Electric quantity dynamic early warning system and method for mobile storage and charging robot

Abstract

The invention discloses an electric quantity dynamic early warning system and method for a mobile storage and charging robot, and the system comprises an electric quantity prediction module which is used for obtaining the consumed electric quantity of a standby robot through a constructed electric quantity consumption prediction model, and calculating the predicted output electric quantity of the standby robot; the task management module is used for acquiring order information and performing task allocation according to the required electricity consumption of the order information and the predicted output electricity quantity of the standby robot; the dynamic early warning module is used for judging whether charging early warning needs to be started or not according to the constructed order prediction model and the predicted output electric quantity extreme value of the standby robot; if yes, charging early warning is started, and the standby robot is charged according to the charging early warning; and if not, closing the charging early warning, and charging the standby robot in the valley price time period. The invention relates to the technical field of new energy storage and charging, and solves the technical problems that in an electric quantity early warning system of an existing mobile storage and charging robot, an early warning mechanism is not intelligent, and a charging management method is single.

Description

Technical Field

[0001] The present invention belongs to the field of new energy storage and charging, and relates to an intelligent dynamic early warning technology for electric quantity, and specifically to a dynamic early warning system and method for electric quantity for a mobile storage and charging robot. Background Art

[0002] With the popularity and development of electric vehicles, mobile charging robots are gradually gaining widespread attention as a flexible and convenient charging solution. In the current technological environment, it is crucial to ensure that electric vehicles can obtain charging services in a timely manner.

[0003] However, the existing charging system has some obvious shortcomings. On the one hand, there is a lack of dynamic early warning mechanism for the power of mobile storage and charging robots. The current system can usually only simply display the remaining power of the robot, and cannot accurately predict the power consumption of the robot in different working environments and when the power may be insufficient. This often leads to the problem of being unable to provide charging services for electric vehicles in a timely manner due to the exhaustion of the robot in practical applications, which affects the user experience and reduces the reliability of the charging service. On the other hand, the existing charging management method is relatively simple and cannot be comprehensively analyzed and dynamically adjusted according to factors such as actual order demand, traffic flow changes, and different date types. For example, during peak hours, order demand and traffic flow are usually large, and the power consumption of mobile storage and charging robots will also increase accordingly, but the existing system often cannot give early warnings and reasonably dispatch robots for charging or task allocation, resulting in possible delays and confusion in charging services. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a dynamic power warning system and method for a mobile storage and charging robot, which is used to solve the technical problems of the unintelligent warning mechanism and the single charging management method in the power warning system of the existing mobile storage and charging robot.

[0005] To achieve the above objectives, the first aspect of the present invention provides a dynamic early warning system for electric quantity of a mobile storage and charging robot, comprising:

[0006] Power prediction module: used to obtain the power consumption of several standby robots through a power consumption prediction model, and obtain the predicted output power of several standby robots according to the power consumption; wherein the power consumption prediction model is constructed based on a deep learning algorithm;

[0007] Task management module: used to obtain order information; wherein the order information includes: the location of the power to be used, the required power consumption; and assign tasks to several standby robots according to the order information and the predicted output power of several standby robots;

[0008] Dynamic warning module: used to determine whether it is necessary to turn on the charging warning based on the order prediction model and the predicted output power maximum values ​​of several standby robots; if yes, the charging warning is turned on and several standby robots are charged according to the charging warning; if no, the charging warning is turned off and the standby robots are charged during the valley price period; the maximum value includes the maximum value and the minimum value, and the order prediction model is built based on the deep learning algorithm;

[0009] It should be noted that the required power consumption in the order information can be obtained by obtaining the total battery capacity of the vehicle based on the vehicle model entered by the user, and can be calculated based on the total battery capacity and the current power percentage entered by the user.

[0010] Based on the above technical modules, dynamic early warning and efficient management of the power of mobile storage and charging robots are realized. By using deep learning algorithms, the output power of standby robots is intelligently predicted, thereby optimizing task allocation and power usage and avoiding resource waste; through the order prediction model, the power demand in the next time period is predicted in advance, and the need for early charging is determined based on the remaining power of the existing robots. In non-emergency situations, the valley price time period is selected to charge the standby robots, so as to meet the task demand while effectively reducing the charging cost.

[0011] Furthermore, the power consumption prediction model is constructed based on a deep learning algorithm, including:

[0012] A1, collect battery capacity attenuation rate, working environment temperature data, distance data, and power consumption data of several robots to obtain a power prediction data set; wherein the distance data represents the driving distance of the robot to complete a charging task, and the power consumption calculation formula is: power consumption = remaining power - output power, and the output power is the required charging amount on the order information;

[0013] A2, after preprocessing the power consumption prediction data set, dividing it according to a preset ratio to obtain a training set and a validation set of a power consumption prediction model;

[0014] A3, construct a convolutional neural network model based on a deep learning algorithm, use the power consumption prediction model training set and verification set to train and verify the convolutional neural network model, and iteratively optimize the convolutional neural network model according to the optimization algorithm to obtain a power consumption prediction model whose input is the battery capacity decay rate, remaining power, working environment temperature, and movement distance of the standby robot, and whose output is the predicted power consumption of the standby robot;

[0015] It is worth noting that the movement distance refers to the movement distance of the robot from the starting point to the order release point and returning to the nearest charging station after completing the order task.

