Methods, apparatus, systems, devices, and media for controlling shared vehicles

Abstract

This disclosure provides a method, apparatus, system, device, and medium for controlling shared vehicles. The method for controlling shared vehicles includes: acquiring riding parameters of the shared vehicle, environmental parameters of the environment in which the shared vehicle is located, and passenger information of the passengers carried in the shared vehicle; based on the riding parameters, environmental parameters, and passenger information, predicting the state of the passengers carried in the shared vehicle using a predictive model to obtain a passenger state prediction result; and controlling the shared vehicle based on the passenger state prediction result. The embodiments of this disclosure can ensure passenger comfort.

Description

Technical Field

[0001] This disclosure relates to the field of shared vehicle technology, and in particular to a method, apparatus, system, device, and medium for controlling shared vehicles. Background Technology

[0002] Currently, an increasing number of shared vehicles are being deployed in the market, greatly facilitating users' travel. Shared vehicles can include, but are not limited to, shared bicycles and shared electric vehicles.

[0003] Some shared vehicles (such as shared family cars) are equipped with passenger seats for passengers to sit in. How to ensure the comfort of passengers is a concern for those skilled in the art. Summary of the Invention

[0004] To address the aforementioned technical problems, this disclosure provides a method, apparatus, system, device, and medium for controlling shared vehicles.

[0005] According to one aspect of the present disclosure, a method for controlling shared vehicles is provided, comprising:

[0006] Obtain the riding parameters of the shared vehicle, the environmental parameters of the environment in which the shared vehicle is located, and the passenger information of the passengers carried by the shared vehicle;

[0007] Based on the riding parameters, the environmental parameters, and the occupant information, the state of the occupants in the shared vehicle is predicted by a prediction model to obtain the occupant state prediction result.

[0008] The shared vehicle is controlled based on the occupant status prediction results.

[0009] According to another aspect of the present disclosure, an apparatus for controlling a shared vehicle is provided, comprising:

[0010] The first acquisition module is used to acquire the riding parameters of the shared vehicle, the environmental parameters of the environment in which the shared vehicle is located, and the passenger information of the passengers carried by the shared vehicle.

[0011] The prediction module is used to predict the state of the passengers in the shared vehicle based on the riding parameters, the environmental parameters and the passenger information, through a prediction model, and obtain the passenger state prediction result.

[0012] The control module is used to control the shared vehicle based on the occupant status prediction results.

[0013] According to another aspect of the present disclosure, a system for controlling shared vehicles is provided, comprising:

[0014] The shared vehicle is equipped with sensors for collecting riding parameters of the shared vehicle.

[0015] The server is used to obtain environmental parameters of the environment in which the shared vehicle is located and passenger information of the passengers in the shared vehicle. Based on the riding parameters, the environmental parameters and the passenger information, the server performs state prediction on the passengers in the shared vehicle through a prediction model to obtain passenger state prediction results, and controls the shared vehicle based on the passenger state prediction results.

[0016] According to another aspect of the present disclosure, an electronic device is provided, the electronic device comprising:

[0017] processor;

[0018] Memory used to store the processor's executable instructions;

[0019] The processor is configured to read the executable instructions from the memory and execute the instructions to implement the method described above for controlling shared vehicles.

[0020] According to another aspect of the present disclosure, a computer-readable storage medium is provided, the storage medium storing a computer program for performing the above-described method for controlling shared vehicles.

[0021] According to another aspect of the present disclosure, a computer program product is provided, including computer program instructions that, when executed by a processor, implement the above-described method for controlling shared vehicles.

[0022] Based on the methods, apparatus, systems, equipment media, and products for controlling shared vehicles provided in the above embodiments of this disclosure, during the riding of a shared vehicle, the state of the passengers in the shared vehicle can be predicted by a prediction model based on the riding parameters of the shared vehicle, the environmental parameters of the environment in which the shared vehicle is located, and the passenger information of the passengers in the shared vehicle, thereby obtaining the passenger state prediction result. The prediction model can be a neural network model trained with a large amount of sample data, and the accuracy and reliability of the passenger state prediction result obtained by the prediction model can be well guaranteed. In this way, based on the passenger state prediction result, it is possible to determine more accurately whether the passengers in the shared vehicle are feeling uncomfortable, so as to adaptively control the shared vehicle and help ensure the riding comfort of the passengers. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the structure of a system for controlling shared vehicles provided by some exemplary embodiments of this disclosure.

[0024] Figure 2 This is a flowchart illustrating a method for controlling shared vehicles provided by some exemplary embodiments of this disclosure.

[0025] Figure 3 This is a flowchart illustrating a method for controlling a shared vehicle based on occupant status prediction results, provided by some exemplary embodiments of this disclosure.

[0026] Figure 4 This is a flowchart illustrating a method for determining a target mood type based on mood prediction values ​​and / or vital sign prediction values, provided by some exemplary embodiments of this disclosure.

[0027] Figure 5 This is a flowchart illustrating a method for controlling a shared vehicle based on occupant status prediction results, provided by some other exemplary embodiments of this disclosure.

[0028] Figure 6 This is a flowchart illustrating a model training method provided by some exemplary embodiments of this disclosure.

[0029] Figure 7 This is a flowchart illustrating a method for obtaining an incremental training dataset provided by some exemplary embodiments of this disclosure.

[0030] Figure 8-1 This is an architecture diagram of a first type of prediction model in some exemplary embodiments of this disclosure.

[0031] Figure 8-2 This is a flowchart illustrating a method for controlling shared vehicles provided by other exemplary embodiments of this disclosure.

[0032] Figure 8-3 This is a flowchart illustrating a model training method provided by some other exemplary embodiments of this disclosure.

[0033] Figure 9 This is a schematic diagram of the structure of a device for controlling shared vehicles provided by some exemplary embodiments of this disclosure.

[0034] Figure 10 This is a schematic diagram of the structure of the control module in some exemplary embodiments of this disclosure.

[0035] Figure 11 This is a schematic diagram of the structure of the control module in some other exemplary embodiments of this disclosure.

