CN117184032A - Driving control method and device of hybrid power fire engine and training method of driving condition prediction model - Google Patents
Driving control method and device of hybrid power fire engine and training method of driving condition prediction model Download PDFInfo
- Publication number
- CN117184032A CN117184032A CN202311149831.0A CN202311149831A CN117184032A CN 117184032 A CN117184032 A CN 117184032A CN 202311149831 A CN202311149831 A CN 202311149831A CN 117184032 A CN117184032 A CN 117184032A
- Authority
- CN
- China
- Prior art keywords
- working condition
- driving
- condition
- period
- training
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000003062 neural network model Methods 0.000 claims description 53
- 238000003064 k means clustering Methods 0.000 claims description 36
- 238000004138 cluster model Methods 0.000 claims description 4
- 239000002699 waste material Substances 0.000 abstract description 4
- 238000005070 sampling Methods 0.000 description 18
- 230000006870 function Effects 0.000 description 8
- 238000013528 artificial neural network Methods 0.000 description 4
- 239000002131 composite material Substances 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 2
- 239000007789 gas Substances 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000007621 cluster analysis Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 239000000446 fuel Substances 0.000 description 1
- 239000000295 fuel oil Substances 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000003921 oil Substances 0.000 description 1
- 239000001301 oxygen Substances 0.000 description 1
- 229910052760 oxygen Inorganic materials 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Landscapes
- Hybrid Electric Vehicles (AREA)
Abstract
The invention discloses a driving control method and device of a hybrid power fire truck and a training method of a driving condition prediction model, wherein the method is based on the driving condition prediction model; the driving condition prediction model is input into driving condition characteristic parameters and rewarding values corresponding to M working condition periods before a target working condition period of the hybrid power fire truck, and output is a predicted driving condition type corresponding to the target working condition period of the hybrid power fire truck. The invention provides a driving working condition prediction model which is used for outputting the prediction of the working condition type after the historical working condition period according to the characteristic parameters and the rewarding value of the historical working condition period, and the speed of a hybrid power fire engine can be adjusted in advance due to the fact that the working condition is predicted in advance based on the prediction function of the model, so that energy waste caused by sudden speed change can be reduced, energy is further saved, meanwhile, electric energy is reserved to be used in a tunnel environment as much as possible, and tunnel rescue requirements are met.
Description
Technical Field
The invention relates to the technical field of hybrid power fire engines, in particular to a running control method and device of a hybrid power fire engine and a training method of a running condition prediction model.
Background
Fire-fighting works need to face various fire-fighting scenes, taking tunnel rescue as an example, when a fire occurs in a tunnel, because oxygen in the tunnel is relatively less, fire-fighting vehicles are not suitable for using fuel oil to provide power in the tunnel, and generated tail gas also aggravates polluted air and shields vision, so that for tunnel rescue, heavy hybrid fire-fighting vehicles are used in more and more fire-fighting works to carry (a large amount of) rescue materials, and only batteries are used in the tunnel. However, the energy of the battery is limited, and it is important to plan the use of oil and gas during driving to avoid consuming the electric energy of the battery without reaching the tunnel.
Disclosure of Invention
The invention aims to solve the technical problems of providing a driving control method and device of a hybrid power fire engine and a training method of a driving condition prediction model, which can predict the condition in advance, so that the speed of the hybrid power fire engine is adjusted in advance, the energy waste caused by sudden speed change is reduced, the energy is saved, the electric energy is reserved in the tunnel environment as much as possible, and the tunnel rescue requirement is met.
In order to solve the technical problems, the first aspect of the invention discloses a driving control method of a hybrid power fire truck, which is based on a driving condition prediction model; the input of the running condition prediction model is a running condition characteristic parameter and a reward value corresponding to M working condition periods before a target working condition period of the hybrid power fire truck, and the output of the running condition characteristic parameter and the reward value is a predicted running condition type corresponding to the target working condition period of the hybrid power fire truck;
the method comprises the following steps:
judging whether the last prediction of the driving working condition prediction model is correct, and determining that the rewarding value corresponding to M working condition periods of the hybrid fire truck before the current working condition period is 1 when the last prediction of the driving working condition prediction model is correct, otherwise, determining that the rewarding value is 0;
acquiring the driving data of the hybrid power fire truck in M working condition periods before the current working condition period, calculating driving working condition characteristic parameters corresponding to each working condition period according to the driving data of the M working condition periods, and inputting the driving working condition characteristic parameters and rewarding values corresponding to the M working condition periods into the driving working condition prediction model;
acquiring the driving condition type output by the driving condition prediction model, and determining the driving condition type as the predicted driving condition type of the hybrid fire truck in the current working condition period;
and controlling the running speed of the hybrid power fire truck according to the predicted running condition type.