[0016] Furthermore, the step of allocating tasks to the standby robots according to the order information and the predicted output power of the standby robots includes:

[0017] B1, extract the predicted output power of several standby robots and the required power consumption in the order information;

[0018] B2, judge whether several standby robots can complete the current order based on the predicted output power of several standby robots and the power consumption required for the order; if yes, dispatch several working robots according to the predicted output power of several standby robots; if not, start the task warning, and charge and assign work to several standby robots according to the task warning.

[0019] It should be noted that if there is a standby robot or a combination of standby robots whose predicted output power is greater than the power required for the order, the current order can be completed; otherwise, no standby robot can complete the current order.

[0020] The power consumption prediction model can accurately predict the power consumption of the robot under different conditions, providing a reliable basis for subsequent task allocation and charging decisions, effectively avoiding task interruptions or equipment damage caused by insufficient power, and improving overall operational efficiency.

[0021] Furthermore, dispatching the standby robots according to the predicted output power of the plurality of standby robots includes:

[0022] B21, arranging the predicted output power of the plurality of standby robots in descending order and storing them in an output power list R; wherein each element in the list R contains the ID of the standby robot and the corresponding predicted output power;

[0023] B22, according to formula C j =R j -E Calculate the difference between the predicted output power and the required power consumption of each standby robot C j ; where j = 1, 2, ..., n, n represents the number of standby robots, R j represents the predicted output power of the jth standby robot, and E represents the required power consumption;

[0024] B23, judgment C j Is it greater than 0? If yes, select C j If the standby robot is greater than 0 and corresponds to the minimum value, the working robot is obtained; otherwise, a working robot combination is generated according to the predicted output power of several standby robots.

[0025] Furthermore, the step of generating a working robot combination according to the predicted output power of a plurality of standby robots includes:

[0026] B23-1, initialize the value of i to 2; i represents the selection of i standby robots;

[0027] B23-2, perform non-repeated permutations and combinations on the i standby robots to obtain a list of standby robot combinations R si (i=2, 3, ..., n);

[0028] B23-3, Calculate the robot combination list R si The predicted output power of each combination is summed up to obtain several combined power, and it is determined whether the values ​​of the several combined power are greater than the power required for the order; if yes, several combinations of robots to be worked are obtained and the process jumps to B23-5; if no, the process jumps to B23-4;

[0029] B23-4, determine whether the value of i is equal to n; if yes, start the task warning, and charge and assign work to several standby robots according to the task warning; if no, increase the value of i by 1 and jump to B23-3;

[0030] B23-5, filter out the waiting robot combination with the lowest combined power, obtain the working robot combination, and assign the current order task to several standby robots in the working robot combination according to the ID.

[0031] Furthermore, the charging and work allocation of several standby robots according to the task warning includes:

[0032] C1: Select the standby robot with the highest predicted power output as the priority robot and dispatch it to the order release point to work. The remaining standby robots turn on the task alarm to remind the staff to charge.

[0033] C2, use the priority robot to charge the order publisher, record the real-time remaining power of the priority robot, and calculate the power according to formula R w =120%×R p -(R1-R2)] Calculate the handover warning power R w ; Among them, R p represents the predicted power consumption of the priority robot, R1 represents the remaining power before the priority robot sets out, and R2 represents the predicted output power of the priority robot;

[0034] C3, determine whether the real-time remaining power of the priority robot is equal to the handover predicted power; if yes, issue a handover signal, and select the standby robot with the highest predicted output power among several standby robots being charged, and mark it as the priority robot after going to the order release point, and repeat C2-C3 until the current order task is completed and the order completion signal is sent; if not, jump to C2.

[0035] During the task allocation process, the most suitable robot is selected to perform the task by calculating the difference between the predicted output power of the standby robot and the power required for the order; when no single robot can meet the order requirements, the best robot combination plan is generated according to the predicted output power of each standby robot, which can minimize power waste and improve the efficiency of robot use; at the same time, when the remaining power of the robot is not enough to complete the order task, the task warning is turned on in time, and charging and work handover are carried out according to the preset strategy to ensure the continuity and stability of the service.