[0036] Figure 12 This is a schematic diagram of a module used to assist model training in some exemplary embodiments of this disclosure.

[0037] Figure 13 This is a schematic diagram of the structure of the third acquisition module in some exemplary embodiments of this disclosure.

[0038] Figure 14 This is a schematic diagram of the structure of a device for controlling shared vehicles provided in some other exemplary embodiments of this disclosure.

[0039] Figure 15 This is a schematic diagram of the structure of an electronic device provided by some exemplary embodiments of this disclosure. Detailed Implementation

[0040] To explain this disclosure, exemplary embodiments of the disclosure will now be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the disclosure, and not all of them. It should be understood that the disclosure is not limited to exemplary embodiments.

[0041] It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments do not limit the scope of this disclosure.

[0042] Application Overview

[0043] like Figure 1 As shown, the system architecture to which some exemplary embodiments of this disclosure are applicable may include a shared vehicle 110, a mobile terminal 120, and a server 130.

[0044] Optionally, the shared vehicle 110 may be equipped with a passenger seat. For example, the shared vehicle 110 may be a shared family bike, in which the passenger can sit in the passenger seat while the user is riding the shared vehicle 110, that is, the shared vehicle 110 can carry passengers.

[0045] Optionally, the mobile terminal 120 may include, but is not limited to, mobile phones, tablets, etc., used by the cyclist. The mobile terminal 120 may have an application provided by the operator of the shared vehicle 110 installed on it. By operating on the interface of the application, users can initiate requests to use or return the shared vehicle 110.

[0046] Optionally, server 130 can be the business server of the operator of shared vehicle 110. Server 130 can communicate remotely with shared vehicle 110. Server 130 can also communicate remotely with mobile terminal 120. For example, server 130 can receive vehicle usage requests, vehicle return requests, etc. from mobile terminal 120.

[0047] In the embodiments of this disclosure, during the riding of the shared vehicle 110, the server 130 can predict the status of the passengers carried in the shared vehicle 110 and obtain the passenger status prediction result, so as to control the shared vehicle based on the passenger status prediction result, thereby ensuring the riding comfort of the passengers.

[0048] Exemplary methods

[0049] Figure 2 This is a flowchart illustrating a method for controlling shared vehicles provided by some exemplary embodiments of this disclosure. Figure 2 The method shown may include steps 210, 220 and 230. Figure 2 The method shown can be applied to electronic devices, such as servers, which can be... Figure 1 The server shown is 130.

[0050] Step 210: Obtain the riding parameters of the shared vehicle, the environmental parameters of the environment where the shared vehicle is located, and the passenger information of the passengers in the shared vehicle.

[0051] Optionally, the riding parameters of the shared vehicle may include at least one of the following: riding speed, riding acceleration, sway amplitude, and riding gradient.

[0052] It should be noted that shared vehicles can be equipped with sensors. Based on the data collected by these sensors, riding parameters can be determined. For example, shared vehicles can be equipped with speed sensors to collect riding speed. As another example, shared vehicles can be equipped with positioning sensors; based on the data collected by these sensors, the riding speed can be determined. Positioning sensors can include, but are not limited to, sensors based on the Global Navigation Satellite System (GNSS) or the Global Positioning System (GPS). For yet another example, shared vehicles can be equipped with accelerometer sensors to collect riding acceleration. Furthermore, shared vehicles can be equipped with inertial measurement units (IMUs); based on the data collected by the IMUs, the sway amplitude and riding gradient can be determined.

[0053] Optionally, the environmental parameters of the environment in which the shared vehicle is located may include at least one of the following: ambient temperature, ambient humidity, and ambient wind speed.

[0054] It should be noted that the current location of the shared vehicle can be determined based on the data collected by the positioning sensor. Next, the application programming interface (API) used to access the weather website can be called to obtain the environmental parameters of the shared vehicle's current location, such as the ambient temperature, humidity, and wind speed.

[0055] Optionally, the mobile device used by the cyclist can have an application provided by the shared vehicle operator installed. Before using the shared vehicle, the cyclist can operate the application interface to input the passenger information for the passengers to be carried in the shared vehicle. The passenger information can be included in the ride request initiated by the cyclist through the mobile device, thus allowing the passenger information to be extracted from the ride request. Passenger information may include, but is not limited to, age range and age.

[0056] Step 220: Based on riding parameters, environmental parameters and occupant information, the state of the occupants in the shared vehicle is predicted by a prediction model to obtain the occupant state prediction result.

[0057] Alternatively, the prediction model can be a neural network model for occupant status prediction that has been pre-trained using a large amount of sample data. As an example, the prediction model can be a Long Short-Term Memory (LSTM) neural network model.

[0058] Optionally, riding parameters, environmental parameters, and occupant information can be used as input data for the prediction model. The prediction model can then perform calculations to predict the state of the occupants in the shared vehicle, thereby obtaining the occupant state prediction results. The occupant state prediction results can include state prediction results corresponding to multiple dimensions. These multiple dimensions may include, but are not limited to, mood dimensions, vital sign dimensions, etc.

[0059] As an example, the state prediction results corresponding to the mood dimension can be in numerical form, with a larger value indicating a better mood and a smaller value indicating a worse mood. Alternatively, the state prediction results corresponding to the mood dimension can be in the form of a ranking system, with a higher ranking indicating a better mood and a lower ranking indicating a worse mood.

[0060] As an example, the vital signs dimension can be further subdivided into heart rate, blood pressure, and body surface temperature dimensions. Similar to the state prediction results corresponding to the mood dimension, the state prediction results corresponding to the vital signs dimension can be in numerical form or in a ranking form.

[0061] Step 230: Control the shared vehicle based on the occupant status prediction results.

[0062] Optionally, the occupant status prediction results can be analyzed to determine whether the occupants of the shared vehicle are feeling unwell. Based on the determination results, the shared vehicle can be controlled. For example, if the occupants of the shared vehicle are feeling unwell, the vehicle can be controlled to alert the rider, allowing the rider to take measures to alleviate the occupants' discomfort. Conversely, if the occupants are not feeling unwell, the vehicle can be controlled to prompt the rider to maintain current riding parameters, ensuring the occupants remain in a more comfortable state.

