Vehicle speed prediction method, device, equipment and storage medium

Abstract

The present disclosure provides a vehicle speed prediction method, device, equipment, and storage medium, belonging to the technical field of vehicle driving condition prediction. The method includes: obtaining a target driving data set during vehicle driving, the target driving data set including multiple driving data groups collected within a first time period, each driving data group including the collection time, the acceleration and speed of the vehicle; using the target driving data set, updating the first vehicle speed prediction model used in the first time period to obtain a second vehicle speed prediction model, the first vehicle speed prediction model and the second vehicle speed prediction model being Markov vehicle speed prediction models; using the second vehicle speed prediction model, predicting the vehicle speed in a second time period, the second time period being a time period after the first time period and adjacent to the first time period. This method can improve the accuracy of vehicle speed prediction.

Description

Technical Field

[0001] The present disclosure relates to the technical field of vehicle driving condition prediction, and in particular to a vehicle speed prediction method, device, equipment and storage medium. Background Art

[0002] Vehicle speed represents a core motion state of the vehicle, influencing numerous functions, including torque request, gear control, and stability control. Accurate speed prediction helps optimize power distribution in hybrid vehicles, thereby improving fuel economy.

[0003] In related technologies, vehicle speed prediction methods include: using historical data for offline training to obtain a vehicle speed prediction model, and using the vehicle speed prediction model to predict the vehicle speed during driving. The historical data generally includes driving data corresponding to standard driving conditions.

[0004] Due to the difference between historical data and the driving data during vehicle driving, the accuracy of the vehicle speed predicted by this method is low. Summary of the Invention

[0005] The embodiments of the present disclosure provide a vehicle speed prediction method, apparatus, device, and storage medium that can improve the accuracy of vehicle speed prediction. The technical solution is as follows:

[0006] In a first aspect, a vehicle speed prediction method is provided, the method comprising: obtaining a target driving data set during vehicle driving, the target driving data set comprising a plurality of driving data groups collected within a first time period, each driving data group comprising a collection time, an acceleration of the vehicle, and a speed of the vehicle; using the target driving data set, updating a first vehicle speed prediction model used within the first time period to obtain a second vehicle speed prediction model, the first vehicle speed prediction model and the second vehicle speed prediction model being both Markov vehicle speed prediction models, the Markov vehicle speed prediction model being used to describe the probability distribution of changing from a first acceleration and a first speed state to a second acceleration at each prediction moment; using the second vehicle speed prediction model, predicting the vehicle speed within a second time period, the second time period being a time period after the first time period and adjacent to the first time period.

[0007] Optionally, the use of the target driving data set to update the first vehicle speed prediction model used in the first time period includes: determining a state transition frequency matrix corresponding to the target driving data set based on the target driving data set; and updating the first vehicle speed prediction model based on the state transition frequency matrix to obtain the second vehicle speed prediction model.

[0008] Optionally, the first vehicle speed prediction model includes a first state transition frequency matrix; the first vehicle speed prediction model is updated according to the state transition frequency matrix to obtain the second vehicle speed prediction model, including: calculating the target state transition frequency matrix according to the set forgetting factor, the first state transition frequency matrix and the state transition frequency matrix; determining the target state transition frequency matrix according to the target state transition frequency matrix to obtain the second vehicle speed prediction model.

[0009] Optionally, calculating the target state transition frequency matrix based on the set forgetting factor, the first state transition frequency matrix and the state transition frequency matrix includes: adding the product of the set forgetting factor and the first state transition frequency matrix to the state transition frequency matrix to obtain the target state transition frequency matrix.

[0010] Optionally, the method also includes: based on the debugging data set, using multiple candidate forgetting factors to update the initial vehicle speed prediction model respectively to obtain multiple test vehicle speed prediction models, the initial vehicle speed prediction model is a vehicle speed prediction model trained using driving data corresponding to the standard driving condition; determining the vehicle speed prediction errors corresponding to the multiple test vehicle speed prediction models, the vehicle speed prediction error is the error between the vehicle speed predicted using the corresponding test vehicle speed prediction model and the actual vehicle speed; and determining the candidate forgetting factor corresponding to the test vehicle speed prediction model when the vehicle speed prediction error is less than the error threshold as the set forgetting factor.

[0011] In a second aspect, a vehicle speed prediction device is provided, which includes: an acquisition module for obtaining a target driving data set during vehicle driving, the target driving data set including multiple driving data groups collected within a first time period, each driving data group including a collection time, an acceleration of the vehicle and a speed of the vehicle; an update module for using the target driving data set to update a first vehicle speed prediction model used in the first time period to obtain a second vehicle speed prediction model, the first vehicle speed prediction model and the second vehicle speed prediction model are both Markov vehicle speed prediction models, and the Markov vehicle speed prediction model is used to describe the probability distribution of changing from a first acceleration and a first speed state to a second acceleration at each prediction moment; a vehicle speed prediction module for using the second vehicle speed prediction model to predict the vehicle speed in a second time period, the second time period being a time period after the first time period and adjacent to the first time period.

