Intelligent scheduling system of shared electric two-wheeled vehicle

Abstract

The invention is suitable for the technical field of intelligent scheduling, and provides an intelligent scheduling system of a shared electric two-wheeled vehicle, which comprises a vehicle-mounted terminal, a cloud computing platform, a scheduling management terminal and a user terminal, the cloud computing platform is respectively connected with the vehicle-mounted terminal, the dispatching management terminal and the user terminal; the vehicle-mounted terminal is used for uploading the collected multi-source data to the cloud computing platform; a demand prediction unit in the cloud computing platform generates vehicle demand prediction results in different time periods; the battery management unit generates a battery state analysis result; the wireless transmission unit is used for transmitting data; the scheduling management terminal is used for generating a scheduling path strategy with scheduling cost, user waiting cost and battery replacement cost as targets; and the user terminal is used for providing multi-dimensional intelligent services. According to the invention, through cooperative work of dynamic perception, accurate prediction and intelligent decision making, the timeliness and comprehensiveness of electric scheduling are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent scheduling, and particularly relates to an intelligent scheduling system for shared electric two-wheel vehicles. Background Art

[0002] With the popularization of the shared travel mode, certain progress has been made in the field of intelligent scheduling of shared electric two-wheel vehicles in existing research. Currently, the existing scheduling system lacks the ability of real-time environment perception, resulting in the problem of imbalance between vehicle supply and demand, where there is a shortage of vehicles in popular areas during peak hours while vehicles are idle in unpopular areas. Secondly, since the battery health monitoring is only based on single index factors such as the number of charge-discharge cycles or voltage thresholds, without considering the dynamic impact of the humid and hot environment on battery attenuation, the deviation of the endurance prediction is large, and users frequently encounter power-off during the journey. In addition, methods for shared bicycles such as vehicle tracking through GPS positioning, or dividing high-demand areas based on clustering algorithms, or detecting low-battery vehicles through regular inspections or user feedback are difficult to directly adapt to the dynamic charging and swapping requirements of electric two-wheel vehicles and the complex urban road network environment, resulting in poor real-time performance of scheduling and incomplete scheduling.

[0003] In view of this, an intelligent scheduling system for shared electric two-wheel vehicles is proposed. Summary of the Invention

[0004] The present invention provides an intelligent scheduling system for shared electric two-wheel vehicles, which is used to solve the problems of poor real-time performance of scheduling and incomplete scheduling.

[0005] The present invention provides an intelligent scheduling system for shared electric two-wheel vehicles, including: On-vehicle terminal, cloud computing platform, scheduling management terminal and user terminal; the cloud computing platform is respectively connected to the on-vehicle terminal, scheduling management terminal and user terminal; The on-vehicle terminal is used to collect vehicle positioning data, vehicle status data, vehicle environment data and battery status data in real time, and upload them to the cloud computing platform; The cloud computing platform includes: Demand prediction unit, which analyzes the vehicle usage rules in each region based on the vehicle positioning data, vehicle status data and vehicle environment data, and generates the predicted results of vehicle demand in different time periods; Battery management unit, which generates the battery status analysis results including remaining life, charging line and charging time based on the battery status data; Wireless transmission unit, which is used to push the vehicle positioning data, vehicle status data and battery status data to the user terminal, and provide the vehicle demand prediction results, battery status analysis results and vehicle positioning data to the scheduling management terminal; The scheduling management terminal is used to generate a scheduling path strategy targeting scheduling cost, user waiting cost, and battery replacement cost based on the vehicle demand prediction result, battery status analysis result, and vehicle positioning data; The user terminal is used to provide vehicle real-time location query, vehicle reservation based on navigation, and vehicle usage status feedback services.

[0006] Furthermore, the in-vehicle terminal includes: A positioning module, used to obtain the vehicle real-time position coordinates and driving trajectory data; A vehicle status perception module, used to detect the running status, usage status, and fault signals of the vehicle; An environment monitoring module, used to collect the temperature data and humidity data of the environment where the vehicle is located; A battery status monitoring module, used to collect the remaining battery power and health status data; A main controller, respectively connected to the positioning module, battery status monitoring module, vehicle status perception module, and environment monitoring module, used to integrate and process multi-source data and generate transmission data; A data transmission module, used to upload the transmission data to the cloud computing platform.

[0007] Furthermore, analyzing the vehicle usage rules of each region based on the vehicle positioning data, vehicle status data, and vehicle environment data, and generating a vehicle demand prediction result for different time periods, including: Respectively extracting the vehicle distribution heat map features of each region based on the vehicle real-time position coordinates and driving trajectory data, calculating the vehicle idle rate, failure rate, and real-time usage frequency based on the running status, usage status, and fault signals of the vehicle, and generating an environment correction coefficient based on the temperature data and humidity data of the environment where the vehicle is located; Taking the vehicle distribution heat map features, idle rate, failure rate, and real-time usage frequency as input features and inputting them into an improved LSTM model, where the environment correction coefficient adjusts the memory weights of the improved LSTM model for temperature and humidity sensitive periods by embedding the gating unit of the improved LSTM model; Outputting the vehicle demand prediction result of each region within a preset time.