[0063] In the embodiments of this disclosure, during the riding of a shared vehicle, the rider's state can be predicted using a prediction model based on the shared vehicle's riding parameters, environmental parameters of the shared vehicle's environment, and rider information. The prediction model can be a neural network model trained with a large amount of sample data, ensuring the accuracy and reliability of the rider state prediction results. Thus, based on the rider state prediction results, it is possible to accurately determine whether the rider is experiencing discomfort, allowing for adaptive control of the shared vehicle and improving rider comfort.

[0064] Figure 3 This is a flowchart illustrating a method for controlling a shared vehicle based on occupant status prediction results, provided by some exemplary embodiments of this disclosure. Figure 3 The method shown may include steps 310, 320 and 330.

[0065] Step 310: Based on the occupant status prediction results, determine the mood prediction value and / or vital sign prediction value.

[0066] Optionally, the occupant state prediction results may include state prediction results corresponding to the mood dimension. These mood prediction results can be in numerical form and can be used as mood prediction values. Similarly, the occupant state prediction results may include state prediction results corresponding to the vital signs dimension. These vital signs prediction results can be in numerical form and can be used as vital signs prediction values.

[0067] Step 320: Determine the target mood type based on the mood prediction value and / or physical sign prediction value.

[0068] Optionally, the target mood type can be determined based solely on mood prediction values ​​or solely on physical characteristic prediction values. Alternatively, the target mood type can be determined by combining mood prediction values ​​and physical characteristic prediction values. The target mood type can have at least two possible outcomes, such as a positive or negative type. If the target mood type is negative, it indicates that the passenger in the shared vehicle is in a bad mood, and therefore it can be determined that the passenger is feeling unwell. If the target mood type is positive, it indicates that the passenger in the shared vehicle is in a good mood, and therefore it can be determined that the passenger is not feeling unwell.

[0069] In some alternative embodiments of this disclosure, such as Figure 4 As shown, step 320 may include steps 410, 420, 430 and 440.

[0070] Step 410: Determine the first distribution information of the mood prediction value relative to the preset numerical range corresponding to the mood dimension.

[0071] Optionally, a preset numerical range can be pre-defined for the mood dimension. The first distribution information of the mood prediction value relative to the preset numerical range corresponding to the mood dimension can be used to characterize whether the mood prediction value falls within the preset numerical range corresponding to the mood dimension.

[0072] Step 420: Determine the mood type that matches the first distribution information.

[0073] Optionally, if the first distribution information is within the preset numerical range corresponding to the mood dimension, it indicates that the predicted mood value is within a reasonable range. In this case, the mood type that matches the first distribution information can be determined to be a positive type. If the first distribution information is outside the preset numerical range corresponding to the mood dimension, it indicates that the predicted mood value exceeds a reasonable range. In this case, the mood type that matches the first distribution information can be determined to be a negative type.

[0074] Step 430: In response to the fact that the mood type that matches the first distribution information is a positive type, the mood type that matches the first distribution information is taken as the target mood type.

[0075] Step 440: In response to the fact that the mood type that matches the first distribution information is a negative type, determine the second distribution information of the predicted value of the vital signs relative to the preset numerical range of the vital signs dimension, determine the mood type that matches the second distribution information, and take the mood type that matches the second distribution information as the target mood type.

[0076] If the mood type that matches the first distribution information is a positive type, the mood type that matches the first distribution information can be directly used as the target mood type. That is, the target mood type is specifically a positive type.

[0077] If the mood type that matches the first distribution information is negative, then a second distribution information can be determined relative to the preset numerical range corresponding to the vital sign dimension for the predicted value. The second distribution information can be used to characterize whether the predicted vital sign value falls within the preset numerical range corresponding to the vital sign dimension. If the second distribution information characterizes the predicted vital sign value as falling within the preset numerical range corresponding to the vital sign dimension, this indicates that the predicted vital sign value is within a reasonable range, and therefore, the mood type that matches the second distribution information can be determined to be positive. If the second distribution information characterizes the predicted vital sign value as falling outside the preset numerical range corresponding to the vital sign dimension, this indicates that the predicted vital sign value exceeds a reasonable range, and therefore, the mood type that matches the second distribution information can be determined to be negative. The mood type that matches the second distribution information can be used as the target mood type.

[0078] Figure 4 In the illustrated implementation, if the predicted mood value is within a reasonable range, a positive type can be used as the target mood type. If the predicted mood value exceeds the reasonable range, but the predicted physical signs value is within the reasonable range, a positive type can still be used as the target mood type. If both the predicted mood value and the predicted physical signs value exceed the reasonable range, a negative type can be used as the target mood type. It should be noted that if the predicted mood value exceeds the reasonable range, it can be assumed that the passenger in the shared vehicle may be feeling unwell. Therefore, further assessment based on the predicted physical signs value can accurately identify the passenger's discomfort, helping to avoid misidentification.

[0079] Of course, the implementation of step 320 is not limited to this. For example, the mood type that matches the first distribution information can be directly used as the target mood type. Or, for example, the mood type that matches the second distribution information can be directly used as the target mood type.

[0080] Step 330: In response to the target mood type being negative, control the shared vehicle to output a prompt message to prompt the rider of the shared vehicle to adjust the riding status.

[0081] Optionally, a buzzer can be installed on the shared vehicle. If the target mood type is negative, the buzzer can be controlled to emit an audible alert to prompt the rider to adjust their riding state, thereby alleviating the discomfort of the passenger. Methods for adjusting riding state may include, but are not limited to, adjusting riding speed, adjusting riding acceleration, adjusting riding route, and stopping to rest.

[0082] The above paragraph describes the situation of outputting prompts in the form of sound. In actual implementation, prompts can also be output in the form of light, vibration, etc. This disclosure does not limit the output form of prompts.

