Method and system for establishing a general energy consumption model for an unattended flow vehicle
By establishing a local energy consumption model on unmanned logistics vehicles and integrating it with the cloud to generate a general model, the lack of sharing in energy consumption management of unmanned logistics vehicles is solved, and the accuracy of energy consumption prediction and operational efficiency are improved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHANGHAI ECAR TECHNOLOGY CO LTD
- Filing Date
- 2026-05-20
- Publication Date
- 2026-07-31
AI Technical Summary
The lack of a universal model for energy consumption management of existing unmanned logistics vehicles results in each vehicle being managed independently, making it impossible to share energy consumption data. Data collection needs to be repeated when a new vehicle is introduced or when entering a new area, which affects operational efficiency and cost control.
A local energy consumption model is established by collecting data through vehicle terminals, and a general energy consumption model is generated by integrating the data with cloud servers. Data is shared among vehicles to optimize the local energy consumption model.
It improved the accuracy and efficiency of energy consumption models, reduced redundant data collection, optimized operational strategies and path planning, and lowered operating costs.
Smart Images
Figure CN122490831A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of autonomous driving technology, and in particular to a method and system for establishing and applying a general energy consumption model for unmanned logistics vehicles. Background Technology
[0002] With the development of the logistics industry and the advancement of autonomous driving technology, unmanned logistics vehicles are being used more and more widely in the field of logistics delivery. However, energy consumption management of unmanned logistics vehicles has always been a key issue restricting their operational efficiency and cost control.
[0003] Some unmanned logistics vehicle systems collect environmental parameters (such as temperature and humidity) to make simple compensations for the vehicle's local energy consumption model. However, in current technology, the local energy consumption models of each unmanned logistics vehicle are unrelated, and each vehicle manages its energy consumption independently. This results in a lack of energy consumption data sharing between vehicles, making it impossible to form a universal energy consumption model. When new unmanned logistics vehicles are put into use, or when older unmanned logistics vehicles enter new operating areas, in order to obtain a local energy consumption model that can more accurately describe and predict the energy consumption during the operation of the unmanned logistics vehicle in the current operating environment, the unmanned logistics vehicle needs to repeat the data collection process to establish or optimize the local energy consumption model.
[0004] To address the aforementioned issues, this invention discloses a method and system for establishing and applying a general energy consumption model for unmanned logistics vehicles. Summary of the Invention
[0005] The purpose of this invention is to provide a method and system for establishing and applying a general energy consumption model for unmanned logistics vehicles, so as to solve the problem that there is no general energy consumption model for unmanned logistics vehicles applicable to various operating environments in the prior art.
[0006] To achieve the above objectives, this invention discloses a method for establishing and applying a general energy consumption model for unmanned logistics vehicles, comprising: when the unmanned logistics vehicle is in the autonomous operation phase, using the vehicle's onboard terminal to collect data to obtain collected data, the collected data including external environmental parameters, vehicle internal parameters, and real-time vehicle energy consumption; establishing a preset local energy consumption model in the onboard terminal, the local energy consumption model including a mapping relationship between external environmental parameters, vehicle internal parameters, and the real-time vehicle energy consumption; calibrating the weight coefficients of the external environmental parameters and vehicle internal parameters in the mapping relationship using the collected data, and sending the calibrated local energy consumption model to a cloud server; the cloud server receiving several local energy consumption models sent by the onboard terminals, integrating the several local energy consumption models to generate a general energy consumption model, and sending the general energy consumption model to several unmanned logistics vehicles, so that the unmanned logistics vehicles optimize the local energy consumption model according to the general energy consumption model.
[0007] Preferably, when the onboard terminal detects a remote driving signal or a manual intervention signal, it determines that the unmanned logistics vehicle has left the autonomous operation phase and suspends data collection. During the non-autonomous operation phase, the unmanned logistics vehicle's operation is under human control, and the data collected at this time is not relevant for calibrating the mapping relationships included in the local energy consumption model.