[0012] Optionally, the updating module is used to determine a state transition frequency matrix corresponding to the target driving data set based on the target driving data set; and update the first vehicle speed prediction model based on the state transition frequency matrix to obtain the second vehicle speed prediction model.

[0013] Optionally, the first vehicle speed prediction model includes a first state transition frequency matrix; the update module is used to calculate a target state transition frequency matrix based on a set forgetting factor, the first state transition frequency matrix and the state transition frequency matrix; and determine a target state transition frequency matrix based on the target state transition frequency matrix to obtain the second vehicle speed prediction model.

[0014] Optionally, the updating module is configured to add the product of the set forgetting factor and the first state transition frequency matrix to the state transition frequency matrix to obtain the target state transition frequency matrix.

[0015] Optionally, the device also includes a determination module, which is used to update the initial vehicle speed prediction model based on the debugging data set using multiple candidate forgetting factors to obtain multiple test vehicle speed prediction models, where the initial vehicle speed prediction model is a vehicle speed prediction model trained using driving data corresponding to a standard driving condition; determine the vehicle speed prediction errors corresponding to the multiple test vehicle speed prediction models, where the vehicle speed prediction error is the error between the vehicle speed predicted by the corresponding test vehicle speed prediction model and the actual vehicle speed; and determine the candidate forgetting factor corresponding to the test vehicle speed prediction model when the vehicle speed prediction error is less than an error threshold as the set forgetting factor.

[0016] According to a third aspect, a computer device is provided, comprising: a processor; and a memory for storing processor-executable instructions; wherein the processor is configured to execute the method described in the first aspect.

[0017] In a fourth aspect, a computer-readable medium is provided. When instructions in the computer-readable medium are executed by a processor of a computer device, the computer device is enabled to execute the method described in the first aspect.

[0018] In a fifth aspect, a computer program product is provided, comprising a computer program / instruction, wherein the computer program / instruction implements the method described in the first aspect when executed by a processor.

[0019] The technical solutions provided by the embodiments of the present disclosure have the following beneficial effects:

[0020] In the disclosed embodiment, a Markov speed prediction model is used as the speed prediction model. A target driving data set for a first time period during vehicle travel is used to update the speed prediction model for predicting the vehicle speed during the first time period, thereby generating a second speed prediction model. The second speed prediction model is then used to predict the vehicle speed during the second time period. In other words, the speed prediction model can be updated using the vehicle's actual driving data while the vehicle is traveling. This allows the speed predicted by the second speed prediction model to more closely approximate the vehicle's actual speed, thereby improving the accuracy of the speed prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0022] Figure 1 is a flow chart of a vehicle speed prediction method provided by an embodiment of the present disclosure;

[0023] Figure 2 is a flow chart of another vehicle speed prediction method provided by an embodiment of the present disclosure;

[0024] Figure 3 This is a structural block diagram of a vehicle speed prediction device provided by an embodiment of the present disclosure;

[0025] Figure 4 This is a structural block diagram of a computer device provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0026] In order to make the objectives, technical solutions and advantages of the present disclosure more clear, the embodiments of the present disclosure will be further described in detail below with reference to the accompanying drawings.

[0027] Hybrid vehicles require energy management strategies to coordinate the distribution of power requirements across multiple energy sources, improving vehicle fuel economy and emissions without sacrificing power. Speed ​​prediction is fundamental to the effective application of energy management strategies.

[0028] Figure 1 This is a flow chart of a vehicle speed prediction method provided by an embodiment of the present disclosure, which can be executed by a vehicle controller. Figure 1 , the method comprising:

[0029] In step 101, a target driving data set during vehicle driving is obtained.

[0030] The target driving data set includes a plurality of driving data groups collected within a first time period.

[0031] The first time period refers to a time period of a set length before the current moment.

[0032] Each driving data set includes a collection time, a vehicle acceleration, and a vehicle speed. For example, the vehicle speed can be obtained by a speed sensor installed on the vehicle, and the vehicle acceleration can be obtained by an acceleration sensor installed on the vehicle.

[0033] In step 102 , the target driving data set is used to update the first vehicle speed prediction model used in the first time period to obtain a second vehicle speed prediction model.