[0083] In the embodiments of this disclosure, based on the occupant status prediction results, mood prediction values ​​and / or vital sign prediction values ​​can be determined efficiently and reliably, providing a valid reference for determining the target mood type. When the target mood type is negative, prompts can be sent to the shared vehicle rider to adjust their riding status, thereby alleviating discomfort and ensuring rider comfort.

[0084] Figure 5 This is a flowchart illustrating a method for controlling a shared vehicle based on occupant status prediction results, provided by some other exemplary embodiments of this disclosure. Figure 5 The method shown may include steps 510, 520 and 530.

[0085] Step 510: Determine the target mood type based on the occupant state prediction results.

[0086] Optionally, the specific implementation of step 510 refers to the above description. Figure 3 The illustrated embodiments and Figure 4 The relevant descriptions of the embodiments shown are sufficient and will not be repeated here.

[0087] Step 520: In response to the target mood type being negative, determine the reason information for the target mood type being negative based on cycling parameters and environmental parameters.

[0088] Step 530: Control the shared vehicle using a control method adapted to the cause information.

[0089] Taking cycling parameters including cycling speed, sway amplitude, and cycling gradient, and environmental parameters including ambient humidity and ambient wind force as an example, reasonable speed range, reasonable amplitude range, reasonable gradient range, reasonable humidity range, and reasonable wind force range can be preset in advance.

[0090] If the target mood type is negative, parameters exceeding a pre-defined reasonable range can be selected from cycling speed, sway amplitude, cycling gradient, ambient humidity, and ambient wind force. Assuming only sway amplitude and ambient wind force exceed the pre-defined reasonable range—that is, sway amplitude exceeds a pre-defined reasonable amplitude range, and ambient wind force exceeds a pre-defined reasonable wind force range—a first value representing the degree of deviation of the sway amplitude from the pre-defined reasonable amplitude range and a second value representing the degree of deviation of the ambient wind force from the pre-defined reasonable wind force range can be determined, and the first and second values ​​can be compared.

[0091] If the first value is greater than the second value, the cause information can be used to characterize the negative mood type as being caused by excessive swaying of the shared vehicle. In this case, the control method adapted to the cause information could be to output a prompt message to remind the shared vehicle rider to improve their riding style or change their riding route so that the shared vehicle can travel smoothly.

[0092] If the first value is less than the second value, the cause information can be used to characterize the target mood type as negative, with the cause being excessive wind. In this case, the control method adapted to the cause information could be: outputting a prompt message to remind shared vehicle riders to ride in sheltered areas as much as possible.

[0093] In some embodiments, if only the swaying amplitude and ambient wind force exceed the corresponding reasonable range set in advance, the cause information can also be directly determined. The cause information can be used to characterize that the cause of the target mood type being negative is that the swaying amplitude of the shared vehicle is too large and the ambient wind force is too strong.

[0094] In some embodiments, the cause information may also be used to characterize the negative mood type as a result of the shared vehicle's excessive riding speed. In this case, the control method adapted to the cause information could be: controlling the shared vehicle's power system to limit the riding speed within a certain range, thus preventing the shared vehicle from riding too fast.

[0095] In the embodiments of this disclosure, when passengers in a shared vehicle feel uncomfortable, the specific cause of the discomfort can be determined by combining riding parameters and environmental parameters, and an appropriate control method can be adopted to control the shared vehicle, which helps to ensure the comfort of the passengers.

[0096] Figure 6 This is a flowchart illustrating a model training method provided by some exemplary embodiments of this disclosure. Figure 6 The method shown may include steps 610, 620, 630 and 640.

[0097] Step 610: Obtain the historical training dataset for training the prediction model.

[0098] Step 620: Obtain the incremental training dataset.

[0099] Optionally, the prediction model can be a neural network model for predicting occupant status, pre-trained using a large amount of sample data. This large set of sample data can be called the historical training dataset. For ease of understanding, the end time of training the model using the historical training dataset can be referred to as the target time. The incremental training dataset can refer to the set of new training data generated after the target time. For example, after the target time, a shared vehicle operator can deploy some shared vehicles to different cities and collect new training data using these newly deployed vehicles to form the incremental training dataset.

[0100] It should be noted that each training data point in the historical training dataset and each training data point in the incremental training dataset can include: training riding parameters, training environment parameters, training rider information, and rider status labels. The composition of training riding parameters, training environment parameters, training rider information, and rider status labels can be referred to the relevant introductions above regarding the composition of riding parameters, environment parameters, rider information, and rider status prediction results, and will not be repeated here.

[0101] Step 630: Integrate the historical training dataset and the incremental training dataset to obtain the target training dataset.

[0102] Optionally, the historical training dataset and the incremental training dataset can be directly merged, and the resulting merged training dataset can be used as the target training dataset. Alternatively, after obtaining the merged training dataset by merging the historical training dataset and the incremental training dataset, the merged training dataset can be processed by removing outlier data, normalizing (e.g., mapping all values ​​to the interval [0, 1]), and the processed merged training dataset can be used as the target training dataset.

[0103] Step 640: Train the prediction model using the target training dataset.

[0104] Optionally, for each training data point in the target training dataset, the training riding parameters, training environment parameters, and training occupant information included in the training data can be provided as input data to the prediction model. The prediction model can then calculate and obtain the corresponding training occupant state prediction result. The training occupant state prediction result corresponding to each training data point in the target training dataset can be considered as prediction data, and the occupant state label included in each training data point in the target training dataset can be considered as ground truth data. Based on the prediction data, ground truth data, and a preset loss function, the model loss value of the prediction model can be determined. The preset loss function may include, but is not limited to, the mean squared error loss function, the mean absolute error loss function, and the mean squared logarithmic error loss function. Based on the model loss value, the prediction model can be trained using methods such as stochastic gradient descent and steepest gradient descent, or an adaptive moment estimation optimizer can be used to train the prediction model. This training can be considered as retraining (because the prediction model itself has already been trained using the historical training dataset).

[0105] Optionally, the target training dataset can be divided into a training set, a validation set, and a test set according to a preset ratio, and the prediction model can be trained based on this. As an example, the preset ratio can be 8:1:1 or 6:2:2.

