane change behaviors that do not meet the requirements in terms of lane change time and lane change offset angle are judged as abnormal lane change behaviors.

[0118] like Figure 4 As shown, after determining the lane change time, the lane change time is compared with a preset time threshold t. If the lane change time is greater than t, the confidence level is increased by 1; otherwise, the confidence level is decreased by 1.

[0119] Then, the determined lane change offset angle is compared with the preset angle threshold θ. If it is greater than θ, the confidence level is increased by 1; otherwise, the confidence level is decreased by 1.

[0120] Determine if the confidence level is equal to 2. If yes, the lane change behavior is considered normal; otherwise, the lane change behavior is considered abnormal (i.e., lane change behavior with too short a lane change time and too large a lane change angle).

[0121] In some embodiments, a list of vehicles exhibiting abnormal lane-changing behavior is generated and output for subsequent querying. This list includes the license plate, the vehicle's position before the lane change, and its position after the lane change. Of course, the list may also include information such as the target ID, which will not be elaborated upon here.

[0122] In some embodiments, warning information is generated and output based on information from the pre-lane change list, the lane change completion list, and / or the rapid lane change list; or, driving habits are generated and output. In this way, users who obtain warning information and / or driving habits can be informed of relevant lane change information in a timely manner, which is conducive to improving vehicle driving safety.

[0123] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0124] Example Two

[0125] The vehicle lane change prediction method corresponding to the above embodiment, Figure 5 The structure block diagram of the vehicle lane change prediction device provided by the embodiment of the application is shown, and only the part related to the embodiment of the application is shown for the convenience of description.

[0126] Referring to Figure 5 The vehicle lane change prediction device 5 comprises a motion vector acquisition module 51, a vector entropy calculation module 52, an RSSI acquisition module 53, and a prediction result output module 54. Wherein:

[0127] The motion vector acquisition module 51 is configured to acquire the motion vector of the target vehicle, wherein the motion vector comprises the tangential velocity and / or the displacement.

[0128] The vector entropy calculation module 52 is configured to calculate the vector entropy of the motion vector, wherein the vector entropy is used to describe the stability of the motion trend of the target vehicle.

[0129] The RSSI acquisition module 53 is configured to acquire the RSSI of the target vehicle.

[0130] The prediction result output module 54 is configured to input the motion vector, the vector entropy and the RSSI into a trained decision tree model to obtain the prediction result output by the trained decision tree model, wherein the prediction result is used to indicate whether the target vehicle changes lane, the trained decision tree model is trained according to a set of data corresponding to a group of vehicles with known lane change, and the data set comprises motion vector samples, vector entropy samples and RSSI samples of the vehicles with known lane change.

[0131] In the embodiment of the application, since the RSSI can reflect the distance between the measurement signal point (i.e. the target vehicle) and the receiving point, the RSSI of the target vehicle can reflect the relative position of the target vehicle and the measurement signal point, i.e. the position of the target vehicle. At the same time, since the motion vector of the target vehicle such as the tangential velocity and / or the displacement can reflect the motion trend of the target vehicle, the vector entropy corresponding to the motion vector can reflect the stability of the motion trend of the target vehicle, and when the vehicle changes lane, its motion trend will inevitably be different from the original motion trend, and the motion trend will continue to be stable for a period of time. Therefore, after inputting the RSSI, the motion vector and the corresponding vector entropy of the target vehicle into the trained decision tree model, an accurate prediction result can be obtained, i.e. the target vehicle can be accurately predicted whether to change lane by the above method.

[0132] It should be noted that the vehicle lane change prediction device of the embodiment of the application can be applied in a radar, can also be applied in an RSU, and can also be applied in a roadside device.

[0133] In some embodiments, the vehicle lane change prediction device 5 provided by the embodiments of the present application further comprises:

[0134] Two probability calculation modules are configured to calculate the lane change probability and the no-lane change probability of the motion vector sample, the vector entropy sample and the RSSI sample respectively.

[0135] A motion vector sample entropy class determination module is configured to determine the entropy class of the motion vector sample according to the lane change probability and the no-lane change probability of the motion vector sample.

[0136] A vector entropy sample entropy class determination module is configured to determine the entropy class of the vector entropy sample according to the lane change probability and the no-lane change probability of the vector entropy sample.

[0137] An RSSI sample entropy class determination module is configured to determine the entropy class of the RSSI sample according to the lane change probability and the no-lane change probability of the RSSI sample.

[0138] A node attribute determination module is configured to determine a target attribute as a root node and other attributes as child nodes according to the sizes of the entropy classes of the motion vector sample, the vector entropy sample and the RSSI sample, wherein the target attribute belongs to a preset attribute set, the preset attribute set comprises a motion vector attribute, a vector entropy attribute and an RSSI attribute, the target attribute is an attribute corresponding to the smallest entropy class, and the other attributes are attributes in the preset attribute set other than the target attribute.