[0008] Specifically, the external environmental parameters include temperature, humidity, wind speed, precipitation, road slope, and road roughness, while the vehicle internal parameters include speed, acceleration, vehicle load, battery status, and motor power. All of these parameters are related to the energy consumption of the unmanned logistics vehicle, and comprehensive data collection can make the calibrated local energy consumption model more accurate.
[0009] Preferably, the local energy consumption model further calculates the estimated energy consumption of the unmanned logistics vehicle based on the mapping relationship, including: the local energy consumption model, based on the mapping relationship, uses a preset optimization algorithm to calculate and adjust the weighting coefficients of the external environmental parameters and vehicle internal parameters on the real-time energy consumption of the vehicle during the energy consumption prediction process in real time, so as to obtain the accurate estimated energy consumption of the unmanned logistics vehicle. This allows for adjustments to operational strategies, planning of delivery routes, and control of operating costs based on the estimated energy consumption.
[0010] Preferably, after the data synchronization conditions are met, it is determined that the local energy consumption model is calibrated; the data synchronization conditions are: the unmanned logistics vehicle's operating time reaches a preset threshold, and its mileage reaches a preset threshold. This method makes the local energy consumption model sent to the cloud server more reliable, providing high-quality data raw materials for generating the general energy consumption model.
[0011] Preferably, integrating several local energy consumption models includes grouping the unmanned logistics vehicles based on vehicle operating factors; Integrate the sub-energy consumption models of each group of unmanned logistics vehicles: Use an ensemble learning algorithm to integrate the weight coefficients of the local energy consumption models of the unmanned logistics vehicles in each group to obtain the weight coefficients of the sub-energy consumption models, thereby forming the sub-energy consumption models. The sub-energy consumption models of all groups of unmanned logistics vehicles are combined into a general energy consumption model applicable to several groups of vehicle operating factors.
[0012] Preferably, before grouping the unmanned logistics vehicles according to vehicle operation factors, the method further includes: performing data cleaning on the local energy consumption models of several unmanned logistics vehicles to remove the local energy consumption models containing abnormal data.
[0013] Specifically, the unmanned logistics vehicles are grouped according to vehicle type and operating area.
[0014] Preferably, the optimization of the local energy consumption model by the unmanned logistics vehicle based on the general energy consumption model includes: when the unmanned logistics vehicle has been operating continuously in an operating area for no longer than a preset duration, or when the unmanned logistics vehicle is newly put into use or enters a new operating area, the corresponding sub-energy consumption model is directly adopted as the local energy consumption model.
[0015] The present invention also discloses an application system for establishing a general energy consumption model for unmanned logistics vehicles, comprising several vehicle-mounted terminals, a cloud server and a communication network. The several vehicle-mounted terminals are installed on the unmanned logistics vehicle, and the several vehicle-mounted terminals and the cloud server are connected through the communication network to realize the application method for establishing a general energy consumption model for unmanned logistics vehicles as described above.
[0016] Compared with existing technologies, this invention utilizes cloud server collaboration to generate a general energy consumption model based on several local energy consumption models. The general energy consumption model includes multiple sub-energy consumption models corresponding to different operating conditions. This allows newly deployed unmanned logistics vehicles or those entering new operating areas to directly adopt sub-energy consumption models that suit their own operating conditions as their local energy consumption models. For unmanned logistics vehicles that have been operating for a certain period of time, the general energy consumption model can also be used to optimize and calibrate their own local energy consumption models. On the one hand, this improves the accuracy of local energy consumption models in describing and predicting the energy consumption of unmanned logistics vehicles during operation. On the other hand, it significantly improves the efficiency of forming accurate local energy consumption models without the need for repetitive data collection to establish or optimize local energy consumption models. Attached Figure Description
[0017] Figure 1 This is a schematic diagram illustrating the steps of establishing and applying the general energy consumption model for unmanned logistics vehicles disclosed in this invention.