[0036] Furthermore, the order prediction model is constructed based on the deep learning algorithm, including:

[0037] D1, count and collect historical hourly traffic data within the service range of the mobile storage and charging robot station, the maximum and minimum power demand of historical orders in several hourly periods, the date type of the corresponding day and weather data, and obtain several traffic flow data, several maximum power demand data, several minimum power demand data, several date type data and several weather data; among which, the date type includes holidays and weekdays, and the weather data includes temperature and rainfall;

[0038] D2, after preprocessing the plurality of vehicle flow data, the plurality of maximum power demand data, the plurality of minimum power demand data and the plurality of date type data, the plurality of vehicle flow data are divided according to a preset ratio to obtain a training set and a validation set of an order prediction model;

[0039] D3, a recurrent neural network model is constructed based on the deep learning model. The training set and validation set of the order prediction model are used to train and validate the time series network model, and the recurrent neural network model is iteratively optimized according to the optimization algorithm. The input is the traffic volume, date type, weather data, and maximum and minimum power demand of the previous hour, and the output is the predicted maximum and minimum power demand of the next hour.

[0040] By collecting statistics and multi-dimensional information such as historical traffic data, order power requirements, date types and weather data, an order prediction model based on deep learning algorithms is built, which can accurately predict power demand in future time periods and help prepare for power reserves and scheduling in advance.

[0041] Furthermore, judging whether it is necessary to start charging warning according to the order prediction model and the predicted maximum output power values ​​of the plurality of standby robots includes:

[0042] E1, obtaining the maximum predicted output power and the minimum predicted output power according to the predicted output power of several standby robots;

[0043] E2, obtains the maximum predicted power demand and the minimum predicted power demand for the next hour according to the order prediction model;

[0044] E3, determine whether the maximum predicted output power is greater than the maximum predicted required power; if yes, jump to E4; if no, start the charging warning, and charge the standby robot with the maximum predicted output power according to the charging warning;

[0045] E4, determine whether the minimum predicted output power is greater than the minimum predicted required power; if yes, turn off the charging warning; if no, turn on the charging warning, and charge the standby robot with the minimum predicted output power according to the charging warning.

[0046] Based on the order forecast results and the predicted output power of the standby robot, it is intelligently determined whether it is necessary to turn on the charging warning to ensure that the robot has enough power to meet user needs during peak hours or in bad weather conditions, avoiding service interruptions caused by insufficient power, which helps to improve the operational efficiency and user satisfaction of the mobile storage and charging robot. At the same time, reasonable power management also helps to reduce operating costs and improve overall economic benefits.

[0047] Preferably, a dynamic power warning system for a mobile charging and storage robot also includes a voltage warning module, which is used to obtain the voltage value of the working robot and start the voltage warning when the voltage value reaches a voltage threshold.

[0048] A second aspect of the present invention provides a method for dynamic early warning of power level of a mobile storage and charging robot, comprising:

[0049] S1, using a deep learning algorithm to build a power consumption prediction model, obtain the power consumption of several standby robots according to the power consumption prediction model, and obtain the predicted output power of several standby robots according to the power consumption;

[0050] S2, obtaining order information, and assigning tasks to the standby robots according to the order information and the predicted output power of the standby robots; wherein the order information includes the standby power location and the required power consumption;

[0051] S3, using a deep learning algorithm to build an order prediction model, and judging whether it is necessary to turn on the charging warning according to the order prediction model and the predicted output power maximum values ​​of several standby robots; if yes, turning on the charging warning, and charging several standby robots according to the charging warning; if no, turning off the charging warning, and charging the standby robots during the valley price period; wherein the maximum value includes the maximum value and the minimum value;

[0052] S4, obtaining the voltage value of the working robot, and starting a voltage warning when the voltage value reaches a voltage threshold.

[0053] Compared with the prior art, the present invention has the following beneficial effects:

[0054] The predicted output power of several standby robots is obtained through the power consumption prediction model, and the power consumption required for the order is obtained according to the order information. The power difference between the required power consumption and the predicted output power of several standby robots is calculated, and the most suitable robot or combination is selected to perform the task through sorting and combination calculation methods to minimize power waste and improve the use efficiency of the robot; determine whether it is necessary to turn on the charging warning according to the order prediction model, and select the valley price time period to charge the standby robot when the charging warning is turned off, which helps to make power reserves and scheduling preparations in advance, optimize charging arrangements, and reduce charging costs; add a voltage warning module, and start the warning when the voltage value of the working robot reaches the voltage threshold, reminding the staff to replace the battery, ensure the performance and safety of the robot, and avoid affecting the service quality due to voltage problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0056] Figure 1 A schematic diagram of a flow chart of a method for dynamic early warning of electric quantity of a mobile storage and charging robot provided by the present invention;

[0057] Figure 2 A schematic diagram of a process of obtaining a working robot in a task management module provided by the present invention;

[0058] Figure 3 A schematic diagram of a process for obtaining a working robot combination in a task management module provided by the present invention;

[0059] Figure 4 A schematic diagram of the process of charging and work handover in the task management module provided by the present invention. DETAILED DESCRIPTION