As an optional implementation manner, in the first aspect of the present invention, the controlling the driving speed of the hybrid fire engine according to the predicted driving condition type includes:
when the predicted running condition type is a condition type that the vehicle speed is smaller than the current vehicle speed, gradually reducing the vehicle speed of the hybrid fire engine along with time so as to enable the vehicle speed to stably reach the vehicle speed corresponding to the predicted running condition type;
when the predicted running condition type is a condition type that the vehicle speed is smaller than the current vehicle speed, the vehicle speed of the hybrid fire engine is gradually increased along with time, so that the vehicle speed stably reaches the vehicle speed corresponding to the predicted running condition type.
As a further optional implementation manner, in the first aspect of the present invention, the driving condition prediction model includes a neural network model trained in advance;
the input layer of the neural network model is provided with M+1 input nodes, and the value of each input node is sequentially a driving working condition characteristic parameter and a reward value corresponding to M working condition periods before a target working condition period of the hybrid fire engine;
the output layer of the neural network model is provided with K output nodes, each output node corresponds to one working condition cluster type one by one, and the value of each output node is the probability that the neural network model predicts that the driving working condition type of the hybrid fire truck under the target working condition period belongs to the working condition cluster type corresponding to the output node.
As a further optional implementation manner, in the first aspect of the present invention, when the driving condition prediction model predicts correctly last time, it is determined that the reward value corresponding to M working condition periods of the hybrid fire truck before the target working condition period is 1, otherwise, it is 0.
As a further optional implementation manner, in the first aspect of the present invention, the driving condition prediction model further includes a determining module; the determining module is used for obtaining values of all output nodes of the neural network model and determining that the predicted driving condition type of the hybrid fire engine corresponding to the target condition period is the condition cluster type corresponding to the output node with the largest value.
The second aspect of the invention discloses a training method of a driving condition prediction model of a hybrid power fire engine, which is used for training the neural network model according to the first aspect of the invention; the method comprises the following steps:
acquiring the travel data of a pre-collected historical working condition period of the hybrid power fire truck, dividing the travel data of the historical working condition period into N travel data of training working condition periods, wherein the duration of each training working condition period is the length of a preset period;
calculating corresponding running condition characteristic parameters for the running data of each training condition period, inputting the running condition characteristic parameters of each training condition period into a predetermined k-means clustering model, and obtaining the running condition type corresponding to the training condition period output by the k-means clustering model; the driving working condition characteristic parameters of the training working condition period comprise the average speed, the parking proportion, the uniform speed proportion, the deceleration proportion and the maximum speed of the hybrid power fire truck under the working condition period corresponding to the driving data of the training working condition period;
in N training working condition periods N 1 The continuous training working condition time periods are target training working condition time periods, and the driving working condition characteristic parameters of M continuous training working condition time periods are started from each target training working condition time periodThe number and the running condition type corresponding to the next training condition period of the M training condition periods are sample data corresponding to the target training condition period, and N is obtained 1 Sample data according to the N 1 Training the neural network model with the sample data; said M and said N 1 Are all smaller than N;
judging whether the training ending condition is met or not;
and when the judgment is satisfied, finishing the training of the neural network model, and determining the current neural network model as a pre-trained neural network model.
In a second aspect of the present invention, as an alternative embodiment, the method according to the N 1 Training the neural network model from the sample data, comprising:
for the first target training working condition period, taking driving working condition characteristic parameters and initial rewards of M continuous training working condition periods which are started by sample data corresponding to the target training working condition period as input of the neural network model, judging whether the driving working condition type corresponding to the maximum value in an output node of the neural network model is consistent with the driving working condition type corresponding to the next training working condition period of the M training working condition periods, and determining that the rewards corresponding to the second target training working condition period is 1 when the driving working condition type is consistent with the driving working condition type corresponding to the next training working condition period of the M training working condition periods, otherwise, determining that the rewards corresponding to the second target training working condition period is 0;
and for each target training working condition period except the first target training working condition period, taking the driving working condition characteristic parameters and corresponding rewards of M continuous training working condition periods which are started by sample data corresponding to the target training working condition period as the input of the neural network model, judging whether the driving working condition type corresponding to the maximum value in the output node of the neural network model is consistent with the driving working condition type corresponding to the next training working condition period of the M training working condition periods, and determining that the rewards corresponding to the next training working condition period is 1 when the driving working condition types are judged to be consistent, otherwise, determining that the rewards corresponding to the next training working condition period is 0.