[0008] Furthermore, the environment correction coefficient adjusts the memory weights of the improved LSTM model for temperature and humidity sensitive periods by embedding the gating unit of the improved LSTM model, including: Mapping the temperature data and humidity data into a temperature influence factor and a humidity influence factor respectively, and generating an environment correction coefficient through linear weighting; Concatenate the environmental correction coefficient with the historical hidden state corresponding to the gating unit to generate temperature and humidity sensitive gating weights; During a preset high temperature and high humidity sensitive period, reduce the forgetting rate of the gating weights to a preset percentage range to enhance the memory strength of the improved LSTM model for the historical vehicle usage patterns during the same period.

[0009] Furthermore, generating a battery state analysis result including remaining life, charging line, and charging time based on the battery state data includes: Obtain the battery internal resistance change rate, capacity attenuation rate, and environmental temperature deviation value based on the battery state data, where the environmental temperature deviation value is the absolute difference between the ambient temperature where the battery is located and a preset reference temperature; Establish a battery health state decay rate model through the correlation analysis of the battery internal resistance change rate, capacity attenuation rate, and temperature deviation value; Calculate the remaining duration of the cycle life based on the difference between the cumulative cycle times of the battery and a preset cycle life threshold; calculate the remaining duration of the capacity life based on the difference between the current health state index and a preset capacity attenuation threshold, and combine the decay rate model; Compare the remaining duration of the cycle life and the remaining duration of the capacity life, and select the smaller value as the target prediction result of the remaining life.

[0010] Furthermore, generating a battery state analysis result including remaining life, charging line, and charging time based on the battery state data further includes: Divide the charging priority levels according to the remaining power of the vehicle and the battery health state; Based on the real-time position of the vehicle, the distribution of charging stations, and the real-time traffic data obtained through a third-party map interface, with the goal of the shortest driving time, the highest charging efficiency, and the lowest battery health loss, use the linear weighted method to allocate weights to multiple objectives to generate the optimal charging path; Analyze and update the charging path in real time according to the traffic state change and the availability of charging piles, and push the updated charging path to the scheduling management terminal.

[0011] Furthermore, generating a battery state analysis result including remaining life, charging line, and charging time based on the battery state data further includes: Dynamically calculate the basic charging duration according to the charging amount required, the charging pile power, and the battery health state, and compensate and correct the charging duration in combination with the environmental temperature data; Based on the peak vehicle usage period output by the demand prediction unit and the time-of-use electricity price data obtained through the power grid server interface, select a period with low demand and low electricity price as the recommended charging window; When the battery temperature is detected to be abnormal or the health status is lower than the safety threshold, a cooling interval is automatically inserted and the charging duration is extended.

[0012] Furthermore, the generation of a scheduling path strategy targeting scheduling cost, user waiting cost, and battery replacement cost includes: The expression of the objective function: Where: is the vehicle number, is the total number of shared electric two-wheel vehicles, is the area number, is the total number of areas, is the spare battery number, is the total number of available batteries, is the unit distance scheduling cost, is the vehicle scheduled to the distance of, is the unit time congestion cost, is the vehicle scheduled to the real-time traffic congestion time of, obtained through the Gaode Map API, is a binary variable. If the vehicle is scheduled from to area , then , otherwise . is the unit cost of user waiting, is the predicted demand of area , is the initial number of vehicles in area , is the single battery replacement cost, is the correction factor for humid and hot climate. The battery wears out faster in high-temperature and high-humidity environments, is a binary variable. If the vehicle uses the spare battery , then , otherwise .

[0013] Furthermore, the constraint conditions of the scheduling path strategy include vehicle allocation constraint, area capacity constraint, battery health constraint, time window constraint, high-demand area service constraint, and charging pile competition constraint.

[0014] Furthermore, the user terminal includes: A positioning query module, used to send a vehicle real-time location query request to the cloud computing platform, receive and display the positioning data of the target vehicle; A navigation reservation module, which is used to initiate a vehicle locking request to a cloud computing platform according to a reservation instruction input by a user and a navigation path planning result, and receive a reservation confirmation message and vehicle status information returned by the cloud computing platform; A status feedback module, which is used to collect feedback data of a user on the usage status of a vehicle and upload the feedback data to the cloud computing platform; Wherein, the navigation reservation module generates a navigation path from the current position of the user to the target vehicle by calling a third-party map service interface, and the vehicle locking request includes a vehicle number, a reservation time, and user identity authentication information.