[0060] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0061] See also Figure 1-Figure 4 The first embodiment of the present invention provides a dynamic early warning system for electric quantity of a mobile storage and charging robot, comprising:

[0062] Power prediction module: used to obtain the power consumption of several standby robots through a power consumption prediction model, and obtain the predicted output power of several standby robots based on the power consumption; wherein the power consumption prediction model is constructed based on a deep learning algorithm;

[0063] Task management module: used to obtain order information; wherein the order information includes: the location of the power to be used, the required power consumption; and assign tasks to several standby robots according to the order information and the predicted output power of several standby robots;

[0064] Dynamic warning module: used to determine whether it is necessary to turn on the charging warning based on the order prediction model and the predicted output power maximum values ​​of several standby robots; if yes, the charging warning is turned on and several standby robots are charged according to the charging warning; if no, the charging warning is turned off and the standby robots are charged during the valley price period; the maximum value includes the maximum value and the minimum value, and the order prediction model is built based on the deep learning algorithm;

[0065] It should be noted that the required power consumption in the order information can be obtained by obtaining the total battery capacity of the vehicle based on the vehicle model entered by the user, and can be calculated based on the total battery capacity and the current power percentage entered by the user.

[0066] In this embodiment, the power consumption prediction model is constructed based on a deep learning algorithm. Specifically, first, in order to construct an accurate power consumption prediction model, it is necessary to comprehensively and meticulously collect historical data, including the battery capacity decay rate of several robots, the working environment temperature of several robots, the driving distance of the robots in each charging task, and the required charging amount obtained based on several order information, and according to the calculation formula of power consumption: power consumption = remaining power - output power, the power consumption of each task is obtained, and the power prediction data set is obtained;

[0067] Then, the original power forecasting data set is preprocessed, including removing outliers and noise, normalizing or standardizing the data to eliminate the influence of different dimensional features and improve the model training efficiency; and the data set is divided into a training set and a validation set according to a preset ratio, such as 7:3;

[0068] Since convolutional neural networks have obvious advantages in processing complex data features, they are very suitable for capturing the nonlinear relationship between battery capacity decay rate, working environment temperature, movement distance and power consumption. Therefore, we choose to use the convolutional neural network model to build a power consumption prediction model;

[0069] In the model training phase, the model is trained using the training set, and the model parameters are continuously adjusted through the back-propagation algorithm to minimize the prediction error. At the same time, the model is periodically verified using the validation set to monitor the model performance and prevent overfitting. In order to further improve the model performance, an optimization algorithm, such as the Adam optimizer, is used to iteratively optimize the model, and the learning rate and other parameters are dynamically adjusted to accelerate the model convergence and improve the prediction accuracy.

[0070] Finally, a power consumption prediction model is obtained, whose input is the remaining power of the standby robot, battery capacity decay rate, working environment temperature, and movement distance, and whose output is the predicted power consumption of the standby robot. This model provides a basis for subsequent power management and task allocation, and improves the accuracy and reliability of power prediction.

[0071] In order to carry out task allocation and charging management reasonably and efficiently, a task warning module is designed, which can perform refined task allocation based on the predicted output power and order information of several standby robots.

[0072] Specifically, we first need to determine whether the current order can be completed by the predicted output power of several standby robots, that is, according to formula C j =R j -E Calculate the difference between the predicted output power and the required power consumption of each standby robot C j , then judge C j Is it greater than 0? If it is, it means that several standby robots can complete the order task and select C j If the standby robot is greater than 0 and corresponds to the minimum value, the working robot is obtained, and the task is assigned to the working robot for execution; if not, it is necessary to further determine whether the combination of several standby robots is greater than the power consumption required by the order. If so, it means that the current several standby robots can complete the order task, and then further processing is performed to obtain the best robot combination to achieve optimal energy allocation; if not, it means that the current several standby robots cannot complete the order task, and it is necessary to enable task warning and conduct timely charging and work handover;

[0073] In the method of obtaining the best robot combination, similarly, several standby robots with the least number, the smallest combined power and greater than the power required by the order are selected as working robots for allocation; if the sum of all the predicted output power of several current standby robots is less than the power required by the order, it means that the task cannot be completed, then a robot with the largest predicted output power is sent to work first, and the remaining standby robots issue a task warning to remind the staff to charge, and then before the previous working robot is about to return, a handover signal is sent in advance, and the robot with the largest predicted output power among the remaining robots is arranged to go to the order release point to carry out the handover of work;

[0074] Specifically, suppose there are 4 standby robots, namely R1, R2, R3, and R4. The power consumption required for the order is calculated to be 22 kWh. The predicted output power of each standby robot is: R1 = 12 kWh, R2 = 8 kWh, R3 = 9 kWh, and R4 = 7 kWh, which are all less than the required power consumption. Therefore, it is necessary to determine whether there is a combination of standby robots that can complete the order task:

[0075] B23-1: Initialize the value of i to 2;

[0076] B23-2: Perform non-repeating combinations based on i=2 to obtain a list of standby robot combinations:

[0077] Combination 1: (R1, R2); Combination 2: (R1, R3); Combination 3: (R1, R4); Combination 4: (R2, R3); Combination 5: (R2, R4); Combination 6: (R3, R4);

[0078] B23-3: Calculate the total predicted output power of each combination in the robot combination list, and determine whether it is greater than the power required for the order:

[0079] Combination 1: 120+80=20 kWh; Combination 2: 120+90=21 kWh; Combination 3: 12+7=19 kWh; Combination 4: 8+9=17 kWh; Combination 5: 8+7=15 kWh; Combination 6: 9+7=16 kWh, all combinations are less than 22 kWh;

[0080] B23-4: Determine whether i is equal to n (n=4), i=2 is not equal to n, increase the value of i by 1, i=3, jump to B23-2;

[0081] B23-2: Perform non-repeating permutations and combinations based on i=3 to obtain a new list of standby robot combinations:

[0082] Combination 1: (R1, R2, R3); Combination 2: (R1, R2, R4); Combination 3: (R1, R3, R4); Combination 4: (R2, R3, R4);

[0083] B23-3: Calculate the total predicted output power of each combination in the robot combination list, and determine whether it is greater than the power required for the order:

[0084] Combination 1: 12+8+9=29 kWh; Combination 2: 12+8+7=27 kWh; Combination 3: 12+9+7=28 kWh; Combination 4: 8+9+7=24 kWh, all greater than 22 kWh;

[0085] B23-5: Filter out the robot combination with the lowest power consumption, and assign the current order task to several standby robots in the working robot combination according to the ID:

[0086] Combination 4 < Combination 2 < Combination 3 < Combination 1, and the combination with the lowest power is Combination 4. Then the order task is assigned to the R2, R3 and R4 robots in Combination 4.

[0087] If it is assumed that there are 3 standby robots, namely R1, R2, and R3, and the power consumption required for the order is 25 kWh, the predicted output power of each standby robot is: R1 = 7 kWh, R2 = 8 kWh, and R3 = 6 kWh; through the above steps, it can be judged that the predicted output power of all current standby robots is less than the power consumption required for the order, so the task warning will be enabled according to the following steps:

[0088] C1: Select the standby robot with the highest predicted power output as the priority robot, that is, R2 (8 kWh), and dispatch R2 to work at the order release point. The remaining standby robots R1 and R3 start task warnings to remind staff to charge.

[0089] C2: Use R2 to charge the order publisher, record the real-time remaining power of R2, and calculate the power according to the formula R w =120%×R p -(R1-R2)] Calculate the handover warning power R w ; Among them, R p represents the predicted power consumption of the priority robot, R1 represents the remaining power before the priority robot sets out, and R2 represents the predicted output power of the priority robot;

[0090] C3: Determine whether the real-time remaining power of R2 is equal to the predicted power for handover. Once it is reached, select the robot with the highest predicted output power among the charging standby robots. Assuming that R1 has been fully charged at this time, select R1 to go to the order release point and mark it as the new priority robot. Repeat C2-C3 until the current order task is completed and the order completion signal is sent.

[0091] Through the above methods and steps, task allocation and energy management are optimized, unnecessary robot deployment and energy consumption are reduced, and the continuity and efficiency of services are guaranteed.

[0092] At the same time, the system also uses a dynamic early warning module to predict the power demand for orders in the next time period, and arranges charging tasks for the standby robots in advance based on the predicted order demand to avoid insufficient service energy during peak order periods. At the same time, through dynamic charging early warning, it intelligently selects a reasonable charging time period. When the remaining power is sufficient to meet the predicted power demand, the charging early warning is turned off, and charging is carried out during the period of low electricity prices to reduce energy costs.

[0093] Specifically, we first need to build an order prediction model to predict the power demand in the next time period:

[0094] Step 1: Count and collect the historical hourly traffic data within the service range of the mobile storage and charging robot station, the maximum and minimum power demand of historical orders in several hourly periods, the date type of the corresponding day, and weather data, and obtain some traffic data, some maximum power demand data, some minimum power demand data, some date type data, and some weather data. Among them, the date type includes holidays and weekdays, and the weather data includes temperature and rainfall; at the same time, the data is preliminarily cleaned and sorted to remove outliers and erroneous data to ensure the accuracy and reliability of the data;

[0095] Step 2: Preprocess the collected traffic flow data, maximum power demand data, minimum power demand data and date type data, including data normalization and standardization, so that different types of data have similar value ranges and distribution characteristics. Then, divide them according to a preset ratio (for example, 7:3 or 8:2) to obtain the training set and validation set of the order prediction model;