As yet another alternative embodiment, in the second aspect of the present invention, the k-means clustering model is predetermined by the following method:
acquiring standard working condition time period data of the hybrid power fire truck under various driving working condition types, and calculating driving working condition characteristic parameters corresponding to each standard working condition time period data; the driving working condition characteristic parameters of each standard working condition period data comprise the average speed, the parking proportion, the uniform speed proportion, the deceleration proportion and the maximum speed of the hybrid power fire truck in the working condition period corresponding to the standard working condition period data;
a plurality of values are pre-fetched for the number of clustering centers of the k-means clustering model to obtain a plurality of k-means clustering models to be determined;
inputting driving condition characteristic parameters of the standard condition period data into each k-means clustering model to be determined, and obtaining a clustering scheme output by the k-means clustering model to be determined; the clustering scheme of the k-means clustering model to be determined is a plurality of groups of standard working condition period data obtained by clustering the standard working condition period data under a plurality of driving working condition types by the k-means clustering model to be determined, wherein each group of standard working condition period data is classified into one working condition clustering type, and the number of the working condition clustering types is equal to the number of clustering centers of the k-means clustering model to be determined;
calculating the contour coefficient of each clustering scheme, and selecting a k-means clustering model to be determined corresponding to the maximum contour coefficient as a predetermined k-means clustering model; and the number K of the output nodes of the neural network model is equal to the number of the working condition cluster types corresponding to the predetermined K-means cluster model.
The third aspect of the present invention discloses a travel control device of a hybrid fire engine, comprising:
a memory storing executable program code;
a processor coupled to the memory;
the processor invokes the executable program code stored in the memory to execute the steps in the running control method of the hybrid fire engine disclosed in the first aspect of the present invention.
A third aspect of the present invention discloses a computer storage medium storing computer instructions for executing the steps of a travel control method of a hybrid fire engine disclosed in the first aspect of the present invention when the computer instructions are called.
Compared with the prior art, the embodiment of the invention has the following beneficial effects:
compared with the prior art, the embodiment of the invention provides the driving condition prediction model which is used for outputting the prediction of the condition type after the historical condition period according to the characteristic parameters and the rewarding value of the historical condition period, and the prediction function based on the model can adjust the speed of the hybrid fire truck in advance due to the fact that the condition is predicted in advance, so that energy waste caused by sudden speed change can be reduced, energy is saved, electric energy is reserved to be used in a tunnel environment as much as possible, and tunnel rescue requirements are met.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a schematic flow chart of a driving control method of a hybrid fire engine according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of a driving condition prediction model of a hybrid fire engine according to an embodiment of the present invention;
FIG. 3 is a schematic flow chart of a training method of a driving condition prediction model of a hybrid fire engine according to an embodiment of the present invention;
FIG. 4 is a chart of 33 typical common operating conditions disclosed in an embodiment of the present invention;
FIG. 5 is a graph of data for a composite operating condition according to an embodiment of the present invention;
FIG. 6 is a diagram of identifying and classifying driving conditions according to an embodiment of the present invention;
fig. 7 is a schematic structural diagram of a driving control device of a hybrid fire engine according to an embodiment of the present invention.
Detailed Description
In order that those skilled in the art will better understand the present invention, a technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Example 1
Referring to fig. 1, an embodiment of the invention discloses a driving control method of a hybrid power fire truck, which is based on a driving condition prediction model; the input of the running condition prediction model is a running condition characteristic parameter and a reward value corresponding to M working condition periods before a target working condition period of the hybrid power fire truck, and the output of the running condition characteristic parameter and the reward value is a predicted running condition type corresponding to the target working condition period of the hybrid power fire truck; the method comprises the following steps:
101. judging whether the prediction of the driving working condition prediction model is correct or not, and determining that the rewarding value corresponding to M working condition periods of the hybrid fire truck before the current working condition period is 1 when the prediction is correct, otherwise, determining that the rewarding value is 0.
102. And acquiring the driving data of the hybrid power fire truck in M working condition periods before the current working condition period, calculating the driving working condition characteristic parameters corresponding to each working condition period according to the driving data of the M working condition periods, and inputting the driving working condition characteristic parameters and the rewarding values corresponding to the M working condition periods into the driving working condition prediction model.