[0015] It can be seen from the above technical solutions that the present invention has the following advantages: The demand prediction unit of the present invention reduces the shortage of vehicles in popular areas and the idle phenomenon in cold areas by alleviating the problem of uneven vehicle distribution; the battery management unit improves the accuracy of the remaining life and the charging path and reduces the risk of mid-course power-off caused by battery problems for users through the embedded humidity and heat environment correction factor and multi-dimensional health status assessment; the dispatching management terminal generates a globally optimal dispatching strategy, which improves the vehicle turnover rate while reducing the operation cost; the user terminal enhances the response speed and resource allocation rationality of the system through real-time data interaction. Through the synergistic effect of dynamic perception, accurate prediction and intelligent decision-making, the present system solves the problems of scheduling lag, rough battery management and large demand prediction deviation in the prior art, and provides efficient and reliable technical support for the operation of shared electric two-wheel vehicles in high-density cities. Description of the Drawings

[0016] Figure 1 It is a schematic structural diagram of an embodiment of an intelligent dispatching system for a shared electric two-wheel vehicle in the present invention; Figure 2 It is a schematic structural diagram corresponding to an in-vehicle terminal in the present invention; Figure 3 It is a schematic flow diagram of an embodiment corresponding to the demand prediction unit in the present invention; Figure 4a It is a schematic flow diagram of an embodiment of the battery management unit in the present invention; Figure 4b It is a schematic flow diagram of another embodiment of the battery management unit in the present invention; Figure 4c It is a schematic flow diagram of another embodiment of the battery management unit in the present invention. Detailed Embodiments

[0017] In the description of the present application and the above-mentioned drawings, terms such as "first", "second", "third", "fourth", etc. (if any) are used to distinguish similar objects and do not necessarily describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances so that the embodiments of the present application described here can be implemented in an order other than those illustrated or described here. In addition, the terms "comprising" and "corresponding to" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device comprising a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0018] Embodiment 1 Please refer to Figure 1 In the intelligent scheduling system of the shared electric two-wheeler in the present invention, it includes: an on-vehicle terminal 1, a cloud computing platform 2, a scheduling management terminal 3, and a user terminal 4; the cloud computing platform 2 is respectively connected to the on-vehicle terminal 1, the scheduling management terminal 3, and the user terminal 4; the on-vehicle terminal 1 is used to collect vehicle positioning data, vehicle status data, vehicle environment data, and battery status data in real time and upload them to the cloud computing platform 2; the cloud computing platform 2 includes a demand prediction unit 201, which analyzes the vehicle usage rules in each area based on the vehicle positioning data, vehicle status data, and vehicle environment data to generate a vehicle demand prediction result for different time periods; a battery management unit 202 generates a battery status analysis result including remaining life, charging line, and charging time based on the battery status data; a wireless transmission unit 203 is used to push the vehicle positioning data, vehicle status data, and battery status data to the user terminal 4 and provide the vehicle demand prediction result, battery status analysis result, and vehicle positioning data to the scheduling management terminal 3; the scheduling management terminal 3 is used to generate a scheduling path strategy targeting scheduling cost, user waiting cost, and battery replacement cost based on the vehicle demand prediction result, battery status analysis result, and vehicle positioning data; the user terminal 4 is used to provide vehicle real-time location query, vehicle reservation based on navigation, and vehicle usage status feedback services.

[0019] The on-vehicle terminal 1 of the present invention collects vehicle multi-source data in real time and uploads it to the cloud computing platform 2. The platform analyzes and generates scheduling basis through the demand prediction unit and the battery management unit; the scheduling management terminal 3 formulates a multi-objective optimization strategy based on the prediction result and real-time positioning data provided by the platform, and the user terminal 4 obtains vehicle information through the platform and realizes navigation reservation and status feedback. Through data interconnection and functional complementarity, a dynamic perception and intelligent decision-making scheduling system is formed, effectively improving the timeliness and comprehensiveness of scheduling.

[0020] Embodiment 2 Please refer to Figure 2, the in - vehicle terminal 1 in the present invention includes: A positioning module 101, which is used to obtain the real - time position coordinates and driving trajectory data of the vehicle; a vehicle state perception module 102, which is used to detect the running state, usage state and fault signals of the vehicle; an environment monitoring module 103, which is used to collect the temperature data and humidity data of the environment where the vehicle is located; a battery state monitoring module 104, which is used to collect the remaining battery power and health state data; a main controller 105, which is respectively connected to the positioning module, the battery state monitoring module, the vehicle state perception module and the environment monitoring module, and is used to integrate and process multi - source data and generate transmission data; a data transmission module 106, which is used to upload the transmission data to the cloud computing platform.