[0139] A decision tree model generation module is configured to generate a decision tree model to be trained according to the target attribute as the root node and the other attributes as the child nodes.

[0140] In some embodiments, the node attribute determination module comprises:

[0141] A root node attribute determination unit is configured to determine the target attribute as the root node according to the sizes of the entropy classes of the motion vector sample, the vector entropy sample and the RSSI sample.

[0142] A conditional probability calculation unit is configured to calculate the lane change probability and the no-lane change probability of the other attributes with the target attribute as a condition.

[0143] A conditional entropy class calculation unit is configured to determine the entropy class corresponding to the other attributes according to the lane change probability and the no-lane change probability of the other attributes.

[0144] The child node attribute determining unit is configured to determine the attribute corresponding to the minimum class entropy in the class entropy corresponding to other attributes as the attribute of the next node, and if the preset attribute set still has attributes that are not used as the attributes of nodes, set the attribute of the next node as a new target attribute, and return to the steps of calculating the lane-changing probability and the non-lane-changing probability of other attributes and subsequent steps based on the target attribute until all attributes in the preset attribute set are used as nodes.

[0145] In some embodiments, the vector entropy calculation module 52 includes:

[0146] The vector entropy calculation unit of the cutting speed is configured to calculate the vector entropy of the cutting speed based on the lane-changing probability and the non-lane-changing probability of the cutting speed if the motion vector includes the cutting speed.

[0147] The vector entropy calculation unit of the displacement is configured to calculate the vector entropy of the displacement based on the lane-changing probability and the non-lane-changing probability of the displacement if the motion vector includes the displacement.

[0148] In some embodiments, the vehicle lane-changing prediction device 5 provided by the embodiments of the present application further includes:

[0149] The pre-lane-changing list output module is configured to output a pre-lane-changing list, and the pre-lane-changing list includes the license plate of the target vehicle, the current lane number of the target vehicle, the lane number to be traveled to, and the position information.

[0150] In some embodiments, the motion vector of the target vehicle is obtained from a radar data frame, and a pre-lane-changing list is generated for each radar data frame. In this case, the vehicle lane-changing prediction device 5 provided by the embodiments of the present application further includes:

[0151] The target license plate determining module is configured to determine the target license plate and the current lane number corresponding to the target license plate from the first pre-lane-changing list, the target license plate is the license plate corresponding to the target vehicle to be judged for whether the lane-changing is completed, and the first pre-lane-changing list is one of the pre-lane-changing lists.

[0152] The target vehicle lane change completion determination module is used to: if the target license plate exists in the second pre-lane change list, obtain the current lane number of the target license plate in the second pre-lane change list; if the current lane number of the target license plate obtained from the first pre-lane change list is the same as the current lane number of the target license plate obtained from the second pre-lane change list, then determine that the vehicle corresponding to the target license plate has not completed the lane change; if the current lane number of the target license plate obtained from the first pre-lane change list is different from the current lane number of the target license plate obtained from the second pre-lane change list, then determine... The vehicle corresponding to the aforementioned target license plate has completed its lane change. Alternatively, if the current lane number of the aforementioned target license plate obtained from the first pre-lane change list is different from the current lane number of the aforementioned target license plate obtained from the second pre-lane change list, and the current lane number of the aforementioned target license plate obtained from the second pre-lane change list is the same as the lane number the aforementioned target license plate intends to go to obtained from the first pre-lane change list, then it is determined that the vehicle corresponding to the aforementioned target license plate has completed its lane change. The second pre-lane change list is a pre-lane change list that is adjacent to the first pre-lane change list and follows the first pre-lane change list.

[0153] The vehicle lane change prediction device 5 provided in this application embodiment also includes:

[0154] The lane change time determination module is used to determine the lane change time and lane change offset angle of the target vehicle after it completes the lane change.

[0155] The lane change behavior abnormality judgment module is used to determine whether the lane change behavior is abnormal based on the lane change time and lane change offset angle mentioned above.

[0156] It should be noted that the information interaction and execution process between the above-mentioned devices / units are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0157] Example Three

[0158] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 6 As shown, the electronic device 6 of this embodiment includes: at least one processor 60 ( Figure 6 The diagram shows only one processor, a memory 61, and a computer program 62 stored in the memory 61 and executable on at least one processor 60. When the processor 60 executes the computer program 62, it implements the steps in any of the above method embodiments.

[0159] The electronic device 6 can be a computing device such as a radar, a road side unit, or a road side device. The electronic device can include, but is not limited to, a processor 60, a memory 61. Those skilled in the art can understand that Figure 6 The electronic device 6 is merely an example and is not intended to limit the electronic device 6, which can include more or less components than shown, or combine some components, or different components, such as an input / output device, a network access device, and the like.

[0160] The processor 60 can be a central processing unit (CPU), and can also be other general-purpose processors, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or the like. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0161] The memory 61 can be an internal storage unit of the electronic device 6, such as a hard disk or a memory of the electronic device 6 in some embodiments. The memory 61 can also be an external storage device of the electronic device 6, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, or the like, in other embodiments. Further, the memory 61 can include both an internal storage unit and an external storage device of the electronic device 6. The memory 61 is used to store an operating system, an application program, a boot loader, data, and other programs, such as program codes of the computer program, and the like. The memory 61 can also be used to temporarily store data that has been output or will be output.

[0162] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional units and modules is exemplified, and in actual application, the above functions can be completed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit or module in the embodiment can be integrated in one processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit, and the integrated unit can be realized in the form of hardware or in the form of software functional unit. In addition, the specific names of each functional unit or module are only for convenient distinction, and do not limit the protection scope of the present application. The specific working process of the units and modules in the system can refer to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0163] The embodiments of the present application further provide a network device, comprising at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the processor implements the steps in any of the method embodiments described above when executing the computer program.

[0164] The embodiments of the present application further provide a computer readable storage medium, which stores a computer program, wherein the computer program is executable by a processor to implement the steps in any of the method embodiments described above.

[0165] The embodiments of the present application provide a computer program product, which, when running on an electronic device, enables the electronic device to implement the steps in any of the method embodiments described above.

[0166] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the present application can implement all or part of the processes in the above-mentioned embodiment methods through a computer program to instruct relevant hardware to complete, and the computer program can be stored in a computer readable storage medium. When the computer program is executed by a processor, the steps of each method embodiment described above can be implemented. The computer program includes computer program code, which can be in the form of source code, object code, executable files or some intermediate forms. The computer readable medium at least includes any entity or device capable of carrying the computer program code to the photographing device / electronic device, recording medium, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium. For example, U disk, mobile hard disk, magnetic disk or optical disk, etc. In some jurisdictions, according to legislation and patent practice, the computer readable medium can not be an electrical carrier signal and a telecommunication signal.

[0167] In the above embodiments, the description of each embodiment has its own focus, and the parts not described or recorded in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0168] Those skilled in the art can appreciate that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0169] In the embodiments provided by the present application, it should be understood that the disclosed apparatus / network device and method can be implemented in other ways. For example, the above-described apparatus / network device embodiments are merely schematic, for example, the division of the modules or units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed each other can be indirect coupling or communication connection through some interfaces, devices or units, and can be electrical, mechanical or other forms.

[0170] The units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, that is, may be located in one place, or may also be distributed to multiple network units. Part or all of the units can be selected to achieve the purpose of the embodiment scheme according to actual needs.

[0171] The above embodiments are only used to illustrate the technical solutions of the present application, but not limit them. Although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those skilled in the art that the technical solutions recorded in the foregoing embodiments can still be modified, or some technical features can be replaced by equivalents. The modification or replacement does not change the essence of the corresponding technical solution, and should be included in the protection scope of the present application.

Claims

1. A vehicle lane change prediction method, characterized by, The method comprises the following steps: acquiring a motion vector of a target vehicle, the motion vector comprising a tangential velocity and / or a displacement; calculating a vector entropy of the motion vector, the vector entropy being used to describe a stability degree of a motion trend of the target vehicle; acquiring an RSSI of the target vehicle, the RSSI reflecting a distance between the target vehicle and a receiving point; inputting the motion vector, the vector entropy and the RSSI into a trained decision tree model to obtain a prediction result output by the trained decision tree model, the prediction result being used to indicate whether the target vehicle changes lane, wherein the trained decision tree model is trained according to a data set corresponding to a group of vehicles with known lane changing or not, and the data set comprises motion vector samples, vector entropy samples and RSSI samples of the vehicles with known lane changing or not.

2. The vehicle lane change prediction method of claim 1, wherein The method further comprises the following steps: respectively calculating a lane changing probability and a non-lane changing probability of the motion vector samples, the vector entropy samples and the RSSI samples; determining a class entropy of the motion vector samples according to the lane changing probability and the non-lane changing probability of the motion vector samples, wherein the class entropy is a value obtained by modifying an information entropy; determining a class entropy of the vector entropy samples according to the lane changing probability and the non-lane changing probability of the vector entropy samples; determining a class entropy of the RSSI samples according to the lane changing probability and the non-lane changing probability of the RSSI samples; determining a target attribute as a root node and other attributes as child nodes according to sizes of the class entropy of the motion vector samples, the class entropy of the vector entropy samples and the class entropy of the RSSI samples, wherein the target attribute belongs to a preset attribute set, the preset attribute set comprising a motion vector attribute, a vector entropy attribute and an RSSI attribute, the target attribute is an attribute corresponding to a minimum class entropy, and the other attributes are attributes in the preset attribute set other than the target attribute; generating a decision tree model to be trained according to the target attribute as the root node and the other attributes as the child nodes.

3. The vehicle lane change prediction method of claim 2, wherein The step of determining the target attribute as the root node and the other attributes as the child nodes according to the sizes of the class entropy of the motion vector samples, the class entropy of the vector entropy samples and the class entropy of the RSSI samples comprises the following steps: determining the target attribute as the root node according to the sizes of the class entropy of the motion vector samples, the class entropy of the vector entropy samples and the class entropy of the RSSI samples; calculating a lane changing probability and a non-lane changing probability of the other attributes under the condition of the target attribute; determining a class entropy corresponding to the other attributes according to the lane changing probability and the non-lane changing probability of the other attributes; determining an attribute corresponding to a minimum class entropy in the class entropy corresponding to the other attributes as an attribute of a next node, setting the attribute of the next node as a new target attribute if there are still attributes in the preset attribute set that are not attributes of nodes, and returning to the step of calculating the lane changing probability and the non-lane changing probability of the other attributes under the condition of the target attribute and subsequent steps until all the attributes in the preset attribute set are attributes of nodes.

4. The vehicle lane change prediction method of claim 1, wherein, The step of calculating the vector entropy of the motion vector comprises the following steps: if the motion vector comprises a cut speed, calculating a vector entropy of the cut speed according to a lane-changing probability and a non-lane-changing probability of the cut speed; if the motion vector comprises a displacement, calculating a vector entropy of the displacement according to a lane-changing probability and a non-lane-changing probability of the displacement.

5. The vehicle lane change prediction method according to any one of claims 1 to 4, characterized in that, if the prediction result indicates that the target vehicle changes lane, further comprising: outputting a pre-lane-changing list, the pre-lane-changing list comprising a license plate of the target vehicle, a current lane number of the target vehicle, a lane number to be arrived at, and position information; 6. The vehicle lane change prediction method of claim 5, wherein acquiring a motion vector of a target vehicle from a radar data frame, and generating a pre-lane-changing list for each radar data frame, the vehicle lane-changing prediction method further comprising: determining a target license plate and a current lane number corresponding to the target license plate from a first pre-lane-changing list, the target license plate being a license plate corresponding to a target vehicle to be judged for whether lane changing is completed, and the first pre-lane-changing list being one of the pre-lane-changing lists; if the target license plate exists in a second pre-lane-changing list, acquiring a current lane number of the target license plate in the second pre-lane-changing list, if the current lane number of the target license plate acquired from the first pre-lane-changing list is the same as the current lane number of the target license plate acquired from the second pre-lane-changing list, determining that the vehicle corresponding to the target license plate has not completed lane changing, if the current lane number of the target license plate acquired from the first pre-lane-changing list is not the same as the current lane number of the target license plate acquired from the second pre-lane-changing list, determining that the vehicle corresponding to the target license plate has completed lane changing, or, if the current lane number of the target license plate acquired from the first pre-lane-changing list is not the same as the current lane number of the target license plate acquired from the second pre-lane-changing list, and the current lane number of the target license plate acquired from the second pre-lane-changing list is the same as the lane number to be arrived at of the target license plate acquired from the first pre-lane-changing list, determining that the vehicle corresponding to the target license plate has completed lane changing, wherein the second pre-lane-changing list is one pre-lane-changing list adjacent to and after the first pre-lane-changing list.

7. The vehicle lane change prediction method of claim 6, wherein further comprising: determining a lane-changing time and a lane-changing offset angle at which the target vehicle completes lane changing; judging whether a lane-changing behavior is abnormal according to the lane-changing time and the lane-changing offset angle.

8. A vehicle lane change prediction device characterized by comprising: comprising: a motion vector acquisition module, configured to acquire a motion vector of a target vehicle, the motion vector comprising a cut speed and / or a displacement; a vector entropy calculation module, configured to calculate a vector entropy of the motion vector, the vector entropy being used to describe a stability degree of a motion trend of the target vehicle; an RSSI acquisition module, configured to acquire an RSSI of the target vehicle, the RSSI reflecting a distance between the target vehicle and a receiving point; A prediction result output module is configured to input the motion vector, the vector entropy, and the RSSI into a trained decision tree model to obtain a prediction result output by the trained decision tree model, the prediction result being used to indicate whether the target vehicle changes lanes.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The computer program is executed by the processor to implement the method in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 9. The computer program is executed by the processor to implement the method in any one of claims 1 to 7.

Images