103. And acquiring the driving condition type output by the driving condition prediction model, and determining the driving condition type as the predicted driving condition type of the hybrid fire truck in the current working condition period.
104. And controlling the running speed of the hybrid power fire truck according to the predicted running condition type.
In this embodiment, the hybrid fire engine uses fuel rather than electric energy preferentially when traveling in a non-tunnel environment.
The embodiment provides a driving condition prediction model, which is used for outputting the prediction of the condition type after the historical condition period according to the characteristic parameters and the rewarding value of the historical condition period, and the speed of the hybrid fire engine can be adjusted in advance due to the fact that the condition is predicted in advance based on the prediction function of the model, so that energy waste caused by sudden speed change can be reduced, energy is further saved, meanwhile, electric energy is reserved to be used in a tunnel environment as much as possible, and tunnel rescue requirements are met.
In this embodiment, the driving condition characteristic parameter of one training condition period includes an average speed, a parking ratio, a uniform speed ratio, a deceleration ratio and a maximum speed of the hybrid fire truck in a condition period corresponding to the driving data of the training condition period, and is specifically calculated according to the following formula:
(1) Average speed = sum of speeds of all sampling points/number of corresponding sampling points, unit km/h;
(2) Maximum speed = maximum value of sampling points during driving, unit km/h;
(3) Deceleration ratio = number of sampling points/number of sampling points in a deceleration state;
(4) The constant speed ratio = the number of sampling points/the number of sampling points in a constant speed state;
(5) Parking ratio = number of sampling points in parking state/number of sampling points.
In an alternative embodiment, the controlling the driving speed of the hybrid fire engine according to the predicted driving condition type includes:
when the predicted running condition type is a condition type that the vehicle speed is smaller than the current vehicle speed, gradually reducing the vehicle speed of the hybrid fire engine along with time so as to enable the vehicle speed to stably reach the vehicle speed corresponding to the predicted running condition type;
when the predicted running condition type is a condition type that the vehicle speed is smaller than the current vehicle speed, the vehicle speed of the hybrid fire engine is gradually increased along with time, so that the vehicle speed stably reaches the vehicle speed corresponding to the predicted running condition type.
In yet another alternative embodiment, referring to FIG. 2, the driving condition prediction model includes a pre-trained neural network model 101;
the input layer of the neural network model is provided with M+1 input nodes, and the value of each input node is sequentially a driving working condition characteristic parameter and a reward value corresponding to M working condition periods before a target working condition period of the hybrid fire engine;
the output layer of the neural network model is provided with K output nodes, each output node corresponds to one working condition cluster type one by one, and the value of each output node is the probability that the neural network model predicts that the driving working condition type of the hybrid fire truck under the target working condition period belongs to the working condition cluster type corresponding to the output node.
In yet another alternative embodiment, when the driving condition prediction model predicts correctly last time, it is determined that the reward value corresponding to M working condition periods of the hybrid fire truck before the target working condition period is 1, otherwise it is 0.
In yet another alternative embodiment, referring to FIG. 2, the driving condition prediction model further includes a determination module 102; the determining module is used for obtaining values of all output nodes of the neural network model and determining that the predicted driving condition type of the hybrid fire engine corresponding to the target condition period is the condition cluster type corresponding to the output node with the largest value.
In this embodiment, M may be 60. The neural network model may be a deep neural network having two hidden layers, wherein the first hidden layer node may be 40, the second hidden layer node may be 40, and the output layer node may be 4. The activation function may be selected as a Sigmoid function. The neural network is updated by adopting a gradient descent method. The sum of squares of the prediction errors of the neural network is expressed as:
wherein,predicting the driving working condition type of the next working condition period as the probability of the i-th working condition cluster type for the neural network.
Example two
Referring to fig. 2, an embodiment of the present invention discloses a training method of a driving condition prediction model of a hybrid fire engine, which is used for training the neural network model according to the first embodiment; the method comprises the following steps:
201. and acquiring the travel data of the pre-collected hybrid power fire truck in the historical working condition period, and dividing the travel data of the historical working condition period into N travel data of training working condition periods.
In this embodiment, the duration of each training condition period is a preset period length.
202. And calculating corresponding running condition characteristic parameters for the running data of each training condition period, inputting the running condition characteristic parameters of each training condition period into a predetermined k-means clustering model, and obtaining the running condition type corresponding to the training condition period output by the k-means clustering model.