[0096] Step 3: Build a recurrent neural network (RNN) model based on the deep learning model, such as a long short-term memory network (LSTM), determine the structure and parameters of the model, including the number of neurons in the input layer, hidden layer, and output layer, activation function, etc.; initialize the model and set appropriate initial weights and biases;

[0097] Step 4: Use the training set of the order prediction model to train the constructed recurrent neural network model; take the traffic volume, date type, weather data, and maximum and minimum power demand of the previous hour as input, and the corresponding maximum and minimum power demand of the next hour as output targets; use appropriate loss functions, such as mean square error, and optimization algorithms, such as stochastic gradient descent, Adam, etc., to update and optimize model parameters, and gradually converge the model through multiple iterative training to reduce the value of the loss function;

[0098] Step 5: Use the validation set to validate the trained model. Evaluate the performance of the model on the validation set, such as calculating the error indicators between the predicted value and the true value, such as mean absolute error, root mean square error, etc.; Based on the validation results, analyze the shortcomings of the model and adjust and optimize the structure and parameters of the model, such as increasing or decreasing the number of neurons in the hidden layer, adjusting the learning rate, etc.;

[0099] Step 6: Iteratively optimize the recurrent neural network model according to the optimization algorithm. Repeat steps 4 and 5 to continuously improve the performance of the model until the preset performance index or the upper limit of the number of iterations is reached; finally, the optimized order prediction model is obtained, whose input is the traffic volume, date type, weather data, and maximum and minimum power demand of the previous hour, and the output is the predicted maximum and minimum power demand of the next hour;

[0100] After obtaining the predicted maximum power demand and the predicted minimum power demand for the next hour by using the order prediction model, it is determined whether and which robots need to be charged in advance according to the predicted maximum output power of the above-mentioned standby robots;

[0101] Specifically, first, determine whether the maximum predicted output power is greater than the maximum predicted power demand; if so, proceed to the next step of determination; if not, start the charging warning and arrange charging for the standby robot with the maximum predicted output power;

[0102] Then, it is determined whether the minimum predicted output power is greater than the minimum predicted required power. If so, the charging warning is turned off, indicating that there is no need to charge the standby robot in advance temporarily; if not, the charging warning is turned on, and charging is arranged for the standby robot with the minimum predicted output power.

[0103] Through the above method, charging warnings can be activated in time when power shortage may occur, providing advance preparation time for charging arrangements, ensuring that the robots have sufficient power to meet order requirements when needed, and improving the reliability and stability of the service; and charging arrangements can be made for specific standby robots in a targeted manner according to different situations, avoiding unnecessary charging operations for all robots, thereby optimizing resource allocation and reducing energy consumption and operating costs.

[0104] The second embodiment provides a dynamic power warning system for mobile storage and charging robots. Different from the above embodiments, due to the special needs of some users, that is, the need for fast charging, when the working robot has a high temperature and low remaining power due to long-term charging, it will cause its output voltage to be small, thereby affecting the charging efficiency of new energy vehicles. Therefore, the system also includes a voltage warning module, which is used to obtain the voltage value of the working robot, and when the voltage value reaches the voltage threshold, it starts the voltage warning, and assigns tasks to several standby robots according to the order information and the predicted output power of several standby robots, so as to avoid the reduction of charging efficiency caused by insufficient voltage, so as to meet the user's need for fast charging.

[0105] Part of the data in the above formula is calculated by removing the dimension and taking its numerical value. The formula is a formula closest to the actual situation obtained by software simulation of a large amount of collected data; the preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained through simulation of a large amount of data.

[0106] See also Figure 2 The third aspect of the present invention provides a dynamic early warning system for a mobile storage and charging robot, comprising:

[0107] S1, using a deep learning algorithm to build a power consumption prediction model, obtain the power consumption of several standby robots according to the power consumption prediction model, and obtain the predicted output power of several standby robots according to the power consumption;

[0108] S2, obtaining order information, and assigning tasks to the standby robots according to the order information and the predicted output power of the standby robots; wherein the order information includes the standby power location and the required power consumption;

[0109] S3, using deep learning algorithm to build an order prediction model, and judging whether it is necessary to turn on the charging warning according to the order prediction model and the predicted output power maximum values ​​of several standby robots; if yes, turn on the charging warning, and charge several standby robots according to the charging warning; if no, turn off the charging warning, and charge the standby robots during the valley price period; wherein the maximum value includes the maximum value and the minimum value;

[0110] S4, obtaining the voltage value of the working robot, and starting a voltage warning when the voltage value reaches a voltage threshold.