[0021] Specifically, the positioning module 101 integrates a GPS / Beidou dual - mode positioning chip, receives satellite signals in real - time and parses the longitude and latitude coordinates, and combines the data of the inertial measurement unit to compensate for the positioning drift when the signal is blocked, so as to generate the vehicle position coordinates and driving trajectory data.

[0022] The vehicle state perception module 102 includes an inertial sensor and a vehicle control bus decoder. The inertial sensor detects the three - axis acceleration and angular velocity of the vehicle to judge the driving state; the vehicle control bus decoder is connected to the vehicle motor controller to decode the CAN bus signal to obtain the motor speed, fault diagnosis code and brake signal. Among them, the running state refers to the physical motion state of the vehicle, including driving speed, acceleration, steering angle, etc.; the usage state refers to whether the vehicle is occupied by the user (judged by the in - vehicle Bluetooth lock / user terminal connection state); the fault signals include motor faults, brake system abnormalities, etc.

[0023] The environment monitoring module 103 adopts a digital temperature and humidity sensor, which is installed in the waterproof cavity at the bottom of the vehicle to collect the ambient temperature and humidity in real - time. The detected temperature and humidity data are used to correct the battery health model (such as high temperature accelerating capacity attenuation) and trigger the overheat protection mechanism.

[0024] The battery state monitoring module 104 is implemented through a battery management system chip, and mainly collects the following remaining power and health state. Among them, the remaining power is calculated in real - time based on the Coulomb counting method, and the health state is comprehensively evaluated through internal resistance detection and capacity attenuation rate (the number of cycles compared with the initial capacity).

[0025] The main controller 105 adopts a low - power microcontroller and runs an embedded real - time operating system to implement the following functions: receiving the original data of each module, performing timestamp alignment and data filtering; encapsulating the data in JSON format, adding the vehicle ID, timestamp, and CRC check code to generate a transmission data packet; when a fault signal is detected, triggering a local alarm and marking the data priority.

[0026] The data transmission module 106 adopts a low-power wide-area network communication module, supports the TCP / IP protocol stack, establishes a long connection with the cloud computing platform through the operator network, and uploads data in an event-triggered or periodic reporting manner, supporting resume from breakpoint and encrypted transmission.

[0027] In the above embodiment, the vehicle-mounted terminal collaboratively collects vehicle position, running status, environmental temperature and humidity, and battery health data in real time through multiple modules, providing high-precision multi-dimensional data for subsequent scheduling; at the same time, effectively reducing the operation and maintenance cost and extending the battery life, ensuring the riding safety and experience of users.

[0028] Embodiment III Please refer to Figure 3 , in the demand prediction unit of the present invention, based on vehicle positioning data, vehicle status data, and vehicle environment data, analyze the vehicle usage patterns in each region, and generate vehicle demand prediction results for different time periods, including the following steps: S31. Extract the vehicle distribution heat characteristics of each region based on the vehicle real-time position coordinates and driving trajectory data respectively, calculate the idle rate, failure rate, and real-time usage frequency of the vehicle based on the running status, usage status, and fault signals of the vehicle, and generate an environment correction coefficient based on the temperature data and humidity data of the vehicle environment; 1. Vehicle distribution heat characteristic extraction: Divide the city into grid regions, count the stay duration and occurrence frequency of vehicles in each region; use the kernel density estimation algorithm to generate a heat distribution map, and the heat value (0~1) of each region reflects the vehicle aggregation degree.

[0029] 2. The idle rate is the proportion of vehicles that are not occupied and not scheduled in the region; the failure rate is the proportion of vehicles reporting fault signals in the region; the real-time usage frequency is the number of times the vehicle is used per unit time.

[0030] 3. Normalize the detected temperature data into a temperature influence factor, normalize the humidity data into a humidity influence factor, and obtain the environment correction coefficient after linear weighting. The weight coefficient of the linear weighting is set according to the local climate characteristics. For example, at 14:00 on a certain day in summer in Nanning, Guangxi, the temperature is 38°C and the humidity is 90%. At this time, the weights corresponding to the temperature and humidity influence factors are 0.6 and 0.4, and the calculated temperature and humidity influence factors are 0.76 and 0.9 respectively. Then the correction coefficient at this time is 0.816.

[0031] S32. Use the vehicle distribution heat characteristics, idle rate, failure rate, and real-time usage frequency as input features and input them into an improved LSTM model. Among them, the environment correction coefficient adjusts the memory weight of the improved LSTM model for temperature and humidity sensitive time periods by embedding the gating unit of the improved LSTM model; Taking the vehicle distribution heat characteristics, idle rate, failure rate, and real-time usage frequency determined above as input features, based on the standard LSTM, an environmental correction coefficient is embedded in the forget gate to dynamically adjust the historical memory weight. The specific implementation is as follows: S321. Map the temperature data and humidity data into temperature influence factors and humidity influence factors respectively, and generate an environmental correction coefficient through linear weighting; S322. Concatenate the environmental correction coefficient with the historical hidden state corresponding to the gating unit to generate a temperature and humidity sensitive gating weight; S323. In the preset high-temperature and high-humidity sensitive period, reduce the forgetting rate of the gating weight to a preset percentage range to enhance the memory strength of the improved LSTM model for the vehicle usage patterns in the same historical period.