In this embodiment, the driving condition characteristic parameters of the training condition period include an average speed, a parking ratio, a uniform speed ratio, a deceleration ratio and a maximum speed of the hybrid fire truck in the condition period corresponding to the driving data of the training condition period;
203. in N training working condition periods N 1 The continuous training working condition time periods are target training working condition time periods, the driving working condition characteristic parameters of M continuous training working condition time periods starting from each target training working condition time period and the driving working condition types corresponding to the next training working condition time periods of the M training working condition time periods are sample data corresponding to the target training working condition time periods, and N is obtained 1 Sample data according to the N 1 Sample dataTraining the neural network model.
In this embodiment, M and N are 1 Are all smaller than N;
204. judging whether the condition of training ending is satisfied.
205. And when the judgment is satisfied, finishing the training of the neural network model, and determining the current neural network model as a pre-trained neural network model.
Alternatively, the condition for ending training may be that the error of the neural network model is smaller than a preset error, or the number of training times reaches a preset number of times.
In an alternative embodiment, said method is based on the N 1 Training the neural network model from the sample data, comprising:
for the first target training working condition period, taking driving working condition characteristic parameters and initial rewards of M continuous training working condition periods which are started by sample data corresponding to the target training working condition period as input of the neural network model, judging whether the driving working condition type corresponding to the maximum value in an output node of the neural network model is consistent with the driving working condition type corresponding to the next training working condition period of the M training working condition periods, and determining that the rewards corresponding to the second target training working condition period is 1 when the driving working condition type is consistent with the driving working condition type corresponding to the next training working condition period of the M training working condition periods, otherwise, determining that the rewards corresponding to the second target training working condition period is 0;
and for each target training working condition period except the first target training working condition period, taking the driving working condition characteristic parameters and corresponding rewards of M continuous training working condition periods which are started by sample data corresponding to the target training working condition period as the input of the neural network model, judging whether the driving working condition type corresponding to the maximum value in the output node of the neural network model is consistent with the driving working condition type corresponding to the next training working condition period of the M training working condition periods, and determining that the rewards corresponding to the next training working condition period is 1 when the driving working condition types are judged to be consistent, otherwise, determining that the rewards corresponding to the next training working condition period is 0.
In this embodiment, the initial prize value is 0. And randomly selecting parameters when the neural network model is initialized.
Optionally, when the reward value is 0, retraining the neural network model according to the sample again until the neural network model output is correct; to improve accuracy.
In yet another alternative embodiment, the k-means cluster model is predetermined by the following method:
acquiring standard working condition time period data of the hybrid power fire truck under various driving working condition types, and calculating driving working condition characteristic parameters corresponding to each standard working condition time period data; the driving working condition characteristic parameters of each standard working condition period data comprise the average speed, the parking proportion, the uniform speed proportion, the deceleration proportion and the maximum speed of the hybrid power fire truck in the working condition period corresponding to the standard working condition period data;
a plurality of values are pre-fetched for the number of clustering centers of the k-means clustering model to obtain a plurality of k-means clustering models to be determined;
inputting driving condition characteristic parameters of the standard condition period data into each k-means clustering model to be determined, and obtaining a clustering scheme output by the k-means clustering model to be determined; the clustering scheme of the k-means clustering model to be determined is a plurality of groups of standard working condition period data obtained by clustering the standard working condition period data under a plurality of driving working condition types by the k-means clustering model to be determined, wherein each group of standard working condition period data is classified into one working condition clustering type, and the number of the working condition clustering types is equal to the number of clustering centers of the k-means clustering model to be determined;
calculating a contour coefficient (SilhouetteCoefficient) of each clustering scheme, and selecting a k-means clustering model to be determined corresponding to the maximum contour coefficient as a predetermined k-means clustering model; and the number K of the output nodes of the neural network model is equal to the number of the working condition cluster types corresponding to the predetermined K-means cluster model.
The contour coefficient (also called Silhouette function) is used for evaluating the rationality of the clustering number, the range of the return value of the function is [ -1,1], and the larger the return value is, the better the classification effect is; when the return value is smaller than 0, the classification effect is not ideal; the function definition is shown in the following formula:
wherein a represents the average distance between the ith working condition period and the same class working condition period; b represents the average distance between the ith operating period and the different category of operating periods.
In this embodiment, the driving condition characteristic parameter of the standard condition period data includes an average speed, a parking ratio, a uniform speed ratio, a deceleration ratio and a maximum speed of the hybrid fire truck in a condition period corresponding to the driving data of the standard condition period data, and is specifically calculated according to the following formula:
(1) Average speed = sum of speeds of all sampling points/number of corresponding sampling points, unit km/h;
(2) Maximum speed = maximum value of sampling points during driving, unit km/h;
(3) Deceleration ratio = number of sampling points/number of sampling points in a deceleration state;
(4) The constant speed ratio = the number of sampling points/the number of sampling points in a constant speed state;
(5) Parking ratio = number of sampling points in parking state/number of sampling points.