[0034] The first and second vehicle speed prediction models are Markov vehicle speed prediction models, which are used to describe the probability distribution of a change from a first acceleration and a first speed state to a second acceleration at each prediction moment. The first vehicle speed prediction model used in the first time period is used to predict the vehicle speed in the first time period.

[0035] Given a known state at a given moment t0, the state at time t>t0 is only related to the state at t0, and not to the state before t0. This random process is called a Markov process. As a vehicle moves, changes in road conditions and the driver's driving intent lead to uncertain changes in vehicle speed. Future vehicle speed can be considered a random variable, and changes in vehicle speed can be considered a Markov process. Therefore, in the disclosed embodiments, a Markov speed prediction model is constructed to predict vehicle speed.

[0036] In step 103, a second vehicle speed prediction model is used to predict the vehicle speed in the second time period.

[0037] In some examples, a vehicle's driving process includes multiple consecutive, adjacent time periods. The first time period is one of the multiple adjacent time periods. The second time period is a time period subsequent to and adjacent to the first time period. For example, the first time period is t+1-ΔT to t, and the second time period is t+1 to t+ΔT. ΔT represents a set length in seconds.

[0038] In the disclosed embodiment, a Markov speed prediction model is used as the speed prediction model. A target driving data set for a first time period during vehicle travel is used to update the speed prediction model for predicting the vehicle speed during the first time period, thereby generating a second speed prediction model. The second speed prediction model is then used to predict the vehicle speed during the second time period. In other words, the speed prediction model can be updated using the vehicle's actual driving data while the vehicle is traveling. This allows the speed predicted by the second speed prediction model to more closely approximate the vehicle's actual speed, thereby improving the accuracy of the speed prediction.

[0039] Figure 2 This is a flow chart of another vehicle speed prediction method provided by an embodiment of the present disclosure, which can be executed by a vehicle controller. Figure 2 , the method comprising:

[0040] In step 201 , a reference driving data set is obtained.

[0041] The reference driving data set includes driving data corresponding to standard driving conditions under four driving scenarios: highway, suburban, urban smooth and urban congestion.

[0042] Exemplarily, the standard driving conditions include, but are not limited to, at least one of the HWFET (Highway Fuel Economy Test Cycle) condition, the US06_HWY (US06_Highway) condition, the INDIA_HWY_SAMPLE (Indian Highway Sample) condition, the WVUINTER (West Virginia Interstate Driving Schedule) condition, the UDDS (Urban Dynamometer Driving Schedule) condition, the INDIA_URBAN_SAMPLE (Indian City Sample) condition, the NYCC (The New York City Cycle) condition, and the MANHATTAN (Manhattan) condition.

[0043] The driving data corresponding to the standard driving condition describes how the vehicle's speed and acceleration change over time. The driving data corresponding to the standard driving condition is set based on the specific standard driving condition and stored in the vehicle controller's memory unit.

[0044] In step 202, an initial vehicle speed prediction model is obtained based on a reference driving data set.

[0045] The initial vehicle speed prediction model is used to predict the vehicle speed in the first time period. That is, when the vehicle is first operated, the initial vehicle speed prediction model can be used to predict the vehicle speed.

[0046] In some embodiments, step 202 includes the following steps:

[0047] The first step is to determine the state space. The state space consists of an acceleration sequence and a velocity sequence. The acceleration sequence and velocity sequence are discrete, equally spaced sequences. The spacing between the acceleration sequence and velocity sequence can be set according to actual needs.

[0048] The acceleration values ​​included in the acceleration sequence are all greater than or equal to a first acceleration threshold and less than or equal to a second acceleration threshold. The first acceleration threshold is a preset minimum acceleration, and the second acceleration threshold is a preset maximum acceleration.

[0049] The speed values ​​included in the speed sequence are all greater than or equal to a first speed threshold and less than or equal to a second speed threshold. The first speed threshold is a preset minimum speed, and the second speed threshold is a preset maximum speed.

[0050] The second step is to determine the state transition frequency matrix corresponding to the reference driving data set based on the state space.

[0051] The state transition frequency matrix corresponding to the reference driving data set describes the number of acceleration state transitions that occur at each predicted moment, given the current speed and acceleration state. In this embodiment of the present application, the reference driving data set corresponds to multiple state transition frequency matrices, each corresponding to a predicted moment.

[0052] For example, the state transition frequency matrix corresponding to the reference driving data set can be expressed using formula (1), which is as follows:

[0053]

[0054] In formula (1), Represents the state transition frequency matrix at the mth prediction moment corresponding to the reference driving data set. m∈{1, 2, 3, …, Lp}, where Lp is the preset vehicle speed prediction time length.