[0106] In the embodiments of this disclosure, historical training datasets and incremental training datasets can be integrated to obtain a target training dataset for training the prediction model. This allows for retraining of the prediction model even with newly generated training data, facilitating further optimization of the model parameters and improving its prediction accuracy.

[0107] Figure 7 This is a flowchart illustrating a method for obtaining an incremental training dataset provided by some exemplary embodiments of this disclosure. Figure 7 The method shown may include steps 710, 720, 730, 740 and 750.

[0108] Step 710: In response to the end of the shared vehicle ride, determine the passenger status statistics during the shared vehicle ride based on the passenger status prediction results.

[0109] Optionally, during the ride of the shared vehicle, timed or untimed actions can be performed. Figure 2The method shown allows for the generation of multiple occupant state predictions, each corresponding to a specific moment. Based on these predictions, occupant state statistics during shared vehicle riding can be determined. These statistics may include the number of times discomfort was felt during the ride, the specific moments of discomfort, etc. In some embodiments, the statistics may further include information on the causes of each specific moment of discomfort.

[0110] Step 720: Send passenger status statistics information to the mobile terminal bound to the shared vehicle.

[0111] Optionally, the mobile terminal linked to the shared vehicle can be the mobile terminal used by the rider of the shared vehicle. The mobile terminal used by the rider can have an application provided by the shared vehicle operator installed. Rider status statistics can be carried in application messages and pushed to the rider's mobile terminal. The rider can view the rider status statistics through the application's interface.

[0112] Step 730: Receive objection information uploaded by the mobile terminal regarding the occupant status statistics information.

[0113] Optionally, riders can determine if there are errors in the occupant status statistics. If errors are found, riders of the shared vehicle can use the mobile application interface to upload objections to the server regarding the occupant status statistics. These objections can indicate which information in the occupant status statistics contains errors.

[0114] Step 740: Based on the objection information, determine the actual occupant status during the shared vehicle ride.

[0115] Optionally, based on the objection information, errors in the passenger status statistics can be corrected, thereby obtaining the true passenger status during the shared vehicle ride.

[0116] In some embodiments, the back-end staff of the shared vehicle operator may also manually verify the passenger status statistics by referring to the objection information in order to determine the true passenger status.

[0117] Step 750: Determine the incremental training dataset based on riding parameters, environmental parameters, rider information, and actual rider status.

[0118] Assuming that occupant state statistics are determined based on multiple occupant state predictions corresponding to multiple time points, then by executing step 740, multiple real occupant states corresponding to multiple time points can be obtained. Thus, for each of these multiple time points, the riding parameters corresponding to that time point can be used as training riding parameters, the environmental parameters corresponding to that time point as training environmental parameters, the occupant information corresponding to that time point as training occupant information, and the real occupant state corresponding to that time point as the occupant state label, thereby obtaining a new training data point corresponding to that time point. In this way, multiple new training data points corresponding to multiple time points can be obtained to form an incremental training dataset.

[0119] In some embodiments, after obtaining multiple new training data corresponding to multiple time points, new training data that duplicates the training data in the historical training dataset can be deleted from the multiple new training data, and the remaining new training data can be used to form an incremental training dataset.

[0120] In the embodiments of this disclosure, after a shared vehicle ride ends, passenger status statistics can be sent to the mobile terminal bound to the shared vehicle. If errors exist in the passenger status statistics, the shared vehicle rider can raise objections. Based on the objections, the actual passenger status during the shared vehicle ride can be determined, and this information, along with riding parameters, environmental parameters, and passenger information, can be used to determine the incremental training dataset. In this way, new training data, corrected by the rider and meeting the accuracy requirements, can be used to retrain the prediction model, facilitating further optimization of the model parameters and improving the prediction accuracy.

[0121] In some alternative examples, the methods provided by embodiments of this disclosure may further include:

[0122] The first type of prediction model is trained periodically within a preset time period;

[0123] After a preset time period, in response to the prediction accuracy of the second type of prediction model being less than or equal to that of the first type of prediction model, the first type of prediction model and the second type of prediction model are periodically trained.

[0124] After a preset time period, in response to the fact that the prediction accuracy of the second type of prediction model is greater than that of the first type of prediction model, the second type of prediction model is periodically trained, and the training of the first type of prediction model is stopped.

[0125] Among them, the second type of prediction model has a higher model complexity than the first type of prediction model; within a preset time period, the occupant state prediction result is generated by the first type of prediction model; after the preset time period, in response to the fact that the training of the first type of prediction model has not stopped, the occupant state prediction result is generated by the first type of prediction model; after the preset time period, in response to the fact that the training of the first type of prediction model has stopped, the occupant state prediction result is generated by the second type of prediction model.

[0126] Optionally, the first type of prediction model can be a neural network model that includes several gated recurrent units (GRUs). For example, such as... Figure 8-1 As shown, the first type of prediction model can be, for example, a neural network model comprising three GRU layers. The first GRU layer may contain, for example, 128 neurons. The second GRU layer may contain, for example, 128 neurons. The third GRU layer may contain, for example, 64 neurons. Optionally, the neuron dropout rate between adjacent layers may be 0.2 to 0.3. Figure 8-1 The method involves two fully connected layers. The first fully connected layer can be initialized with random weights and an activation function can be used to solve the gradient vanishing problem. The second fully connected layer can serve as the output layer. Alternatively, the second type of prediction model can be an LSTM neural network model. Of course, the first and second types of prediction models are not limited to these, as long as the model complexity of the second type of prediction model is higher than that of the first type. This disclosure does not impose any limitations on this.

[0127] Optionally, the start time of the preset time period can be the moment when the shared vehicle operator first deploys the shared vehicles, and the duration of the preset time period can be a preset duration. The preset duration can be, for example, six months, one year, two years, etc. The following explanation will use the case of a preset duration of one year as an example.