[0018] Figure 2 This is a schematic diagram of the architecture of the application system for establishing a general energy consumption model for unmanned logistics vehicles disclosed in this invention. Detailed Implementation
[0019] To illustrate the technical content, structural features, objectives, and effects of the present invention in detail, the following description is provided in conjunction with the embodiments and accompanying drawings.
[0020] refer to Figure 1 This invention discloses a method for establishing and applying a general energy consumption model for unmanned logistics vehicles, including steps S10-S40.
[0021] S10, when the unmanned logistics vehicle is in the autonomous operation phase, the on-board terminal of the unmanned logistics vehicle is used to collect data to obtain the collected data, which includes external environmental parameters, vehicle internal parameters and real-time energy consumption of the vehicle.
[0022] It should be noted that when the vehicle-mounted terminal detects a remote driving signal or a manual intervention signal, it determines that the unmanned logistics vehicle has left the autonomous operation phase and suspends data collection. During the non-autonomous operation phase, the unmanned logistics vehicle's operation is under human control, and the data collected at this time is not relevant for calibrating the mapping relationships included in the local energy consumption model.
[0023] Specifically, the external environmental parameters include temperature, humidity, wind speed, precipitation, road slope, and road roughness, while the vehicle internal parameters include speed, acceleration, vehicle load, battery status, and motor power. All of these parameters are related to the energy consumption of the unmanned logistics vehicle. Comprehensive data collection can make the calibrated local energy consumption model more accurate. Of course, the energy consumption-related parameters that need to be collected are not limited to the above-mentioned parameters, and this embodiment does not impose any particular restrictions.
[0024] S20, a preset local energy consumption model is established in the vehicle terminal. The local energy consumption model includes the mapping relationship between external environmental parameters, vehicle internal parameters and the real-time energy consumption of the vehicle.
[0025] The local energy consumption model further calculates the estimated energy consumption of the unmanned logistics vehicle based on the mapping relationship. Specifically, the local energy consumption model uses a preset optimization algorithm to calculate and adjust the weighting coefficients of external environmental parameters and vehicle internal parameters on the real-time energy consumption of the vehicle during the energy consumption prediction process, based on the mapping relationship, to obtain an accurate estimated energy consumption for the unmanned logistics vehicle. This allows for adjustments to operational strategies, delivery route planning, and cost control based on the estimated energy consumption.
[0026] In this embodiment, the optimization algorithm is the gradient descent algorithm.
[0027] For example, when an increase in external wind speed is detected, the influence weight of the drag coefficient on vehicle energy consumption is calculated and adjusted in real time using a preset optimization algorithm based on the mapping relationship; when the cargo load increases, the influence weight of the load on vehicle energy consumption is adjusted based on the mapping relationship to obtain an accurate predicted energy consumption of the unmanned logistics vehicle under the current environment. Based on an accurate local energy consumption model, accurate basis is provided for the system's path planning, operational decisions, etc.
[0028] In reality, unmanned logistics vehicles often face various environments during operation, leading to inaccurate predictions of vehicle energy consumption. This results in a series of problems, such as the lack of optimal delivery route planning, insufficient vehicle range during delivery, and increased delivery costs. Existing energy consumption models cannot fully handle these situations. When one or more parameters, either external environmental parameters or internal vehicle parameters, change, the local energy consumption model adjusts the weights of each parameter's influence on energy consumption in real time based on the mapping relationship. This achieves an accurate description and prediction of the unmanned logistics vehicle's energy consumption during operation, providing a theoretical basis for vehicle decision-making.
[0029] S30, using the collected data to calibrate the weight coefficients of the external environment parameters and vehicle internal parameters in the mapping relationship, and sending the calibrated local energy consumption model to the cloud server.
[0030] Preferably, after the data synchronization conditions are met, the local energy consumption model is determined to be calibrated. The data synchronization conditions are: the unmanned logistics vehicle's operating time reaches a preset threshold, and its mileage reaches a preset threshold. The above method makes the local energy consumption model sent to the cloud server more reliable, providing high-quality data raw materials for generating the general energy consumption model.