[0055] i and j represent the i-th and j-th discrete acceleration states respectively, i∈{1, 2, 3, ..., La}, j∈{1, 2, 3, ..., La}, La represents the number of discrete acceleration states corresponding to the acceleration sequence, and La is less than or equal to the number of discrete accelerations corresponding to the first acceleration sequence.

[0056] n represents the nth discrete speed state, n∈{1, 2, 3, ..., Lv}, Lv represents the state number of discrete speeds corresponding to the speed sequence, and Lv is less than or equal to the number of discrete speeds corresponding to the speed sequence.

[0057] a(k) and v(k) represent the acceleration and speed of the vehicle at the time k in the reference driving data set. k represents the current time, and k traverses the driving data corresponding to the reference driving data set at a set time interval Δt (for example, Δt=1). a i Indicates the i-th discrete acceleration corresponding to a(k) in the acceleration sequence. n Indicates the nth discrete acceleration corresponding to v(k) in the velocity sequence. i and v n The nearest-neighborhood method (NNM) is used to obtain a(k) and v(k) respectively.

[0058] The following takes a(k) as an example to illustrate how to determine a i The process is as follows: determine the difference between a(k) and each discrete acceleration value in the acceleration sequence; determine the discrete acceleration value corresponding to the minimum absolute value of the difference between the acceleration sequence and a(k) as a i .

[0059] a(k+m) represents the acceleration of the vehicle corresponding to the mth prediction moment when the travel time is k (that is, the acceleration of the vehicle corresponding to the (k+m) moment).

[0060] N(a(k+m)=a j |a(k)=a i , v(k)=v n ) means that when a(k) is equal to a i And v(k) is equal to v n When a(k+m) is a j The number of times. j represents the jth discrete acceleration in the acceleration sequence.

[0061] In some examples, the state transition frequency matrix at the mth prediction moment is a matrix with a size of La rows and (La*Lv) columns. Where (La*Lv) represents the total number of states corresponding to the acceleration sequence and the speed sequence, and La represents the number of discrete acceleration states of the acceleration sequence. The element N in the pth row and qth column of the state transition frequency matrix corresponding to the reference driving data set is pq Indicates the number of transitions from state q to state p.

[0062] The third step is to determine the state transition frequency matrix corresponding to the reference driving data set based on the state transition frequency matrix corresponding to the reference driving data set.

[0063] The state transition frequency matrix corresponding to the reference driving data set is used to describe the probability distribution of the vehicle's acceleration state transition at each predicted moment, given the current speed and acceleration state. In this embodiment of the present application, the reference driving data set corresponds to multiple state transition frequency matrices, each corresponding to a predicted moment.

[0064] For example, the state transition frequency matrix corresponding to the reference driving data set is calculated using formula (2), which is as follows:

[0065]

[0066] In formula (2), It represents the state transition frequency matrix of the reference driving data set corresponding to the mth prediction moment. For the relevant contents of i, j, m and n, please refer to the aforementioned formula (1), and the detailed description is omitted here. Represents the state transition frequency matrix in formula (1). When the state number of the acceleration sequence is La, when a(k) is equal to a i And v(k) is equal to v n In the case of , the sum of the number of times all discrete acceleration values ​​are transferred to the acceleration sequence.

[0067] In some examples, the state transition frequency matrix at the mth prediction moment corresponding to the reference driving data set is a matrix with a size of La rows and (La*Lv) columns. Where (La*Lv) represents the total number of states corresponding to the acceleration sequence and the velocity sequence, and La represents the number of discrete acceleration states of the acceleration sequence. The element N in the pth row and qth column of the state transition frequency matrix corresponding to the reference driving data set is pq ′ represents the probability of transitioning from state q to state p.

[0068] Step 4: Obtain the initial vehicle speed prediction model based on the state transition frequency matrix corresponding to the reference driving data set.

[0069] In the disclosed embodiment, the initial vehicle speed prediction model is a single-order multi-step model. In the single-order multi-step model, the state of the i-th vehicle speed is related to the state of the (i-1)-th vehicle speed.

[0070] In some implementations, the initial vehicle speed prediction model is obtained using formula (3), which is as follows:

[0071] v(t+m)=v(t+m-1)+a(m)*Δt*3.6 (3)

[0072] In formula (3), v(t+m) represents the speed of the vehicle predicted at the mth prediction moment when the current moment is t, and v(t+m-1) represents the speed of the vehicle predicted at the (m-1)th prediction moment. a(m) represents the expected acceleration at the mth prediction moment. Δt represents the time interval between adjacent prediction moments, for example, Δt is 1s. Since the time unit in the embodiment of the present disclosure is s and the unit of speed is km / h, 3.6 is a conversion coefficient used to convert the unit of a(m)*Δt to km / h. a(m) is obtained using formula (4), which is as follows:

[0073]

[0074] In formula (4), a(m) represents the expected acceleration at the mth prediction moment. j represents the jth discrete acceleration in the acceleration sequence, La represents the number of discrete acceleration states in the acceleration sequence, It represents the state transition frequency matrix corresponding to the reference driving data set obtained by the above formula (2).