[0128] In the embodiments of this disclosure, after the shared vehicle operator first deploys shared vehicles at a certain time (hereinafter referred to as the initial time), for one year after the initial time, a first type of prediction model can be continuously used to predict the state of the occupants in the shared vehicles to obtain occupant state prediction results. Furthermore, the first type of prediction model can be trained periodically, for example, every one or two months, to ensure the prediction accuracy of the first type of prediction model. After another year following the initial time, a transition phase can be considered to have begun. During the transition phase, not only can the first type of prediction model be trained periodically, but a second type of prediction model can also be trained periodically. Since the model complexity of the second type of prediction model is higher than that of the first type of prediction model, it requires a relatively longer time to train the second type of prediction model to a better state. Therefore, for a period of time after entering the transition phase, the first type of prediction model can be used first to predict the state of the occupants in the shared vehicles. When the training time of the second type of prediction model is long enough, its prediction accuracy will surpass that of the first type. At this point, it is possible to switch to using the second type of prediction model for occupant status prediction. Subsequent training only requires periodic training of the second type of prediction model, without needing to train the first type again. Optionally, each training iteration of both the first and second type of prediction models can use the latest obtained target training dataset.

[0129] By employing the embodiments of this disclosure, a prediction model with high prediction accuracy can be consistently used for occupant status prediction, which helps to ensure the accuracy and reliability of the occupant status prediction results.

[0130] In some optional examples, such as Figure 8-2 As shown, during the ride of a shared bicycle, the server can collect parameters such as the bicycle's acceleration, sway amplitude, and gradient at regular intervals. The server can also obtain parameters such as ambient humidity and wind speed at the bicycle's current location from weather websites. By inputting these parameters and the passenger information into the prediction model, the server can predict the passenger's heart rate, blood pressure, body temperature, and mood. If the predicted mood is within a reasonable range, the passenger's mood type is determined to be positive. If the predicted mood exceeds a reasonable range, the system can accurately identify a negative mood type based on whether the predicted heart rate, blood pressure, and body temperature are within their respective reasonable ranges. If the passenger's mood is negative, a buzzer can be used to remind the rider to adjust their riding posture. This helps ensure passenger comfort.

[0131] Optionally, such as Figure 8-3 As shown, after a shared vehicle ride ends, the server can push rider status statistics to the user's mobile device. The user can then upload objections to these statistics via their mobile device. Based on these objections, the server can determine the true rider status and apply it to the training of the prediction model.

[0132] In some optional examples, to train the predictive model, shared vehicles used to collect training data can be deployed in different cities, and health data can be collected from passengers of different ages. For example, passengers can wear fitness watches, and while sitting in the passenger seats of the shared vehicle, health data such as heart rate, blood pressure, mood, and body temperature can be collected through the fitness watches. At different times and on different road sections, various riding methods can be used to collect data such as cycling acceleration, cycling speed, sway amplitude, and cycling gradient, and upload them to the server. The server can then input passenger health data, weather conditions, and other data at the same time into the initial training dataset.

[0133] Next, the initial training dataset can be preprocessed. For example, outlier data can be removed, the data can be arranged in chronological order of collection time, and the data can be normalized.

[0134] Next, feature processing can be performed. For example, fields such as cycling acceleration, sway amplitude, ambient humidity, ambient wind force, slope, and rider age after data preprocessing can be stored in the sequence dataset, while fields such as heart rate, blood pressure, mood, and body surface temperature can be stored in the label dataset. Optionally, the sequence dataset and label dataset can be set as training, validation, and test sets in an 8:1:1 ratio. To speed up training, batch data can be constructed to obtain training batch data, validation batch data, and test batch data. Based on the training batch data, validation batch data, and test batch data, training can be performed for 30 to 50 rounds.

[0135] Optionally, the prediction model can be overfitted. For example, the following overfitting optimization scheme can be used:

[0136] (1) Cross-validation: Divide the training set into multiple parts for training, use the validation set to test the performance of these training results, adjust the model hyperparameters and continue to use the validation set for validation.

[0137] (2) When iterating the training of the prediction model, the performance of each iteration can be measured. When the validation loss starts to increase, the model training should be stopped.

[0138] (3) Reduce model complexity through regularization. For example, the first fully connected layer of the prediction model can be regularized.

[0139] It should be emphasized that the collection, storage, use, processing, transmission, provision, disclosure, and deletion of data (such as data carrying users' personal information) involved in the technical solution disclosed herein comply with relevant laws and regulations and do not violate public order and good morals. Furthermore, the collection and use of users' personal information involved in the technical solution disclosed herein are all conducted with the user's knowledge and authorization, and do not involve the illegal collection or use of users' personal information.

[0140] Exemplary device

[0141] Figure 9 This is a schematic diagram of the structure of a device for controlling shared vehicles provided by some exemplary embodiments of this disclosure. Figure 9 The apparatus shown may include:

[0142] The first acquisition module 910 is used to acquire the riding parameters of the shared vehicle, the environmental parameters of the environment in which the shared vehicle is located, and the passenger information of the passengers carried by the shared vehicle.

[0143] The prediction module 920 is used to predict the state of the passengers in the shared vehicle based on riding parameters, environmental parameters and passenger information, and obtain the passenger state prediction result.

[0144] The control module 930 is used to control the shared vehicle based on the occupant status prediction results.

[0145] In some optional examples, such as Figure 10 As shown, the control module 930 includes:

[0146] The first determining submodule 1010 is used to determine the mood prediction value and / or vital sign prediction value based on the occupant state prediction result;

[0147] The second determining submodule 1020 is used to determine the target mood type based on the mood prediction value and / or the vital sign prediction value;

[0148] The first control submodule 1030 is used to control the shared vehicle to output a prompt message in response to the target mood type being negative, so as to prompt the rider of the shared vehicle to adjust the riding status.

[0149] In some optional examples, the second determining submodule 1020 includes:

[0150] The first determining unit is used to determine the first distribution information of the mood prediction value relative to the preset numerical range corresponding to the mood dimension;

[0151] The second determining unit is used to determine the mood type that matches the first distribution information;

[0152] The third determining unit is used to determine the target mood type in response to the mood type that matches the first distribution information being a positive type.