[0032] S33. Output the predicted results of the vehicle demand in each area within the preset time.

[0033] Concatenate the environmental correction coefficient with the historical hidden state of the forget gate, and calculate the output of the forget gate through a weight matrix and bias. The preset high-temperature and high-humidity sensitive period here is set as the sensitive period corresponding to when the temperature is greater than 35% and the humidity is greater than 80%. In the sensitive period, multiply the output of the forget gate by an attenuation factor to reduce the forgetting rate of the gating weight to a preset percentage range, which can be set to 50% here; after reducing the forgetting rate, the model retains more historical data in the same period (such as the vehicle usage pattern in the same period last year), enhancing the adaptability to demand fluctuations in the hot and humid environment. Input the final hidden state of the improved LSTM into the fully connected layer to output the predicted demand values for each area in the next T (24) hours, that is, output the predicted demand for each area.

[0034] The above embodiment improves the LSTM model through environmental correction, making the improved model more suitable for scenarios with large fluctuations in temperature and humidity changes, and effectively improving the accuracy of the demand prediction model.

[0035] Embodiment 4 In the present invention, the battery management unit generates a battery state analysis result including remaining life, charging line, and charging time based on battery state data, which respectively include the following steps: S411. Obtain the battery internal resistance change rate, capacity attenuation rate, and environmental temperature deviation value based on the battery state data, where the environmental temperature deviation value is the absolute difference between the environmental temperature where the battery is located and the preset reference temperature; S412. Establish a battery health state attenuation rate model through the correlation analysis of the battery internal resistance change rate, capacity attenuation rate, and temperature deviation value; S413. Calculate the remaining duration of the cycle life based on the difference between the cumulative cycle count of the battery and the preset cycle life threshold; calculate the remaining duration of the capacity life based on the difference between the current health state index and the preset capacity attenuation threshold, and combine the attenuation rate model; S414. Compare the remaining duration of the cycle life and the remaining duration of the capacity life, and select the smaller value as the target prediction result of the remaining life.

[0036] Specifically, the battery internal resistance change rate measures the battery AC impedance periodically through the BMS chip, that is, calculates the internal resistance growth rate per unit time; the capacity attenuation rate is based on the charge-discharge cycle data to calculate the capacity attenuation percentage (such as the capacity drops by 2% per 100 cycles); the ambient temperature deviation value is based on the real-time temperature collected by the ambient temperature sensor, and is the absolute difference from the reference temperature (25°C) (such as the deviation value is 10°C at 35°C). Use multiple linear regression analysis to establish the relationship equation between the internal resistance change rate, the capacity attenuation rate and the temperature deviation value, and calculate the attenuation rate.

[0037] Here, the preset cycle life threshold is set according to the battery type, and the preset capacity attenuation threshold is the industry general standard. For example, when the capacity drops to 80% of the initial value, the battery is determined to fail. The final remaining life takes the smaller value of the cycle life and the capacity life to ensure conservative prediction.

[0038] S421. Divide the charging priority levels according to the remaining power of the vehicle and the battery health state; S422. Based on the real-time position of the vehicle, the distribution of charging stations and the real-time traffic data obtained through the third-party map interface, with the shortest driving time, the highest charging efficiency and the lowest battery health loss as the optimization goals, use the linear weighted method to allocate weights to multiple objectives and generate the optimal charging path; S423. Analyze and update the charging path in real time according to the traffic state change and the availability of charging piles, and push the updated charging path to the dispatching management terminal.

[0039] Specifically, according to the remaining power of the vehicle (low power first) and the battery health state (low capacity first), the vehicle is divided into three levels: emergency charging (power < 20%), regular charging (20% - 50%), and delayed charging (> 50%). The dispatching management terminal accesses the Gaode Map API to obtain real-time traffic data, combines the distribution of charging station locations, and takes the shortest driving time, the highest charging pile efficiency, and the smallest battery health loss as the goals, and uses the linear weighted method to allocate weights to generate the global optimal path. Detect the traffic state change and the availability of charging piles every 5 minutes, automatically re-plan the path and push it to the dispatching terminal to ensure the timeliness of dispatching.