[0075] Optionally, in the embodiment of the present disclosure, steps 201 and 202 are optional steps. In other embodiments, the initial vehicle speed prediction model obtained based on the reference driving data set is directly stored in the storage unit of the vehicle controller, and the vehicle controller can directly obtain the initial vehicle speed prediction model from the storage unit.

[0076] In step 203, a target driving data set during the vehicle driving process is obtained.

[0077] The target driving data set includes multiple driving data groups collected in the first time period. For details about the target driving data set, refer to the aforementioned step 101, and a detailed description is omitted here.

[0078] In step 204 , the target driving data set is used to update the first vehicle speed prediction model used in the first time period to obtain a second vehicle speed prediction model.

[0079] When the speed prediction model for predicting the vehicle speed in the first time period is updated for the first time, the speed prediction model for predicting the vehicle speed in the first time period is trained using the reference driving data set, i.e., the aforementioned initial speed prediction model. When the speed prediction model for predicting the vehicle speed in the first time period is updated for the nth time, the speed prediction model for predicting the vehicle speed in the first time period is the second speed prediction model obtained after the previous update (i.e., the n-1th update), where n is an integer greater than 1.

[0080] For example, the vehicle travel process includes time periods T1-T2, T2-T3, and T3-T4. The speed prediction model for predicting the vehicle speed in time period T1-T2 is the aforementioned initial speed prediction model; the speed prediction model for predicting the vehicle speed in time period T2-T3 is a second speed prediction model obtained by updating the speed model for predicting time period T1-T2; and the speed prediction model for predicting the vehicle speed in time period T3-T4 is a second speed prediction model obtained by updating the speed model for predicting time period T2-T3.

[0081] In the embodiment of the present disclosure, step 204 includes: a first step of determining a state transition frequency matrix corresponding to the target driving data set based on the target driving data set; a second step of updating the first vehicle speed prediction model based on the state transition frequency matrix corresponding to the target driving data set to obtain a second vehicle speed prediction model.

[0082] The state transition frequency matrix corresponding to the target driving data set describes the number of acceleration state transitions that occur at each predicted moment, given the vehicle's current speed and acceleration state. In this embodiment of the present application, the target driving data set corresponds to multiple state transition frequency matrices, each corresponding to a predicted moment.

[0083] The state transition frequency matrix corresponding to the target driving data set can be expressed using formula (5), which is as follows:

[0084]

[0085] In formula (5), Represents the state transition frequency matrix at the mth prediction moment corresponding to the target driving data set. For details about m, see formula (1).

[0086] i and j represent the i-th and j-th discrete acceleration states, respectively. i∈{1, 2, 3, ..., La} and j∈{1, 2, 3, ..., La}, where La represents the number of discrete acceleration states corresponding to the acceleration sequence, and La is less than or equal to the number of discrete accelerations corresponding to the second acceleration sequence. n represents the n-th discrete velocity state, n∈{1, 2, 3, ..., Lv}, where Lv represents the number of discrete velocity states corresponding to the velocity sequence, and Lv is less than or equal to the number of discrete velocities corresponding to the velocity sequence.

[0087] a(k) and v(k) represent the acceleration and speed of the vehicle at the time k in the target driving data set. k represents the current time, and k traverses the driving data corresponding to the target driving data set at a set time interval Δt (for example, Δt=1). a i Indicates the i-th discrete acceleration corresponding to a(k) in the acceleration sequence.n Indicates the nth discrete velocity corresponding to v(k) in the velocity sequence. i and υ n The NNM rule is applied to a(k) and v(k) respectively. The relevant contents of the NNM rule are referred to the aforementioned step 202, and the detailed description is omitted here.

[0088] a(k+m) represents the acceleration of the vehicle corresponding to the mth predicted moment when the travel time is k (that is, the acceleration of the vehicle corresponding to the moment (k+m)). N(a(k+m)=a j |a(k)=a i , v(k)=v n ) means that when a(k) is equal to a i And v(k) is equal to υ n When a(k+m) is a j The number of times. j represents the jth discrete acceleration in the acceleration sequence.