[0111] Working principle of the present invention:

[0112] (1) Power prediction: Use deep learning algorithms to build a power consumption prediction model, use the model to predict the power consumption of the standby robot, and then obtain the predicted output power;

[0113] (2) Task management: Obtain order information, including the location of the waiting power supply and the required power consumption, and assign tasks to the waiting robots based on the order information and the predicted output power, ensuring that the least number of best robots are selected to complete the task;

[0114] (3) Dynamic warning: Use deep learning algorithms to build an order prediction model. The order prediction model is used to predict the maximum order power demand in the next time period. Based on the predicted maximum power demand and the predicted maximum output power of the standby robot, it is determined whether it is necessary to turn on the charging warning. If charging is required, turn on the warning and arrange charging. If not, turn off the warning and charge the robot during the low electricity price period to reduce energy costs.

[0115] (4) Task warning and handover: During the task allocation process, if the current standby robot cannot complete the order task alone or in combination, the task warning will be activated; the robot with the highest predicted power output will be selected to perform the task first, and the remaining robots will prepare for charging. Before the previous working robot is about to return, a handover signal will be sent in advance, and the robot with the highest predicted power output among the remaining robots will be arranged to go to the order release point for work handover.

[0116] (5) Voltage warning: When the working robot is charging the new energy vehicle, the voltage value of the working robot is obtained; when the voltage value reaches the preset voltage threshold, the voltage warning is activated, and tasks are assigned to several standby robots based on the order information and the predicted output power of several standby robots to avoid reduced charging efficiency due to insufficient voltage.

[0117] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A dynamic early warning system for electric quantity of a mobile storage and charging robot, characterized in that: include: Power prediction module: used to obtain the power consumption of several standby robots through a power consumption prediction model, and obtain the predicted output power of several standby robots according to the power consumption; wherein the power consumption prediction model is constructed based on a deep learning algorithm; Task management module: used to obtain order information; wherein the order information includes: the location of the power to be used, the required power consumption; and assign tasks to several standby robots according to the order information and the predicted output power of several standby robots; Dynamic warning module: used to determine whether it is necessary to turn on the charging warning based on the order prediction model and the predicted maximum output power values ​​of several standby robots; if yes, turn on the charging warning and charge several standby robots according to the charging warning; if no, turn off the charging warning and charge the standby robots during the valley price period; the maximum value includes the maximum value and the minimum value, and the order prediction model is built based on the deep learning algorithm.

2. A dynamic early warning system for electric quantity of a mobile storage and charging robot according to claim 1, characterized in that: The power consumption prediction model is built based on a deep learning algorithm and includes: A1, collect battery capacity attenuation rate, working environment temperature data, distance data, and power consumption data of several robots to obtain a power prediction data set; wherein the distance data represents the driving distance of the robot to complete a charging task, and the power consumption calculation formula is: power consumption = remaining power - output power, and the output power is the required charging amount on the order information; A2, after preprocessing the power consumption prediction data set, dividing it according to a preset ratio to obtain a training set and a validation set of a power consumption prediction model; A3, builds a convolutional neural network model based on the deep learning algorithm, uses the power consumption prediction model training set and verification set to train and verify the convolutional neural network model, and iteratively optimizes the convolutional neural network model according to the optimization algorithm to obtain a power consumption prediction model whose input is the battery capacity decay rate, remaining power, working environment temperature, and movement distance of the standby robot, and whose output is the predicted power consumption of the standby robot.

3. A dynamic early warning system for electric quantity of a mobile storage and charging robot according to claim 1, characterized in that: The step of allocating tasks to the standby robots according to the order information and the predicted output power of the standby robots includes: B1, extract the predicted output power of several standby robots and the required power consumption in the order information; B2, judge whether several standby robots can complete the current order based on the predicted output power of several standby robots and the power consumption required for the order; if yes, dispatch several working robots according to the predicted output power of several standby robots; if not, start the task warning, and charge and assign work to several standby robots according to the task warning.

4. A dynamic early warning system for electric quantity of a mobile storage and charging robot according to claim 3, characterized in that: The dispatching of standby robots according to the predicted output power of the plurality of standby robots comprises: B21, arranging the predicted output power of the plurality of standby robots in descending order and storing them in an output power list R; wherein each element in the list R contains the ID of the standby robot and the corresponding predicted output power; B22, according to formula C j =R j -E Calculate the difference between the predicted output power and the required power consumption of each standby robot C j ; where j = 1, 2, ..., n, n represents the number of standby robots, R j represents the predicted output power of the jth standby robot, and E represents the required power consumption; B23, judgment C j Is it greater than 0? If yes, select C j If the standby robot is greater than 0 and corresponds to the minimum value, the working robot is obtained; otherwise, a working robot combination is generated according to the predicted output power of several standby robots.