[0040] S431. Dynamically calculate the basic charging duration according to the required charging amount, the charging pile power, and the battery health status, and compensate and correct the charging duration in combination with the ambient temperature data; S432. Based on the regional vehicle peak hours output by the demand prediction unit and the time-of-use electricity price data obtained through the grid server interface, select the time periods with low demand and low electricity price as the recommended charging windows; S433. When it is detected that the battery temperature is abnormal or the health status is lower than the safety threshold, automatically insert a cooling interval and extend the charging duration.

[0041] Specifically, generate the basic charging duration according to the required charging amount, the charging pile power, and the battery health status. For example, slow charging takes 2 hours and fast charging takes 1 hour. When the ambient temperature is higher than 25 °C, the charging duration increases by 10% for every 5 °C increase; when it is lower than 10 °C, the duration increases by 5%.

[0042] Here, the low-demand time periods are the time periods other than the regional vehicle peak hours output by the demand prediction unit (such as 7-9 am and 5-7 pm), such as 1-3 pm in the afternoon and 11 pm-1 am at night; the low-electricity-price time periods obtain the time-of-use electricity price data through the grid interface, and the time periods when the electricity price is lower than the preset threshold (such as 0.5 yuan / kWh), such as the valley electricity period from 0:00 to 8:00. Finally, select the intersection time periods that meet both low demand and low electricity price (such as 1-3 am).

[0043] Here, the temperature safety threshold is set as the battery temperature ≥ 50 °C (triggering forced cooling) or ≤ 0 °C (triggering preheating); the health status safety threshold is set at 70%. When it is less than 70%, immediately stop using and replace the battery; when it is detected that the temperature or capacity exceeds the limit, automatically insert a 15-minute cooling interval, extend the charging duration, and push an alarm to the dispatching terminal at the same time.

[0044] The above embodiments improve the accuracy of the prediction results by quantifying the comprehensive influence of internal resistance, capacity, and temperature on the attenuation rate, and combining the cycle and capacity life thresholds; dynamically optimize the path based on real-time data, balance time, efficiency, and battery health, and adapt to areas with relatively complex road networks; combine demand prediction and time-of-use electricity price, complete charging during low operating cost periods, and ensure battery reliability through safety thresholds.

[0045] Embodiment Five The dispatching management terminal of the present invention is used to generate a dispatching path strategy targeting dispatching cost, user waiting cost, and battery replacement cost according to the vehicle demand prediction result, the battery status analysis result, and the vehicle positioning data, including the following: The objective function of the dispatching path strategy is as follows: The expression of the objective function: Among them: is the vehicle number, is the total number of shared electric two-wheel vehicles, is the area number, is the total number of areas, is the spare battery number, is the total number of available batteries, is the unit distance scheduling cost, is the vehicle scheduled to distance, is the unit time congestion cost, is the vehicle scheduled to real-time traffic congestion time, obtained through the Amap API, is a binary variable. If the vehicle is scheduled from to area , then , otherwise . is the unit cost of user waiting, is the area predicted demand, is the area initial number of vehicles, is the single battery replacement cost, is the correction factor for humid and hot climate. The battery wears out faster in high-temperature and high-humidity environments, is a binary variable. If the vehicle uses the spare battery , then , otherwise .

[0046] The constraint conditions include vehicle allocation constraints, area capacity constraints, battery health constraints, time window constraints, high-demand area service constraints, and charging pile competition constraints. The expressions are as follows: Vehicle allocation constraints: Where: is whether the vehicle is scheduled. Each vehicle is only assigned to one area to avoid repeated scheduling; Area capacity constraints: Where: is the maximum parking capacity of area to prevent vehicle accumulation in the area; Battery health constraints: Where: is the vehicle The battery health index, is the health threshold, which forces the replacement of low-health batteries to avoid mid-course failures; Time window constraint: where: is the vehicle scheduled to area estimated time, is the maximum allowable scheduling time to ensure scheduling timeliness and avoid delays; High-demand area service constraint: where: is the demand threshold, is the minimum service rate to ensure a demand satisfaction rate of over 80% in popular areas; where: is the maximum number of vehicles that a single charging pile can serve, to avoid overcrowding at the same charging pile and causing queuing delays.

[0047] The above high-complexity constraints such as regional capacity constraints and charging pile competition constraints are relaxed into the objective function through Lagrange multipliers; the original problem is decomposed into two independent sub-problems of vehicle scheduling and battery replacement, and solved separately; updated through the sub-gradient method to gradually approach the optimal solution; terminated when the change rate of the objective function is less than the threshold or the maximum number of iterations is reached.

[0048] The above embodiments effectively reduce the computational complexity and are suitable for large-scale vehicle scheduling scenarios such as Nanning, Guangxi; by dynamically adjusting the multiplier weights, they can flexibly handle emergencies such as traffic congestion. They can achieve low-cost and high-response intelligent scheduling, significantly improving vehicle utilization and user experience.