[0031] This embodiment does not impose any special restrictions on the data synchronization conditions. For example, in order to make the data contained in the local energy consumption model more comprehensive, the following conditions can be added to the original data synchronization conditions: when collecting data during the operation of the unmanned logistics vehicle, at least three different weather conditions should be covered, such as sunny, cloudy, light rain, and light snow.
[0032] S40, the cloud server receives local energy consumption models sent by several vehicle terminals, integrates several local energy consumption models to generate a general energy consumption model, and sends the general energy consumption model to several unmanned logistics vehicles so that the unmanned logistics vehicles optimize the local energy consumption model according to the general energy consumption model.
[0033] The integration of several local energy consumption models includes: grouping the unmanned logistics vehicles according to vehicle operation factors; Integrate the sub-energy consumption models of each group of unmanned logistics vehicles: Use an ensemble learning algorithm to integrate the weight coefficients of the local energy consumption models of the unmanned logistics vehicles in each group to obtain the weight coefficients of the sub-energy consumption models, thereby forming the sub-energy consumption models. The sub-energy consumption models of all groups of unmanned logistics vehicles are combined into a general energy consumption model applicable to several groups of vehicle operating factors.
[0034] Preferably, before grouping the unmanned logistics vehicles based on vehicle operating factors, the method further includes: performing data cleaning on the local energy consumption models of several unmanned logistics vehicles to remove local energy consumption models containing abnormal data. For the local energy consumption models of several unmanned logistics vehicles, one or more of them may have significant differences compared with other local energy consumption models. Removing such local energy consumption models can further improve the data yield of the remaining local energy consumption models used to generate a general energy consumption model.
[0035] Specifically, the unmanned logistics vehicles are grouped according to vehicle type and operating area.
[0036] In this embodiment, the ensemble learning algorithm can be the random forest algorithm.
[0037] Specifically, the optimization of the local energy consumption model by the unmanned logistics vehicle based on the general energy consumption model includes: when the unmanned logistics vehicle has been operating continuously in an operating area for no longer than a preset duration, or when the unmanned logistics vehicle is newly put into use or enters a new operating area, the corresponding sub-energy consumption model is directly adopted as the local energy consumption model.
[0038] In practice, when the unmanned logistics vehicle operates continuously in a certain area for a considerable period, the local energy consumption model can also be optimized by calibrating some mapping relationships based on the general energy consumption model. Furthermore, when the unmanned logistics vehicle operates continuously in a certain area for a sufficiently long time, and the collected data covers various situations during its operation, its local energy consumption model is already accurate enough and does not require optimization using the general energy consumption model. The general energy consumption model is more commonly used when the unmanned logistics vehicle is newly deployed or enters a new operating area; compared to re-collecting data, directly applying the general energy consumption model can significantly improve operational efficiency.
[0039] Preferably, the local energy consumption model does not permanently end the data collection process after a preset time of data collection, and the general energy consumption model is not generated once or a few times and then stopped. Instead, the calibration of the local energy consumption model and the generation and distribution of the general energy consumption model are long-term, continuous system behaviors.
[0040] For a further example, several unmanned logistics vehicles, under the aforementioned data synchronization conditions, periodically (e.g., every two weeks) send the calibrated local energy consumption model to the cloud server. The cloud server continuously receives new local energy consumption models and periodically (e.g., monthly) updates the general energy consumption model. During the update process, valid information from historical general energy consumption models is retained, while the optimization results of newly received local energy consumption models are integrated to generate a new general energy consumption model. The updated general energy consumption model is then verified using multiple unmanned logistics vehicles to ensure its accuracy under various environments and operating conditions. Finally, the verified general energy consumption model is sent to each unmanned logistics vehicle.
[0041] refer to Figure 2 The present invention also discloses an application system for establishing a general energy consumption model for unmanned logistics vehicles, including several vehicle-mounted terminals 10, a cloud server 20 and a communication network 30. The several vehicle-mounted terminals 10 are installed on the unmanned logistics vehicle, and the several vehicle-mounted terminals 10 and the cloud server 20 are connected to each other through the communication network 30 to realize the application method for establishing a general energy consumption model for unmanned logistics vehicles as described above.