5. A dynamic early warning system for electric quantity of a mobile storage and charging robot according to claim 4, characterized in that: The method of generating a working robot combination according to the predicted output power of a plurality of standby robots includes: B23-1, initialize the value of i to 2; i represents the selection of i standby robots; B23-2, perform non-repeated permutations and combinations on the i standby robots to obtain a list of standby robot combinations R si (i=2, 3, ..., n); B23-3, Calculate the robot combination list R si The predicted output power of each combination is summed up to obtain several combined power, and it is determined whether the values ​​of the several combined power are greater than the power required for the order; if yes, several combinations of robots to be worked are obtained and the process jumps to B23-5; if no, the process jumps to B23-4; B23-4, determine whether the value of i is equal to n; if yes, start the task warning, and charge and assign work to several standby robots according to the task warning; if no, increase the value of i by 1 and jump to B23-3; B23-5, filter out the waiting robot combination with the lowest combined power, obtain the working robot combination, and assign the current order task to several standby robots in the working robot combination according to the ID.

6. A dynamic early warning system for electric quantity of a mobile storage and charging robot according to claim 1, characterized in that: The charging and work allocation of several standby robots according to the task warning includes: C1: Select the standby robot with the highest predicted power output as the priority robot and dispatch it to the order release point to work. The remaining standby robots turn on the task alarm to remind the staff to charge. C2, use the priority robot to charge the order publisher, record the real-time remaining power of the priority robot, and calculate the power according to formula R w =120%×R p -(R1-R2)] Calculate the handover warning power R w ; Among them, R p represents the predicted power consumption of the priority robot, R1 represents the remaining power before the priority robot sets out, and R2 represents the predicted output power of the priority robot; C3, determine whether the real-time remaining power of the priority robot is equal to the handover predicted power; if yes, issue a handover signal, and select the standby robot with the highest predicted output power among several standby robots being charged, and mark it as the priority robot after going to the order release point, and repeat C2-C3 until the current order task is completed and the order completion signal is sent; if not, jump to C2.

7. A dynamic early warning system for electric quantity of a mobile storage and charging robot according to claim 1, characterized in that: The order prediction model is constructed based on the deep learning algorithm, including: D1, count and collect historical hourly traffic data within the service range of the mobile storage and charging robot station, the maximum and minimum power demand of historical orders in several hourly periods, the date type of the corresponding day and weather data, and obtain several traffic flow data, several maximum power demand data, several minimum power demand data, several date type data and several weather data; among which, the date type includes holidays and weekdays, and the weather data includes temperature and rainfall; D2, after preprocessing the plurality of vehicle flow data, the plurality of maximum power demand data, the plurality of minimum power demand data and the plurality of date type data, the plurality of vehicle flow data are divided according to a preset ratio to obtain a training set and a validation set of an order prediction model; D3, a recurrent neural network model is constructed based on the deep learning model. The training set and validation set of the order prediction model are used to train and validate the time series network model, and the recurrent neural network model is iteratively optimized according to the optimization algorithm. The input is the traffic volume, date type, weather data, and maximum and minimum power demand of the previous hour, and the output is the predicted maximum and minimum power demand of the next hour.

8. The power dynamic warning system for a mobile storage and charging robot according to claim 1 is characterized in that: The determining whether to enable charging warning according to the order prediction model and the predicted maximum output power of a plurality of standby robots includes: E1, obtaining the maximum predicted output power and the minimum predicted output power according to the predicted output power of several standby robots; E2, obtains the maximum predicted power demand and the minimum predicted power demand for the next hour according to the order prediction model; E3, determine whether the maximum predicted output power is greater than the maximum predicted required power; if yes, jump to E4; if no, start the charging warning, and charge the standby robot with the maximum predicted output power according to the charging warning; E4, determine whether the minimum predicted output power is greater than the minimum predicted required power; if yes, turn off the charging warning; if no, turn on the charging warning, and charge the standby robot with the minimum predicted output power according to the charging warning.

9. The power dynamic warning system for a mobile storage and charging robot according to claim 1, characterized in that: Also includes: The voltage warning module is used to obtain the voltage value of the working robot and start the voltage warning when the voltage value reaches the voltage threshold.

10. A method for dynamic early warning of electric quantity for a mobile storage and charging robot, applied to a dynamic early warning system for electric quantity for a mobile storage and charging robot as claimed in any one of claims 1 to 9, characterized in that: include: S1, using a deep learning algorithm to build a power consumption prediction model, obtain the power consumption of several standby robots according to the power consumption prediction model, and obtain the predicted output power of several standby robots according to the power consumption; S2, obtaining order information, and assigning tasks to the standby robots according to the order information and the predicted output power of the standby robots; wherein the order information includes the standby power location and the required power consumption; S3, using a deep learning algorithm to build an order prediction model, and judging whether it is necessary to turn on the charging warning according to the order prediction model and the predicted output power maximum values ​​of several standby robots; if yes, turning on the charging warning, and charging several standby robots according to the charging warning; if no, turning off the charging warning, and charging the standby robots during the valley price period; wherein the maximum value includes the maximum value and the minimum value; S4, obtaining the voltage value of the working robot, and starting a voltage warning when the voltage value reaches a voltage threshold.

Images