[0089] In some embodiments, the first vehicle speed prediction model includes a first state transition frequency matrix, and the second vehicle speed prediction model includes a target state transition frequency matrix. The first vehicle speed prediction model is updated based on the state transition frequency matrix corresponding to the target driving data set to obtain the second vehicle speed prediction model, including:

[0090] The first step is to calculate a target state transition frequency matrix based on a set forgetting factor, the first state transition frequency matrix, and the state transition frequency matrix corresponding to the target driving data set. The set forgetting factor is used to modify the first vehicle speed prediction model used during the first time period. For example, the target state transition frequency matrix can be obtained by adding the product of the set forgetting factor and the first state transition frequency matrix to the state transition frequency matrix corresponding to the target driving data set.

[0091] In the embodiment of the present disclosure, formula (6) is used to calculate the target state transition frequency matrix. Formula (6) is as follows:

[0092]

[0093] In formula (6), μ represents the set forgetting factor, μ is greater than 0 and less than 1, and the set μ is stored in the storage unit of the vehicle controller. K represents the number of updates of the vehicle speed prediction model, and the initial value of K is 0. TPM(K) represents the state transition frequency matrix corresponding to the vehicle speed prediction model for predicting the vehicle speed in the first time period. TPM(0) is the state transition frequency matrix corresponding to the reference driving data set in step 202. TFM(K) represents the state transition frequency matrix corresponding to the target driving data set. TFM(K+1) represents the updated target state transition frequency matrix.

[0094] Since the value μ of the set forgetting factor is greater than 0 and less than 1, and the value in TPM(K) (state transition frequency matrix) is also greater than 0 and less than 1, by multiplying the reciprocal of the set forgetting factor with TPM(K), TPM(K) can be converted into a state transition frequency matrix at a specified ratio.

[0095] The second step is to determine the target state transfer frequency matrix according to the target state transfer frequency matrix to obtain the second vehicle speed prediction model.

[0096] In some implementations, the target state transition frequency matrix is ​​calculated using formula (7), which is as follows:

[0097]

[0098] In formula (7), Represents the target state transition frequency matrix at the mth prediction moment. represents the target state transition frequency matrix at the mth prediction moment corresponding to formula (6), When the state number of the acceleration sequence is La, when a(k) is equal to a i And v(k) is equal to υ n In the case of , the sum of the number of times all discrete acceleration values ​​are transferred to the acceleration sequence.

[0099] The third step is to determine the second vehicle speed prediction model based on the target state transfer frequency matrix.

[0100] It should be noted that the difference between the second vehicle speed prediction model and the initial vehicle speed prediction model is that the state transition frequency matrix used to predict acceleration is different.

[0101] In the embodiment of the present disclosure, the forgetting factor can be pre-set before the vehicle speed prediction method of the embodiment of the present disclosure is executed. In practical applications, the forgetting factor can be set according to the vehicle speed prediction error obtained based on the debugging data set. For example, the forgetting factor can be set in the following manner: based on the debugging data set, multiple candidate forgetting factors are used to update the initial vehicle speed prediction model respectively (the updating method refers to the updating method of the target vehicle speed prediction model, which will not be repeated here), and multiple test vehicle speed prediction models are obtained, and the initial vehicle speed prediction model is a vehicle speed prediction model obtained by training the driving data corresponding to the standard driving condition; the vehicle speed prediction error corresponding to the multiple test vehicle speed prediction models is determined, and the vehicle speed prediction error is the error between the vehicle speed predicted by the corresponding test vehicle speed prediction model and the actual vehicle speed; the candidate forgetting factor corresponding to the test vehicle speed prediction model when the vehicle speed prediction error is less than the error threshold is determined as the set forgetting factor.

[0102] In some examples, the vehicle speed prediction error can be calculated using the following formula (8).

[0103]

[0104] In formula (8), represents the speed prediction error, that is, the mean of the root mean square error of the speed at all prediction moments. p Indicates the number of times the vehicle speed is predicted during the speed prediction experiment, Lp indicates the length of the speed prediction time, and υ p (t) represents the predicted vehicle speed, v r (t) represents the actual vehicle speed collected.

[0105] Exemplarily, the error threshold is set according to actual needs, and the embodiments of the present disclosure do not limit this.

[0106] In the embodiment of the present disclosure, the candidate forgetting factor when the vehicle speed prediction error is less than the error threshold is used as the set forgetting factor, which can make the updated second vehicle speed prediction model more accurate.

[0107] In step 205 , the second vehicle speed prediction model is used to predict the vehicle speed in the second time period.

[0108] The second time period is a time period after the first time period and adjacent to the first time period.