[0042] The vehicle-mounted terminal 10 includes a data acquisition module, a model optimization module, and a data synchronization module. The data acquisition module collects data to obtain collected data, which is then sent to the model optimization module. The collected data includes external environmental parameters, vehicle internal parameters, and real-time vehicle energy consumption. The model optimization module establishes a preset local energy consumption model, which includes a mapping relationship between external environmental parameters, vehicle internal parameters, and the real-time vehicle energy consumption. The model optimization module uses the collected data to calibrate the weighting coefficients of the external environmental parameters and vehicle internal parameters in the mapping relationship. After meeting the data synchronization conditions, the data synchronization module sends the calibrated local energy consumption model to the cloud server 20.
[0043] In this embodiment, the data acquisition module includes various sensors for acquiring various parameters. For example, among the sensors for acquiring external environmental parameters, the temperature and humidity sensor acquires the ambient temperature and humidity, the wind speed sensor acquires the wind speed, and the road condition sensor acquires the road surface slope and roughness. Among the sensors for acquiring vehicle internal parameters, the speed sensor acquires the vehicle speed, the acceleration sensor acquires the driving acceleration, the load sensor acquires the cargo weight, the battery management system acquires the battery status, and the motor controller acquires the motor power, etc.
[0044] The cloud server 20 receives local energy consumption models sent by several vehicle terminals 10, integrates the local energy consumption models to generate a general energy consumption model, and sends the general energy consumption model to several unmanned logistics vehicles so that the unmanned logistics vehicles can optimize the local energy consumption model based on the general energy consumption model.
[0045] The communication network 30 connects several vehicle-mounted terminals 10 and cloud server 20 to realize data transmission between the vehicle-mounted terminals 10 and cloud server 20.
[0046] In this embodiment, the communication network 30 can be a 4G network or a 5G network, etc.
[0047] In this embodiment, from the perspective of electronic devices, the vehicle-mounted terminal 10 includes several memories, a processor, and one or more programs, wherein the memories and the processor are used to store and execute the programs, respectively. Similarly, the cloud server 20 includes several memories, a processor, and one or more programs, wherein the memories and the processor are used to store and execute the programs, respectively. The vehicle-mounted terminal 10 and the cloud server 20 execute the above-mentioned programs to jointly implement the application method for establishing a general energy consumption model for unmanned logistics vehicles as described above.
[0048] Compared with existing technologies, this invention utilizes cloud server collaboration to generate a general energy consumption model based on several local energy consumption models. The general energy consumption model includes multiple sub-energy consumption models corresponding to different operating conditions. This allows newly deployed unmanned logistics vehicles or those entering new operating areas to directly adopt sub-energy consumption models that suit their own operating conditions as their local energy consumption models. For unmanned logistics vehicles that have been operating for a certain period of time, the general energy consumption model can also be used to optimize and calibrate their own local energy consumption models. On the one hand, this improves the accuracy of local energy consumption models in describing and predicting the energy consumption of unmanned logistics vehicles during operation. On the other hand, it significantly improves the efficiency of forming accurate local energy consumption models without the need for repetitive data collection to establish or optimize local energy consumption models.
[0049] The above-disclosed embodiments are merely preferred embodiments of the present invention and should not be construed as limiting the scope of the present invention. Therefore, any equivalent variations made in accordance with the scope of the present invention are still within the scope of the present invention.
Claims
1. A method for establishing and applying a general energy consumption model for unmanned logistics vehicles, characterized in that, include: When the unmanned logistics vehicle is in the autonomous operation phase, the on-board terminal of the unmanned logistics vehicle is used to collect data to obtain the collected data, which includes external environmental parameters, vehicle internal parameters and real-time energy consumption of the vehicle. A preset local energy consumption model is established in the vehicle terminal. The local energy consumption model includes the mapping relationship between external environmental parameters, vehicle internal parameters and the real-time energy consumption of the vehicle. The collected data is used to calibrate the weighting coefficients of the external environment parameters and vehicle internal parameters in the mapping relationship, and the calibrated local energy consumption model is sent to the cloud server. The cloud server receives local energy consumption models sent by several vehicle terminals, integrates the local energy consumption models to generate a general energy consumption model, and sends the general energy consumption model to several unmanned logistics vehicles so that the unmanned logistics vehicles can optimize the local energy consumption models based on the general energy consumption model.