[0153] The fourth determining unit is used to determine the second distribution information of the predicted value of vital signs relative to the preset numerical range of the corresponding vital sign dimension in response to the mood type that matches the first distribution information being a negative type, determine the mood type that matches the second distribution information, and take the mood type that matches the second distribution information as the target mood type.

[0154] In some optional examples, such as Figure 11 As shown, the control module 930 includes:

[0155] The third determination submodule 1110 is used to determine the target mood type based on the occupant state prediction results;

[0156] The fourth determination submodule 1120 is used to determine the reason information for the target mood type being negative, based on cycling parameters and environmental parameters, in response to the target mood type being negative.

[0157] The second control submodule 1130 is used to control the shared vehicle using a control method adapted to the cause information.

[0158] In some optional examples, such as Figure 12 As shown, the apparatus provided in the embodiments of this disclosure may further include:

[0159] The second acquisition module 1210 is used to acquire the historical training dataset for training the prediction model;

[0160] The third acquisition module 1220 is used to acquire the incremental training dataset;

[0161] Integration module 1230 is used to integrate the historical training dataset and the incremental training dataset to obtain the target training dataset;

[0162] The first training module 1240 is used to train the prediction model using the target training dataset;

[0163] Each training data point in the historical training dataset and each training data point in the incremental training dataset includes: training riding parameters, training environment parameters, training rider information, and rider status labels.

[0164] In some optional examples, such as Figure 13 As shown, the third acquisition module 1220 includes:

[0165] The fifth determining submodule 1310 is used to determine the passenger status statistics during the shared vehicle ride based on the passenger status prediction results in response to the end of the shared vehicle ride.

[0166] The sending submodule 1320 is used to send passenger status statistics information to the mobile terminal bound to the shared vehicle;

[0167] The receiving submodule 1330 is used to receive objection information uploaded by the mobile terminal regarding the occupant status statistics information;

[0168] The sixth determination submodule 1340 is used to determine the actual occupant status during the shared vehicle ride based on the objection information;

[0169] The seventh determination submodule 1350 is used to determine the incremental training dataset based on riding parameters, environmental parameters, rider information, and actual rider status.

[0170] In some optional examples, such as Figure 14 As shown, the apparatus provided in the embodiments of this disclosure further includes:

[0171] The second training module 1410 is used to periodically train the first type of prediction model within a preset time period.

[0172] The third training module 1420 is used to periodically train the first type of prediction model and the second type of prediction model after a preset time period, in response to the prediction accuracy of the second type of prediction model being less than or equal to that of the first type of prediction model.

[0173] The fourth training module 1430 is used to periodically train the second type of prediction model after a preset time period, in response to the prediction accuracy of the second type of prediction model being greater than that of the first type of prediction model, and to stop training the first type of prediction model.

[0174] Among them, the second type of prediction model has a higher model complexity than the first type of prediction model; within a preset time period, the occupant state prediction result is generated by the first type of prediction model; after the preset time period, in response to the fact that the training of the first type of prediction model has not stopped, the occupant state prediction result is generated by the first type of prediction model; after the preset time period, in response to the fact that the training of the first type of prediction model has stopped, the occupant state prediction result is generated by the second type of prediction model.

[0175] In some optional examples, the cycling parameters include at least one of the following: cycling speed, cycling acceleration, sway amplitude, and cycling gradient.

[0176] In some optional examples, the environmental parameters include at least one of the following: ambient temperature, ambient humidity, and ambient wind speed.

[0177] In the apparatus disclosed herein, the various optional embodiments, optional implementation methods and optional examples disclosed above can be flexibly selected and combined as needed to achieve the corresponding functions and effects, and this disclosure does not list them all.

[0178] Exemplary System

[0179] Figure 1 This is a schematic diagram of the structure of a system for controlling a shared vehicle 110 provided by some exemplary embodiments of this disclosure. Figure 1 The system shown may include:

[0180] Shared vehicle 110 is equipped with sensors that are used to collect riding parameters of shared vehicle 110;

[0181] Server 130 is used to obtain environmental parameters of the environment where the shared vehicle 110 is located and passenger information of the passengers carried by the shared vehicle 110. Based on the riding parameters, environmental parameters and passenger information, the server 130 performs state prediction on the passengers carried by the shared vehicle 110 through a prediction model to obtain passenger state prediction results. Based on the passenger state prediction results, the server 130 controls the shared vehicle 110.

[0182] In the system disclosed herein, the various optional embodiments, optional implementation methods and optional examples disclosed above can be flexibly selected and combined as needed to achieve the corresponding functions and effects, and this disclosure does not list them all.

[0183] Exemplary electronic devices

[0184] Figure 15 The illustration shows a block diagram of an electronic device according to an embodiment of the present disclosure. The electronic device 1500 includes one or more processors 1510 and memory 1520.

[0185] The processor 1510 may be a central processing unit (CPU) or other form of processing unit with data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 1500 to perform desired functions.

[0186] The memory 1520 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and / or cache memory. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 1510 may execute one or more computer program instructions to implement the methods of the various embodiments of this disclosure described above and / or other desired functions.

[0187] In one example, the electronic device 1500 may also include an input device 1530 and an output device 1540, which are interconnected via a bus system and / or other forms of connection mechanism (not shown).

[0188] The input device 1530 may also include, for example, a keyboard, a mouse, etc.

[0189] The output device 1540 can output various information to the outside, including, for example, a display, a speaker, a printer, and a communication network and its connected remote output devices, etc.

[0190] Of course, for the sake of simplicity, Figure 15 Only some of the components of the electronic device 1500 relevant to this disclosure are shown, omitting components such as buses, input / output interfaces, etc. In addition, the electronic device 1500 may include any other suitable components depending on the specific application.

[0191] Exemplary computer program products and computer-readable storage media

[0192] In addition to the methods and apparatus described above, embodiments of this disclosure may also be computer program products comprising computer program instructions that, when executed by a processor, cause the processor to perform the steps in the methods according to various embodiments of this disclosure as described in the "Exemplary Methods" section of this specification.