In this embodiment, the standard working condition period data may be working condition period data conforming to the driving working condition of the type, which is simulated by the simulation software.
The following is illustrative:
the 33 representative typical conditions commonly used in the automobile simulation are selected, as shown in fig. 4, and composite conditions are constructed, which add up to 16680s, as shown in fig. 5. In this embodiment, 60s is selected as the working condition time period duration, the composite working condition data of fig. 5 is divided into segments, and one working condition time period of 0 is removed to obtain 277 working condition time period data. And then, calculating the characteristic parameters of each working condition period according to the five characteristic parameters of the average speed, the maximum speed, the deceleration proportion, the constant speed proportion and the parking proportion defined in the first part. And then clustering the working condition time periods by using a K-means clustering algorithm, and evaluating the clustered working condition types by using a Silhouette function. Through multiple calculations and evaluations, k=4 was selected in this example. And the cluster centers of the four clusters are shown in the following formula:
and analyzing two parameters of average speed and parking proportion, wherein the central coordinate of the cluster 1 is (0.0801, 25.882) and represents the working condition of the rural area. The central coordinates of cluster 2 are (0.238, 10.723), and the parking proportion of such conditions is relatively large and the vehicle speed is relatively small, which represents the urban resident conditions. The center coordinates of cluster 3 are (0.004, 59.940), the parking ratio of such conditions is small and the vehicle speed is fast, and the urban suburban conditions are indicated. The central coordinate of the cluster 4 is (0.635,2.437), and the parking proportion of the working conditions is large and the vehicle speed is small, so that the central working condition of the urban area is indicated.
And carrying out cluster analysis on the composite driving working condition time period to obtain the category of each segment as shown in fig. 6. Wherein 1 represents a city center working condition; 2 represents the residential area working condition of the urban area; 3 represents a rural working condition; 4 indicates urban suburban conditions.
Example III
The embodiment of the invention discloses a computer storage medium which stores computer instructions for executing the steps in the driving control method of the hybrid fire engine described in the first embodiment of the invention when the computer instructions are called.
Example IV
Referring to fig. 7, fig. 7 is a schematic structural diagram of a driving control device of a hybrid fire engine according to an embodiment of the present invention, and the driving control device of the hybrid fire engine may include:
a memory 401 storing executable program codes;
a processor 402 coupled with the memory 401;
the processor 402 invokes executable program codes stored in the memory 401 to perform the steps in a travel control method of a hybrid fire engine according to the first embodiment of the present invention.
The disclosure of the embodiments of the present invention is merely a preferred embodiment of the present invention, and is merely for illustrating the technical scheme of the present invention, but not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art will understand that; the technical scheme recorded in the various embodiments can be modified or part of technical features in the technical scheme can be replaced equivalently; such modifications and substitutions do not depart from the spirit and scope of the corresponding technical solutions.
Claims (9)
1. A driving control method of a hybrid power fire engine is characterized in that the method is based on a driving condition prediction model; the input of the running condition prediction model is a running condition characteristic parameter and a reward value corresponding to M working condition periods before a target working condition period of the hybrid power fire truck, and the output of the running condition characteristic parameter and the reward value is a predicted running condition type corresponding to the target working condition period of the hybrid power fire truck;
the method comprises the following steps:
judging whether the last prediction of the driving working condition prediction model is correct, and determining that the rewarding value corresponding to M working condition periods of the hybrid fire truck before the current working condition period is 1 when the last prediction of the driving working condition prediction model is correct, otherwise, determining that the rewarding value is 0;
acquiring the driving data of the hybrid power fire truck in M working condition periods before the current working condition period, calculating driving working condition characteristic parameters corresponding to each working condition period according to the driving data of the M working condition periods, and inputting the driving working condition characteristic parameters and rewarding values corresponding to the M working condition periods into the driving working condition prediction model;
acquiring the driving condition type output by the driving condition prediction model, and determining the driving condition type as the predicted driving condition type of the hybrid fire truck in the current working condition period;
and controlling the running speed of the hybrid power fire truck according to the predicted running condition type.