2. The method for establishing and applying a general energy consumption model for unmanned logistics vehicles as described in claim 1, characterized in that, When the vehicle terminal detects a remote driving signal or a manual intervention signal, it determines that the unmanned logistics vehicle has left the autonomous operation phase and suspends data collection.
3. The method for establishing and applying the general energy consumption model for unmanned logistics vehicles as described in claim 1, characterized in that, The external environmental parameters include temperature, humidity, wind speed, precipitation, road slope, and road roughness, while the vehicle internal parameters include speed, acceleration, vehicle load, battery status, and motor power.
4. The method for establishing and applying a general energy consumption model for unmanned logistics vehicles as described in claim 1, characterized in that, The local energy consumption model also calculates the expected energy consumption of the unmanned logistics vehicle based on the mapping relationship, including: the local energy consumption model calculates and adjusts the weighting coefficients of the external environmental parameters and vehicle internal parameters on the real-time energy consumption of the vehicle in the energy consumption prediction process in real time using a preset optimization algorithm based on the mapping relationship, so as to obtain the accurate expected energy consumption of the unmanned logistics vehicle.
5. The method for establishing and applying a general energy consumption model for unmanned logistics vehicles as described in claim 1, characterized in that, After the data synchronization conditions are met, it is determined that the local energy consumption model is calibrated. The data synchronization conditions are: the running time of the unmanned logistics vehicle reaches a preset threshold, and the driving mileage reaches a preset threshold.
6. The method for establishing and applying a general energy consumption model for unmanned logistics vehicles as described in claim 1, characterized in that, Integrating several of the aforementioned local energy consumption models includes: The unmanned logistics vehicles are grouped according to vehicle operation factors; Integrate the sub-energy consumption models of each group of unmanned logistics vehicles: Use an ensemble learning algorithm to integrate the weight coefficients of the local energy consumption models of the unmanned logistics vehicles in each group to obtain the weight coefficients of the sub-energy consumption models, thereby forming the sub-energy consumption models. The sub-energy consumption models of all groups of unmanned logistics vehicles are combined into a general energy consumption model applicable to several groups of vehicle operating factors.
7. The method for establishing and applying a general energy consumption model for unmanned logistics vehicles as described in claim 6, characterized in that, Before grouping the unmanned logistics vehicles based on vehicle operation factors, the process further includes: performing data cleaning on the local energy consumption models of several unmanned logistics vehicles to remove local energy consumption models containing abnormal data.
8. The method for establishing and applying the general energy consumption model for unmanned logistics vehicles as described in claim 6, characterized in that, The unmanned logistics vehicles are grouped according to vehicle type and operating area.
9. The method for establishing and applying a general energy consumption model for unmanned logistics vehicles as described in claim 6, characterized in that, The optimization of the local energy consumption model based on the general energy consumption model for the unmanned logistics vehicle includes: when the unmanned logistics vehicle has been operating continuously in an operating area for no longer than a preset duration, or when the unmanned logistics vehicle is newly put into use or enters a new operating area, the corresponding sub-energy consumption model is directly adopted as the local energy consumption model.
10. A system for establishing a general energy consumption model for unmanned logistics vehicles, characterized in that, It includes several vehicle-mounted terminals, a cloud server, and a communication network. Several of the vehicle-mounted terminals are installed on unmanned logistics vehicles, and the several vehicle-mounted terminals and the cloud server are connected through the communication network to implement the method for establishing and applying a general energy consumption model for unmanned logistics vehicles as described in claims 1-9.