[0193] Computer program products can be written in any combination of one or more programming languages ​​to perform the operations of embodiments of this disclosure. These programming languages ​​include object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on a user's computing device, partially on a user's computing device, as a standalone software package, partially on a user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0194] Furthermore, embodiments of this disclosure may also be computer-readable storage media storing computer program instructions thereon, which, when executed by a processor, cause the processor to perform the steps in the methods according to various embodiments of this disclosure described in the "Exemplary Methods" section above.

[0195] The computer-readable storage medium may be any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may, for example, include, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0196] The basic principles of this disclosure have been described above with reference to specific embodiments. However, the advantages, benefits, and effects mentioned in this disclosure are merely examples and not limitations, and should not be considered as essential features of each embodiment of this disclosure. The specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the scope of this disclosure to the necessity of employing the specific details described above.

[0197] Various modifications and variations can be made to this disclosure without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this disclosure and their equivalents, this disclosure is also intended to include such modifications and variations.

Claims

1. A method for controlling shared vehicles, comprising: Obtain the riding parameters of the shared vehicle, the environmental parameters of the environment in which the shared vehicle is located, and the passenger information of the passengers carried by the shared vehicle; Based on the riding parameters, the environmental parameters, and the occupant information, the state of the occupants in the shared vehicle is predicted by a prediction model to obtain the occupant state prediction result. The shared vehicle is controlled based on the occupant status prediction results.

2. The method according to claim 1, wherein, The control of the shared vehicle based on the occupant status prediction result includes: Based on the occupant status prediction results, determine the mood prediction value and / or vital sign prediction value; Based on the predicted mood values ​​and / or predicted vital signs values, the target mood type is determined; In response to the target mood type being negative, the shared vehicle is controlled to output a prompt message to remind the rider to adjust their riding status.

3. The method according to claim 2, wherein, Determining the target mood type based on the predicted mood value and / or the predicted vital signs value includes: Determine the first distribution information of the mood prediction value relative to the preset numerical range corresponding to the mood dimension; Determine the mood type that matches the first distribution information; In response to the fact that the mood type that matches the first distribution information is a positive type, the mood type that matches the first distribution information is taken as the target mood type. In response to the fact that the mood type that matches the first distribution information is a negative type, a second distribution information is determined for the predicted value of the vital signs relative to the preset numerical range corresponding to the vital signs dimension. The mood type that matches the second distribution information is determined and the mood type that matches the second distribution information is taken as the target mood type.

4. The method according to claim 1, wherein, The control of the shared vehicle based on the occupant status prediction result includes: Based on the occupant state prediction results, the target mood type is determined; In response to the target mood type being negative, the reason information for the target mood type being negative is determined based on the cycling parameters and the environmental parameters; The shared vehicle is controlled using a control method adapted to the aforementioned cause information.

5. The method according to any one of claims 1-4, further comprising: Obtain the historical training dataset used for training the prediction model; Obtain the incremental training dataset; The historical training dataset and the incremental training dataset are integrated to obtain the target training dataset; The prediction model is trained using the target training dataset; Each training data point in the historical training dataset and each training data point in the incremental training dataset includes: training riding parameters, training environment parameters, training rider information, and rider status labels.

6. The method according to claim 5, wherein, The process of obtaining the incremental training dataset includes: In response to the end of the shared vehicle ride, based on the occupant status prediction result, determine the occupant status statistics during the shared vehicle ride. Send the passenger status statistics information to the mobile terminal bound to the shared vehicle; Receive objection information uploaded by the mobile terminal regarding the occupant status statistics; Based on the objection information, the actual occupant status during the shared vehicle ride is determined; Based on the riding parameters, the environmental parameters, the rider information, and the actual rider status, an incremental training dataset is determined.

7. The method according to any one of claims 1-4, further comprising: The first type of prediction model is trained periodically within a preset time period; After the preset time period, in response to the prediction accuracy of the second type of prediction model being less than or equal to that of the first type of prediction model, the first type of prediction model and the second type of prediction model are periodically trained. After the preset time period, in response to the fact that the prediction accuracy of the second type of prediction model is greater than that of the first type of prediction model, the second type of prediction model is periodically trained, and the training of the first type of prediction model is stopped. Wherein, the second type of prediction model has a higher model complexity than the first type of prediction model; within the preset time period, the occupant state prediction result is generated via the first type of prediction model; after the preset time period, in response to the fact that the training of the first type of prediction model has not stopped, the occupant state prediction result is generated via the first type of prediction model; after the preset time period, in response to the fact that the training of the first type of prediction model has stopped, the occupant state prediction result is generated via the second type of prediction model.

8. A device for controlling shared vehicles, comprising: The first acquisition module is used to acquire the riding parameters of the shared vehicle, the environmental parameters of the environment in which the shared vehicle is located, and the passenger information of the passengers carried by the shared vehicle. The prediction module is used to predict the state of the passengers in the shared vehicle based on the riding parameters, the environmental parameters and the passenger information, through a prediction model, and obtain the passenger state prediction result. The control module is used to control the shared vehicle based on the occupant status prediction results.

9. A system for controlling shared vehicles, comprising: The shared vehicle is equipped with sensors for collecting riding parameters of the shared vehicle. The server is used to obtain environmental parameters of the environment in which the shared vehicle is located and passenger information of the passengers in the shared vehicle. Based on the riding parameters, the environmental parameters and the passenger information, the server performs state prediction on the passengers in the shared vehicle through a prediction model to obtain passenger state prediction results, and controls the shared vehicle based on the passenger state prediction results.

10. An electronic device, the electronic device comprising: processor; Memory used to store the processor's executable instructions; The processor is configured to read the executable instructions from the memory and execute the instructions to implement the method for controlling a shared vehicle as described in any one of claims 1-7.

11. A computer-readable storage medium storing a computer program for performing the method for controlling a shared vehicle as described in any one of claims 1-7.

12. A computer program product comprising computer program instructions that, when executed by a processor, implement the method for controlling a shared vehicle as described in any one of claims 1-7.

Images