[0049] Embodiment Six In this embodiment, the user terminal 4 includes a positioning and query module 401 for sending a vehicle real-time position query request to the cloud computing platform, receiving and displaying the positioning data of the target vehicle; a navigation and reservation module 402 for initiating a vehicle locking request to the cloud computing platform according to the reservation instruction input by the user and the navigation path planning result, and receiving the reservation confirmation information and vehicle status information returned by the cloud computing platform; a status feedback module 403 for collecting the feedback data of the user on the vehicle usage status and uploading the feedback data to the cloud computing platform; wherein, the navigation and reservation module generates the navigation path from the user's current position to the target vehicle by calling the third-party map service interface, and the vehicle locking request includes the vehicle number, reservation time, and user authentication information.

[0050] The user sends a real-time location query request to the cloud computing platform through the location query module. After receiving the request, the platform obtains the GPS coordinates and status information of the target vehicle from the in-vehicle terminal and returns them. After the user terminal parses the data, it dynamically displays the vehicle location on the map interface; the navigation reservation module calls the Amap API to generate the optimal path from the user's current location to the target vehicle. After the user confirms the reservation, the terminal encapsulates the encrypted vehicle number, reservation timestamp, and user identity token into a locking request and sends it to the cloud platform. After verifying the legitimacy of the token, the cloud platform locks the vehicle and returns the reservation success status and remaining battery power; after use, the status feedback module collects the text or image data submitted by the user through a preset form, compresses and encrypts it, and then uploads it to the cloud platform database, triggering subsequent processing by the battery management unit or the scheduling decision unit, thus forming a closed-loop service link.

[0051] It can be understood that those skilled in the art can, under the guidance of the above embodiments, combine various implementation manners in the above various embodiments to obtain technical solutions of various implementation manners.

[0052] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An intelligent scheduling system for shared electric two-wheel vehicles, characterized in that, Including: An in-vehicle terminal, a cloud computing platform, a dispatching management terminal, and a user terminal; the cloud computing platform is respectively connected to the in-vehicle terminal, the dispatching management terminal, and the user terminal; The in-vehicle terminal is used to collect vehicle positioning data, vehicle status data, vehicle environment data, and battery status data in real time, and upload them to the cloud computing platform; The cloud computing platform includes: A demand prediction unit, which analyzes the vehicle usage patterns in each region based on the vehicle positioning data, vehicle status data, and vehicle environment data, and generates a vehicle demand prediction result for different time periods; A battery management unit, which generates a battery status analysis result including remaining life, charging line, and charging time based on the battery status data; A wireless transmission unit, which is used to push vehicle positioning data, vehicle status data, and battery status data to the user terminal, and provide the vehicle demand prediction result, battery status analysis result, and vehicle positioning data to the dispatching management terminal; The dispatching management terminal is used to generate a dispatching path strategy targeting dispatching cost, user waiting cost, and battery replacement cost based on the vehicle demand prediction result, battery status analysis result, and vehicle positioning data; The user terminal is used to provide vehicle real-time location query, vehicle reservation based on navigation, and vehicle usage status feedback services.

2. The intelligent dispatching system of the shared electric two-wheeler according to claim 1, characterized in that, The in-vehicle terminal includes: A positioning module, which is used to obtain vehicle real-time position coordinates and driving trajectory data; A vehicle status sensing module, which is used to detect the running status, usage status, and fault signals of the vehicle; An environment monitoring module, which is used to collect temperature data and humidity data of the vehicle environment; A battery status monitoring module, which is used to collect battery remaining power and health status data; A main controller, which is respectively connected to the positioning module, battery status monitoring module, vehicle status sensing module, and environment monitoring module, and is used to integrate and process multi-source data and generate transmission data; A data transmission module, which is used to upload the transmission data to the cloud computing platform.

3. The intelligent dispatching system for shared electric two-wheel vehicles according to claim 2, characterized in that, The analysis of the vehicle usage patterns in each region based on the vehicle positioning data, vehicle status data, and vehicle environment data, and the generation of a vehicle demand prediction result for different time periods includes: Respectively extracting the vehicle distribution heat characteristics in each region based on the vehicle real-time position coordinates and driving trajectory data, calculating the vehicle idle rate, failure rate, and real-time usage frequency based on the running status, usage status, and fault signals of the vehicle, and generating an environment correction coefficient based on the temperature data and humidity data of the vehicle environment; Taking the vehicle distribution heat characteristics, idle rate, failure rate, and real-time usage frequency as input features and inputting them into an improved LSTM model, where the environment correction coefficient adjusts the memory weights of the improved LSTM model for temperature- and humidity-sensitive time periods by being embedded in the gated unit of the improved LSTM model; Outputting the vehicle demand prediction result in each region within a preset time.