2. The travel control method according to claim 1, wherein the controlling the travel speed of the hybrid fire truck according to the predicted travel condition type includes:
when the predicted running condition type is a condition type that the vehicle speed is smaller than the current vehicle speed, gradually reducing the vehicle speed of the hybrid fire engine along with time so as to enable the vehicle speed to stably reach the vehicle speed corresponding to the predicted running condition type;
when the predicted running condition type is a condition type that the vehicle speed is smaller than the current vehicle speed, the vehicle speed of the hybrid fire engine is gradually increased along with time, so that the vehicle speed stably reaches the vehicle speed corresponding to the predicted running condition type.
3. The travel control method according to claim 1, characterized in that the travel condition prediction model includes a neural network model trained in advance;
the input layer of the neural network model is provided with M+1 input nodes, and the value of each input node is sequentially a driving working condition characteristic parameter and a reward value corresponding to M working condition periods before a target working condition period of the hybrid fire engine;
the output layer of the neural network model is provided with K output nodes, each output node corresponds to one working condition cluster type one by one, and the value of each output node is the probability that the neural network model predicts that the driving working condition type of the hybrid fire truck under the target working condition period belongs to the working condition cluster type corresponding to the output node.
4. The travel control method according to claim 1, wherein when the travel condition prediction model predicts correctly last time, it is determined that the reward value corresponding to M condition periods of the hybrid fire truck before the target condition period is 1, otherwise it is 0.
5. The travel control method of claim 3, wherein the travel condition prediction model further comprises a determination module; the determining module is used for obtaining values of all output nodes of the neural network model and determining that the predicted driving condition type of the hybrid fire engine corresponding to the target condition period is the condition cluster type corresponding to the output node with the largest value.
6. A method for training a driving condition prediction model of a hybrid fire engine, characterized in that the method is used for training the neural network model according to claim 3 or 5; the method comprises the following steps:
acquiring the travel data of a pre-collected historical working condition period of the hybrid power fire truck, dividing the travel data of the historical working condition period into N travel data of training working condition periods, wherein the duration of each training working condition period is the length of a preset period;
calculating corresponding running condition characteristic parameters for the running data of each training condition period, inputting the running condition characteristic parameters of each training condition period into a predetermined k-means clustering model, and obtaining the running condition type corresponding to the training condition period output by the k-means clustering model; the driving working condition characteristic parameters of the training working condition period comprise the average speed, the parking proportion, the uniform speed proportion, the deceleration proportion and the maximum speed of the hybrid power fire truck under the working condition period corresponding to the driving data of the training working condition period;
in N training working condition periods N 1 The continuous training working condition time periods are target training working condition time periods, the driving working condition characteristic parameters of M continuous training working condition time periods starting from each target training working condition time period and the driving working condition types corresponding to the next training working condition time periods of the M training working condition time periods are sample data corresponding to the target training working condition time periods, and N is obtained 1 Sample data according to the N 1 Training the neural network model with the sample data; said M and said N 1 Are all smaller than N;
judging whether the training ending condition is met or not;
and when the judgment is satisfied, finishing the training of the neural network model, and determining the current neural network model as a pre-trained neural network model.
7. The training method of claim 6, wherein the training method is based on the N 1 Training the neural network model from the sample data, comprising:
for the first target training working condition period, taking driving working condition characteristic parameters and initial rewards of M continuous training working condition periods which are started by sample data corresponding to the target training working condition period as input of the neural network model, judging whether the driving working condition type corresponding to the maximum value in an output node of the neural network model is consistent with the driving working condition type corresponding to the next training working condition period of the M training working condition periods, and determining that the rewards corresponding to the second target training working condition period is 1 when the driving working condition type is consistent with the driving working condition type corresponding to the next training working condition period of the M training working condition periods, otherwise, determining that the rewards corresponding to the second target training working condition period is 0;
and for each target training working condition period except the first target training working condition period, taking the driving working condition characteristic parameters and corresponding rewards of M continuous training working condition periods which are started by sample data corresponding to the target training working condition period as the input of the neural network model, judging whether the driving working condition type corresponding to the maximum value in the output node of the neural network model is consistent with the driving working condition type corresponding to the next training working condition period of the M training working condition periods, and determining that the rewards corresponding to the next training working condition period is 1 when the driving working condition types are judged to be consistent, otherwise, determining that the rewards corresponding to the next training working condition period is 0.