[0109] In the disclosed embodiment, a Markov speed prediction model is used as the speed prediction model. A target driving data set for a first time period during vehicle travel is used to update the speed prediction model for predicting the vehicle speed during the first time period, thereby generating a second speed prediction model. The second speed prediction model is then used to predict the vehicle speed during the second time period. In other words, the speed prediction model can be updated using the vehicle's actual driving data while the vehicle is traveling. This allows the speed predicted by the second speed prediction model to more closely approximate the vehicle's actual speed, thereby improving the accuracy of the speed prediction.

[0110] Figure 3 is a structural block diagram of a vehicle speed prediction device 300 provided by an embodiment of the present disclosure, such as Figure 3 As shown, the device includes: an acquisition module 301, an update module 302 and a vehicle speed prediction module 303.

[0111] Among them, the acquisition module 301 is used to obtain a target driving data set during the vehicle's driving process, and the target driving data set includes multiple driving data groups collected within a first time period, and each driving data group includes the collection time, the acceleration of the vehicle, and the speed of the vehicle. The updating module 302 is used to use the target driving data set to update the first vehicle speed prediction model used in the first time period to obtain a second vehicle speed prediction model. The first vehicle speed prediction model and the second vehicle speed prediction model are both Markov vehicle speed prediction models. The Markov vehicle speed prediction model is used to describe the probability distribution of changing from the first acceleration and the first speed state to the second acceleration at each prediction moment. The vehicle speed prediction module 303 is used to use the second vehicle speed prediction model to predict the vehicle speed in the second time period. The second time period is a time period after the first time period and adjacent to the first time period.

[0112] Optionally, the updating module 302 is used to determine a state transition frequency matrix corresponding to the target driving data set based on the target driving data set; and update the first vehicle speed prediction model based on the state transition frequency matrix to obtain the second vehicle speed prediction model.

[0113] Optionally, the first vehicle speed prediction model includes a first state transition frequency matrix; the update module 302 is used to calculate a target state transition frequency matrix based on a set forgetting factor, the first state transition frequency matrix and the state transition frequency matrix; and determine a target state transition frequency matrix based on the target state transition frequency matrix to obtain the second vehicle speed prediction model.

[0114] Optionally, the updating module 302 is configured to add the product of the set forgetting factor and the first state transition frequency matrix to the state transition frequency matrix to obtain the target state transition frequency matrix.

[0115] Optionally, the device also includes a determination module 304, which is used to, based on the debugging data set, use multiple candidate forgetting factors to update the initial vehicle speed prediction model respectively to obtain multiple test vehicle speed prediction models, where the initial vehicle speed prediction model is a vehicle speed prediction model trained using driving data corresponding to a standard driving condition; determine the vehicle speed prediction errors corresponding to the multiple test vehicle speed prediction models, where the vehicle speed prediction error is the error between the vehicle speed predicted by the corresponding test vehicle speed prediction model and the actual vehicle speed; and determine the candidate forgetting factor corresponding to the test vehicle speed prediction model when the vehicle speed prediction error is less than the error threshold as the set forgetting factor.

[0116] It should be noted that the vehicle speed prediction device provided in the above embodiment is merely illustrated by the division of the aforementioned functional modules when performing vehicle speed prediction. In actual applications, the aforementioned functions can be assigned to different functional modules as needed, i.e., the internal structure of the device can be divided into different functional modules to perform all or part of the functions described above. Furthermore, the vehicle speed prediction device provided in the above embodiment and the vehicle speed prediction method embodiment are based on the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0117] Figure 4 This is a block diagram of the computer device provided by the embodiment of the present disclosure. Figure 4 As shown, the computer device 400 may be a vehicle-mounted computer, etc. The computer device 400 includes: a processor 401 and a memory 402 .

[0118] The processor 401 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 401 may be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), or PLA (Programmable Logic Array). The processor 401 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the awake state, also known as a CPU (Central Processing Unit); the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor 401 may be integrated with a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 401 may also include an AI (Artificial Intelligence) processor, which is used to process computing operations related to machine learning.

[0119] Memory 402 may include one or more computer-readable media, which may be non-transitory. Memory 402 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices or flash memory storage devices. In some embodiments, the non-transitory computer-readable medium in memory 402 is used to store at least one instruction, which is executed by processor 401 to implement the vehicle speed prediction method provided in the embodiments of the present disclosure.

[0120] Those skilled in the art will understand that Figure 4 The structure shown in the figure does not constitute a limitation on the computer device 400, and the computer device 400 may include more or fewer components than shown in the figure, or combine some components, or adopt a different component arrangement.

[0121] The embodiment of the present disclosure further provides a non-transitory computer-readable medium. When the instructions in the medium are executed by the processor of the computer device 400, the computer device 400 is enabled to execute the vehicle speed prediction method provided in the embodiment of the present disclosure.

[0122] The embodiment of the present disclosure further provides a computer program product, including a computer program / instruction, which implements the vehicle speed prediction method provided in the embodiment of the present disclosure when the computer program / instruction is executed by a processor.

[0123] The above are merely optional embodiments of the present disclosure and are not intended to limit the present disclosure. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present disclosure shall be included in the scope of protection of the present disclosure.

Claims

1. A vehicle speed prediction method, characterized in that: The method comprises: Obtaining a target driving data set during vehicle driving, the target driving data set comprising a plurality of driving data groups collected within a first time period, each driving data group comprising a collection time, an acceleration of the vehicle, and a speed of the vehicle; the vehicle driving process comprising a plurality of consecutive adjacent time periods, the first time period being one of the plurality of consecutive adjacent time periods; and predicting a vehicle speed within the first time period using a first vehicle speed prediction model; Determining a state transition frequency matrix corresponding to the target driving data set based on the target driving data set; the state transition frequency matrix is ​​used to describe the number of acceleration state transitions corresponding to each predicted time point of the vehicle under the current speed and acceleration state; updating the first vehicle speed prediction model according to the state transition frequency matrix to obtain a second vehicle speed prediction model, wherein the first vehicle speed prediction model and the second vehicle speed prediction model are both Markov vehicle speed prediction models, and the Markov vehicle speed prediction models are used to describe the probability distribution of changing from the first acceleration and the first speed state to the second acceleration at each prediction moment; The second vehicle speed prediction model is used to predict the vehicle speed in a second time period, where the second time period is a time period after the first time period and adjacent to the first time period.

2. The method according to claim 1, characterized in that The first vehicle speed prediction model includes a first state transition frequency matrix; The updating of the first vehicle speed prediction model according to the state transition frequency matrix to obtain the second vehicle speed prediction model includes: Calculating a target state transition frequency matrix according to the set forgetting factor, the first state transition frequency matrix, and the state transition frequency matrix; According to the target state transition frequency matrix, a target state transition frequency matrix is ​​determined to obtain the second vehicle speed prediction model.

3. The method according to claim 2, characterized in that The step of calculating a target state transition frequency matrix according to the set forgetting factor, the first state transition frequency matrix, and the state transition frequency matrix includes: The product of the set forgetting factor and the first state transition frequency matrix is ​​added to the state transition frequency matrix to obtain the target state transition frequency matrix.

4. The method according to claim 2 or 3, characterized in that The method further comprises: Based on the debugging data set, multiple candidate forgetting factors are used to update the initial vehicle speed prediction model to obtain multiple test vehicle speed prediction models, wherein the initial vehicle speed prediction model is a vehicle speed prediction model trained using driving data corresponding to a standard driving condition; Determining vehicle speed prediction errors corresponding to the multiple test vehicle speed prediction models, wherein the vehicle speed prediction error is the error between the vehicle speed predicted by the corresponding test vehicle speed prediction model and the actual vehicle speed; The candidate forgetting factor corresponding to the test vehicle speed prediction model when the vehicle speed prediction error is less than the error threshold is determined as the set forgetting factor.

5. A vehicle speed prediction device, characterized in that: The device comprises: an acquisition module, configured to obtain a target driving data set during vehicle travel, the target driving data set comprising a plurality of driving data groups collected within a first time period, each driving data group comprising a collection time, an acceleration of the vehicle, and a speed of the vehicle; the vehicle travel process comprising a plurality of consecutive adjacent time periods, the first time period being one of the plurality of consecutive adjacent time periods; and within the first time period, predicting the vehicle speed using a first vehicle speed prediction model; an updating module for determining, based on the target driving data set, a state transition frequency matrix corresponding to the target driving data set; the state transition frequency matrix is ​​used to describe, under the current speed and acceleration state, the number of acceleration state transitions corresponding to each prediction moment; and updating the first vehicle speed prediction model based on the state transition frequency matrix to obtain a second vehicle speed prediction model, wherein both the first vehicle speed prediction model and the second vehicle speed prediction model are Markov vehicle speed prediction models, and the Markov vehicle speed prediction model is used to describe the probability distribution of changing from the first acceleration and the first speed state to the second acceleration at each prediction moment; The vehicle speed prediction module is used to use the second vehicle speed prediction model to predict the vehicle speed in a second time period, where the second time period is a time period after the first time period and adjacent to the first time period.

6. A computer device, characterized in that: include: processor; a memory for storing processor-executable instructions; The processor is configured to execute the method according to any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that When the instructions in the computer-readable storage medium are executed by a processor of a computer device, the computer device is enabled to perform the method according to any one of claims 1 to 4.

8. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the method according to any one of claims 1 to 4 is implemented.

Images