4. The intelligent scheduling system of the shared electric two-wheeler according to claim 3, wherein, The environment correction coefficient adjusts the memory weights of the improved LSTM model for temperature- and humidity-sensitive time periods by being embedded in the gated unit of the improved LSTM model, including: Map the temperature data and humidity data to a temperature impact factor and a humidity impact factor respectively, and generate an environmental correction coefficient through linear weighting; Concatenate the environmental correction coefficient with the historical hidden state corresponding to the gating unit to generate a temperature and humidity sensitive gating weight; During a preset high temperature and high humidity sensitive period, reduce the forgetting rate of the gating weight to a preset percentage range to enhance the memory strength of the improved LSTM model for the vehicle usage patterns in the same historical period.

5. The intelligent dispatching system for shared electric two-wheel vehicles according to claim 2, characterized in that, The generating of the battery state analysis result including remaining life, charging line and charging time based on the battery state data includes: Obtain the battery internal resistance change rate, capacity attenuation rate and environmental temperature deviation value based on the battery state data, where the environmental temperature deviation value is the absolute difference between the ambient temperature where the battery is located and the preset reference temperature; Establish a battery health state decay rate model through the correlation analysis of the battery internal resistance change rate, capacity attenuation rate and temperature deviation value; Calculate the remaining duration of the cycle life based on the difference between the cumulative cycle times of the battery and the preset cycle life threshold; calculate the remaining duration of the capacity life based on the difference between the current health state index and the preset capacity attenuation threshold, and combine the decay rate model; Compare the remaining duration of the cycle life and the remaining duration of the capacity life, and select the smaller value as the target prediction result of the remaining life.

6. The intelligent dispatching system of the shared electric two-wheeler according to claim 5, characterized in that, The generating of the battery state analysis result including remaining life, charging line and charging time based on the battery state data further includes: Divide the charging priority levels according to the remaining power of the vehicle and the battery health state; Based on the real-time position of the vehicle, the distribution of charging stations and the real-time traffic data obtained through the third-party map interface, with the shortest driving time, the highest charging efficiency and the lowest battery health loss as the optimization objectives, use the linear weighting method to allocate weights to multiple objectives and generate the optimal charging path; Analyze and update the charging path in real time according to the traffic state change and the availability of charging piles, and push the updated charging path to the scheduling management terminal.

7. The intelligent scheduling system of the shared electric two-wheeler according to claim 6, wherein The generating of the battery state analysis result including remaining life, charging line and charging time based on the battery state data further includes: Dynamically calculate the basic charging duration according to the charging amount required, the charging pile power and the battery health state, and compensate and correct the charging duration in combination with the environmental temperature data; Based on the peak vehicle usage period output by the demand prediction unit and the time-of-use electricity price data obtained through the power grid server interface, select the period with low demand and low electricity price as the recommended charging window; When it is detected that the battery temperature is abnormal or the health state is lower than the safety threshold, automatically insert a cooling interval and extend the charging duration.

8. The intelligent dispatching system of the shared electric two-wheeler according to claim 1, characterized in that, The generating of the scheduling path strategy with the scheduling cost, user waiting cost and battery replacement cost as the objectives includes: The expression of the objective function: Wherein: is the vehicle number, is the total number of shared electric two-wheel vehicles, is the area number, is the total number of areas, is the spare battery number, is the total number of available batteries, is the scheduling cost per unit distance, is the vehicle scheduled to area distance, is the congestion cost per unit time, is the vehicle scheduled to area real-time traffic congestion time, obtained through the Amap API, is a binary variable. If the vehicle is scheduled from to area , then otherwise . is the unit cost for user waiting, is the predicted demand of area , is the initial number of vehicles in area , is the single battery replacement cost, is the correction factor for humid and hot climate. The battery wears out faster in high temperature and high humidity environment, is a binary variable. If the vehicle uses the spare battery , then otherwise .

9. The intelligent scheduling system of the shared electric two-wheeler according to claim 8, wherein The constraint conditions of the scheduling path strategy include vehicle allocation constraint, regional capacity constraint, battery health constraint, time window constraint, high demand area service constraint and charging pile competition constraint.

10. The intelligent scheduling system of the shared electric two-wheeler according to claim 1, wherein, The user terminal includes: A positioning query module, configured to send a vehicle real-time position query request to the cloud computing platform, and receive and display the positioning data of the target vehicle; A navigation reservation module, which is used to initiate a vehicle locking request to the cloud computing platform according to the reservation instruction input by the user and the navigation path planning result, and receive the reservation confirmation information and vehicle status information returned by the cloud computing platform; A status feedback module, which is used to collect the feedback data of the user on the vehicle usage status and upload the feedback data to the cloud computing platform; Among them, the navigation reservation module generates a navigation path from the user's current location to the target vehicle by calling a third-party map service interface, and the vehicle locking request includes a vehicle number, a reservation time, and user authentication information.

Images