8. The training method of claim 6, wherein the k-means clustering model is predetermined by the method of:
acquiring standard working condition time period data of the hybrid power fire truck under various driving working condition types, and calculating driving working condition characteristic parameters corresponding to each standard working condition time period data; the driving working condition characteristic parameters of each standard working condition period data comprise the average speed, the parking proportion, the uniform speed proportion, the deceleration proportion and the maximum speed of the hybrid power fire truck in the working condition period corresponding to the standard working condition period data;
a plurality of values are pre-fetched for the number of clustering centers of the k-means clustering model to obtain a plurality of k-means clustering models to be determined;
inputting driving condition characteristic parameters of the standard condition period data into each k-means clustering model to be determined, and obtaining a clustering scheme output by the k-means clustering model to be determined; the clustering scheme of the k-means clustering model to be determined is a plurality of groups of standard working condition period data obtained by clustering the standard working condition period data under a plurality of driving working condition types by the k-means clustering model to be determined, wherein each group of standard working condition period data is classified into one working condition clustering type, and the number of the working condition clustering types is equal to the number of clustering centers of the k-means clustering model to be determined;
calculating the contour coefficient of each clustering scheme, and selecting a k-means clustering model to be determined corresponding to the maximum contour coefficient as a predetermined k-means clustering model; and the number K of the output nodes of the neural network model is equal to the number of the working condition cluster types corresponding to the predetermined K-means cluster model.
9. A travel control device for a hybrid fire engine, comprising:
a memory storing executable program code;
a processor coupled to the memory;
the processor invokes the executable program code stored in the memory to perform a travel control method of a hybrid fire engine as claimed in any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311149831.0A CN117184032A (en) | 2023-09-06 | 2023-09-06 | Driving control method and device of hybrid power fire engine and training method of driving condition prediction model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311149831.0A CN117184032A (en) | 2023-09-06 | 2023-09-06 | Driving control method and device of hybrid power fire engine and training method of driving condition prediction model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117184032A true CN117184032A (en) | 2023-12-08 |
Family
ID=89002912
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311149831.0A Pending CN117184032A (en) | 2023-09-06 | 2023-09-06 | Driving control method and device of hybrid power fire engine and training method of driving condition prediction model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117184032A (en) |
-
2023
- 2023-09-06 CN CN202311149831.0A patent/CN117184032A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110775065B (en) | Hybrid electric vehicle battery life prediction method based on working condition recognition | |
| CN110610260B (en) | Driving energy consumption prediction system, method, storage medium and equipment | |
| CN107169567B (en) | Method and device for generating decision network model for automatic vehicle driving | |
| CN111267830B (en) | Hybrid power bus energy management method, device and storage medium | |
| US8706416B2 (en) | System and method for determining a vehicle route | |
| WO2020056157A1 (en) | Systems and methods for managing energy storage systems | |
| CN111845701A (en) | HEV energy management method based on deep reinforcement learning in car following environment | |
| US20230139003A1 (en) | Systems and methods for managing velocity profiles | |
| CN112668799A (en) | Intelligent energy management method and storage medium for PHEV (Power electric vehicle) based on big driving data | |
| CN110910531B (en) | Rapid pavement friction coefficient detection method based on vehicle-mounted OBD information | |
| CN112026744B (en) | Series-parallel hybrid power system energy management method based on DQN variants | |
| CN112339756B (en) | New energy automobile traffic light intersection energy recovery optimization speed planning algorithm based on reinforcement learning | |
| JP2022089806A (en) | Modeling operation profiles of vehicle | |
| CN114074680B (en) | Vehicle channel change behavior decision method and system based on deep reinforcement learning | |
| CN113657432A (en) | Commercial vehicle driving behavior risk level identification method based on Internet of vehicles data | |
| CN115422747A (en) | Method and device for calculating discharge amount of pollutants in tail gas of motor vehicle | |
| CN116187161A (en) | Intelligent energy management method and system for hybrid electric bus in intelligent networking environment | |
| Montazeri-Gh et al. | Driving condition recognition for genetic-fuzzy HEV control | |
| CN115759462A (en) | Charging behavior prediction method and device for electric vehicle user and electronic equipment | |
| CN113415288A (en) | Sectional type longitudinal vehicle speed planning method, device, equipment and storage medium | |
| CN115107733A (en) | Energy management method and system for hybrid electric vehicle | |
| CN113246958A (en) | TD 3-based multi-target HEV energy management method and system | |
| CN117184032A (en) | Driving control method and device of hybrid power fire engine and training method of driving condition prediction model | |
| CN116538286A (en) | Commercial vehicle gear shifting system and method considering NVH characteristics | |
| CN114278727B (en) | Vehicle automatic gear shifting control method, system, medium and equipment under limit working condition |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |