Control system and control method of vehicle-mounted power grid service robot

Abstract

The invention provides a control system and a control method of a vehicle-mounted power grid service robot, and belongs to the technical field of mobile energy storage robots. Comprising a remote cloud server and a plurality of service robots. Each service robot is provided with a controller; and each service robot performs data interaction with the remote cloud server and the controllers on other service robots through the controller. The remote cloud server is used for receiving a rescue request and related information of a to-be-rescued vehicle and issuing the received information to the controller of each service robot; the controller of each service robot executes calculation operation, performs length comparison with rescue distances of other service robots and determines whether rescue conditions are met or not; if the rescue condition is met, rescue is executed; in the rescue process, the temperature of the vehicle is predicted in real time based on the convolutional hybrid neural network, so that the safety is ensured. According to the invention, on the basis of ensuring the charging safety, the rapid charging rescue of the to-be-rescued vehicle can be realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of mobile energy storage robots, and in particular relates to a control system and a control method for a vehicle-mounted power grid service robot. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] With the enhancement of modern environmental awareness and the transformation of energy structure, new energy vehicles (especially pure electric vehicles) have been rapidly developed and popularized. The demand for charging new energy vehicles has surged. Traditional fixed charging piles can no longer meet the growing demand for charging, and charging often requires queuing. Therefore, many new energy vehicle owners will choose to give up charging if they think the vehicle itself has enough power and the waiting time for charging is long; and the mileage of new energy vehicles will be affected by various factors (such as weather, road conditions, etc.), so the power displayed on the odometer of new energy vehicles is sometimes "false power", which often causes new energy vehicles to stop on the road and cannot continue to drive due to lack of power. Therefore, it is very necessary to study a service robot that can provide power replenishment and rescue for new energy vehicles that are parked on the road due to lack of power.

[0004] As a kind of service robot, the vehicle-mounted power grid service robot can replace humans to complete the charging service work of new energy vehicles, improve charging efficiency and service quality; however, the inventors found that the existing vehicle-mounted power grid service robots have the following technical problems: (1) The on-board power grid service robot needs to be dispatched through the cloud platform. The entire dispatch calculation process is implemented by the cloud platform, which requires a lot of calculation. This results in low real-time dispatch efficiency of the on-board power grid service robot, and it is difficult for rescue vehicles to get power replenishment and rescue in time.

[0005] (2) When performing power supply rescue, the existing on-board power grid service robots lack the ability to judge the safety status of the rescued vehicle during the charging process, making it difficult to detect charging abnormalities in a timely manner, and the safety of power supply rescue is relatively low. Summary of the invention

[0006] In order to overcome the above-mentioned deficiencies of the prior art, the present invention provides a control system and a control method for a vehicle-mounted power grid service robot, which can realize rapid power replenishment and rescue for vehicles to be rescued while ensuring the safety of power replenishment.

[0007] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions: A first aspect of the present invention provides a control system for a vehicle-mounted power grid service robot.

[0008] A control system for a vehicle-mounted power grid service robot, comprising: Remote cloud servers and multiple service robots; Each of the service robots is provided with a controller with the same function; each of the service robots realizes data interaction with the remote cloud server and the controllers on other service robots through the controller, specifically: The remote cloud server is used to receive the rescue request, location information and vehicle information of the vehicle to be rescued; when receiving the rescue request, the remote cloud server sends the location information and vehicle information of the vehicle to be rescued to the controller of each service robot; The controller of each service robot performs calculation operations to calculate the distance between the current position of the corresponding service robot and the position of the vehicle to be rescued, that is, the rescue distance; then, it exchanges data with the controllers on other service robots to obtain the rescue distances of other service robots; Compare the rescue distance with other service robots and determine whether the rescue conditions are met in combination with vehicle information; when the rescue conditions are met, determine to perform the rescue work; During the power replenishment and rescue process, the vehicle temperature is predicted in real time based on the convolutional hybrid neural network. When the vehicle temperature is predicted to be abnormal, the power replenishment is stopped in time to ensure the safety of the power replenishment and rescue process.

[0009] Furthermore, the vehicle information includes the vehicle model and the battery capacity of the vehicle model.

[0010] Furthermore, a length comparison is made with the rescue distance of other service robots. Specifically: if the number of service robots is n, the controller sorts and numbers the corresponding service robots and other service robots from 1 to n from near to far according to the comparison of the length of the rescue distance; first, service robot No. 1 is used as the standby service robot, and the controller on the standby service robot determines whether the rescue conditions are met according to the current state of the standby service robot; when the rescue conditions are not met, the service robot with the next number is selected as the new standby service robot for re-judgment.

[0011] Furthermore, the current state of the service preparation robot includes an action state and a current state of stored energy.

[0012] Furthermore, there are two standards for the standby service robot to meet the rescue conditions, and both standards must be met at the same time; among them, the first is that the working state of the standby service robot is idle, and the second is that the current state of the energy storage capacity of the standby service robot is not less than a first preset power threshold; the first preset power threshold is higher than the battery capacity of the vehicle to be rescued.

[0013] Furthermore, each service robot is equipped with a vehicle for carrying the service robot.

[0014] Furthermore, the convolutional hybrid neural network consists of two parts: a convolutional neural network and a bidirectional gated recurrent unit.

[0015] Furthermore, the convolutional neural network includes a self-attention layer, a convolution layer, a pooling layer and a deep feature layer; the bidirectional gated recurrent unit includes a front hidden layer and a rear hidden layer.

[0016] Furthermore, during the power replenishment rescue process, a sliding window is used to analyze the temperature residual of the collected vehicle charging data and set a first abnormal power replenishment threshold and a second abnormal power replenishment threshold; the analysis value of the temperature residual is compared with the first abnormal power replenishment threshold and the second abnormal power replenishment threshold to determine whether the vehicle temperature is abnormal.

[0017] A second aspect of the present invention provides a control method for a vehicle-mounted power grid service robot.

[0018] A control method for a vehicle-mounted power grid service robot, comprising: Receive the location information and vehicle information of the rescue vehicle sent by the remote cloud server; Calculate the rescue distance between the current location and the location of the vehicle to be rescued; Obtain the rescue distance of other service robots and compare the length with the rescue distance of other service robots; determine whether the rescue conditions are met in combination with the vehicle information; when the rescue conditions are met, determine to perform the rescue work; During the power replenishment and rescue process, the vehicle temperature is predicted in real time based on the convolutional hybrid neural network. When the vehicle temperature is predicted to be abnormal, the power replenishment is stopped in time to ensure the safety of the power replenishment and rescue process. One or more of the above technical solutions have the following beneficial effects: (1) In the present invention, the remote cloud server is only used to receive the rescue request, location information and vehicle information of the vehicle to be rescued. When receiving the rescue request, the remote cloud server will send the location information and vehicle information of the vehicle to be rescued to the controller on each service robot; and the relevant calculations for the specific scheduling problems of the service robot (distance judgment, selection of the service robot for rescue) are all implemented by the service robot. Therefore, based on the idea of ​​edge computing, the present invention relieves the computing pressure of the remote cloud server; effectively improves the real-time scheduling efficiency of the vehicle-mounted power grid service robot, and the vehicle to be rescued can be recharged and rescued in time.

[0019] (2) During the power supply and rescue process, the vehicle-mounted power grid service robot of the present invention can make real-time predictions on the vehicle temperature based on a convolutional hybrid neural network. When it is predicted that the vehicle temperature is abnormal, the power supply can be stopped in time to ensure the safety of the power supply and rescue process.

[0020] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0022] Figure 1 This is a schematic diagram of the structure of a control system of an on-board power grid service robot in Embodiment 1 of the present invention.

[0023] Figure 2 Schematic diagram of the structural composition of the convolutional hybrid neural network in Example 1 of the present invention.

[0024] Figure 3 This is a flow chart of a control method for a vehicle-mounted power grid service robot in Embodiment 2 of the present invention. DETAILED DESCRIPTION

[0025] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0026] It should be noted that the terms used herein are for describing specific embodiments only and are not intended to be limiting of exemplary embodiments according to the present invention.

[0027] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0028] Embodiment 1 This embodiment discloses a control system for a vehicle-mounted power grid service robot.

[0029] like Figure 1 As shown, a control system of a vehicle-mounted power grid service robot includes: Remote cloud servers and multiple service robots; Each of the service robots is provided with a controller with the same function; each of the service robots realizes data interaction with the remote cloud server and the controllers on other service robots through the controller, specifically: The remote cloud server is used to receive the rescue request, location information and vehicle information of the vehicle to be rescued; when receiving the rescue request, the remote cloud server sends the location information and vehicle information of the vehicle to be rescued to the controller of each service robot; The controller of each service robot performs calculation operations to calculate the distance between the current position of the corresponding service robot and the position of the vehicle to be rescued, that is, the rescue distance; then, it exchanges data with the controllers on other service robots to obtain the rescue distances of other service robots; Compare the rescue distance with other service robots and determine whether the rescue conditions are met in combination with vehicle information; when the rescue conditions are met, determine to perform the rescue work; During the power replenishment and rescue process, the vehicle temperature is predicted in real time based on the convolutional hybrid neural network. When the vehicle temperature is predicted to be abnormal, the power replenishment is stopped in time to ensure the safety of the power replenishment and rescue process.

[0030] Based on the above design, the present invention can realize rapid power supply rescue for the vehicle to be rescued on the basis of ensuring the safety of power supply. To facilitate the understanding of the technical solution of the present invention, the control system of the vehicle-mounted power grid service robot of the present invention is further explained and illustrated below, specifically: The controller is integrated into the service robot, and each service robot is equipped with a vehicle for carrying the service robot. The remote cloud server is used to receive the rescue request, location information and vehicle information of the vehicle to be rescued; when receiving the rescue request, the remote cloud server sends the location information and vehicle information of the vehicle to be rescued to the controller of each service robot.

[0031] Specifically, the owner of the vehicle to be rescued can initiate a rescue request through the relevant application installed on the smart terminal (mobile phone, iPad, etc.). When initiating the rescue request, the owner also needs to add the location information and vehicle information of the vehicle to be rescued on the application; the vehicle information includes the model and the battery capacity of the model. After the owner confirms the sending, the application will send the location information and vehicle information of the vehicle to be rescued together with the rescue request to the remote cloud server. When receiving the rescue request, the remote cloud server will send the location information and vehicle information of the vehicle to be rescued to the controller of each service robot; the controller of each service robot exchanges data with the remote cloud server through the Internet of Things.

[0032] When the controller receives the position information and vehicle information of the vehicle to be rescued, it starts to calculate the distance between the current position of the corresponding service robot and the position of the vehicle to be rescued, that is, the rescue distance.

[0033] Specifically, the controller needs to rely on map software to calculate the rescue distance; first, select one of the existing map software as the connection object; then, connect the controller to the API interface provided by the selected map software through the process of registering a developer account, obtaining an API key, configuring interface parameters, etc.; after the connection is established, the controller sends a request to the map software through the API interface, obtains the current position of the corresponding service robot and uses the position of the vehicle to be rescued as the target position, and the rescue distance can be calculated. In this embodiment, the selected map software is Baidu Map. Since establishing a connection with the map software belongs to the existing technology, this embodiment will not be described in detail here.

[0034] After the rescue distance of the corresponding service robot is calculated, the controller on each service robot exchanges data with the controllers on other service robots; therefore, any service robot can obtain the rescue distance of other service robots through the controller.

[0035] After each service robot obtains the rescue distance of other service robots, it compares the length with the rescue distance of other service robots and determines whether the rescue conditions are met in combination with the vehicle information; when the rescue conditions are met, it determines to perform the rescue work.

[0036] Specifically, after each service robot obtains the rescue distance of other service robots, if the total number of service robots used for power replenishment rescue is n, the controller on the service robot will sort and number itself and other service robots from 1 to n from near to far according to the comparison of the length of the rescue distance; first, service robot No. 1 (the service robot with the closest rescue distance) is used as the reserve service robot, and the controller on the reserve service robot determines whether the rescue conditions are met according to the current state of the reserve service robot; wherein the current state of the reserve service robot includes the action state and the existing state of the energy storage power. When the rescue conditions are not met, the next numbered service robot is selected as the new reserve service robot for re-judgment until a service robot that meets the rescue conditions is found. It should be noted that when the reserve service robot is selected, whether the reserve service robot meets the rescue conditions is only judged by the reserve service robot itself, and other service robots do not participate in the judgment of whether the conditions are met; and when the reserve service robot does not meet the rescue conditions, the next numbered service robot is judged as the new reserve service robot.

[0037] Furthermore, there are two criteria for the preparatory service robot to meet the rescue conditions, and both criteria must be met at the same time for the preparatory service robot to be considered to meet the rescue conditions; specifically, the first criterion is that the working state of the preparatory service robot is idle, and the second criterion is that the existing state of the energy storage of the preparatory service robot is not less than the first preset threshold of electric energy; the first preset threshold of electric energy is higher than the battery capacity of the vehicle to be rescued. In this embodiment, the first preset threshold of electric energy is set to 1.5 times the battery capacity of the vehicle to be rescued, that is, only when the preparatory service robot is in an idle state and its remaining energy storage is not less than 1.5 times the battery capacity of the vehicle to be rescued can it be considered to have the conditions for power-on rescue; when the rescue conditions are met, it is determined to perform the rescue work.

[0038] A service robot that meets the rescue conditions is used as a service robot for power replenishment rescue. Because the service robot is installed on the vehicle, on the actual way to the rescue destination, the staff responsible for driving the vehicle can drive according to the navigation route obtained by the service robot through connecting to map software, or they can navigate by themselves.

[0039] When the service robot for power-on rescue arrives at the rescue destination, it officially starts power-on rescue for the rescue vehicle. The controller on the service robot is embedded with a convolutional hybrid neural network and connected to a warning light. Therefore, during the power-on rescue process, the controller can make a real-time prediction of the vehicle's temperature based on the convolutional hybrid neural network. When it is predicted that the vehicle's temperature is abnormal, the power-on rescue is stopped in time to ensure the safety of the power-on rescue process.

[0040] Furthermore, the convolutional hybrid neural network consists of two parts: a convolutional neural network and a bidirectional gated recurrent unit. Figure 2 As shown. In this embodiment, first, the controller extracts features from the historical charging data of the rescue vehicle based on a convolutional neural network, wherein the convolutional neural network includes a self-attention layer, a convolution layer, a pooling layer, and a deep feature layer; then, the extracted features are sent to a bidirectional gated recurrent unit BiGRU for model training, wherein the bidirectional gated recurrent unit includes a front hidden layer and a rear hidden layer; finally, the historical parameters output by the training enter a sliding window; by setting a double threshold, the double threshold is used to judge the charging status of the rescue vehicle through the real-time data of the sliding window.

[0041] Specifically, the self-attention mechanism layer Self-attention consists of two parts: Multi-Head Attention structure and ScaldDot-Product Attention structure; among them, the Scald Dot-Product Attention structure is used to calculate the degree of relationship between the current word and other words, and derive the weight vector of other words based on this; then, the weight vector is used as the feature representation. The Multi-Head Attention structure is equivalent to the repeated operation of the Scald Dot-Product Attention structure, so it can be used to obtain more feature representations. In this embodiment, first, by adding a self-attention mechanism layer on the basis of the basic convolutional neural network, the input text data (charging data of the rescue vehicle) is vectorized and a vector matrix is ​​established; then the more important words are extracted through the Self-attention layer, and weights are assigned to them by calculating the similarity between the query and each key value to increase their importance. At the same time, the semantic features of each word in each sentence can be captured, and it is easier to obtain the medium and long-distance interdependence between words. Subsequently, a convolutional layer is added after the self-attention mechanism layer to perform convolution operations on the output multi-dimensional feature matrix to obtain local features of the text; the weights of the convolutional filters of the convolutional layer are adjusted by back propagation. The pooling layer is located after the convolutional layer, and performs dimensionality reduction and feature selection on the feature map output by the convolutional layer; the deep feature layer is a deep feature extraction structure formed by multiple combinations of convolutional layers and pooling layers. Through multiple convolution and pooling operations, the deep feature layer can extract deep features from the input data; in other words, the data connection relationship in the deep feature layer is gradually established through multiple convolution and pooling operations, that is, the output of each layer in the deep feature layer is used as the input of the next layer, and features are extracted and abstracted layer by layer to finally form a deep feature representation.

[0042] In the actual power replenishment and rescue process, the battery management system of the rescued vehicle can send the charging data of the vehicle battery before charging to the controller of the service robot through the CAN bus. The temperature residual of the collected vehicle charging data is analyzed using a sliding window and the first abnormal power replenishment threshold and the second abnormal power replenishment threshold are set; wherein the temperature residual includes the temperature residual mean and the temperature residual standard deviation coefficient. The two analysis values ​​of the temperature residual are compared with the first abnormal power replenishment threshold and the second abnormal power replenishment threshold to determine whether the vehicle temperature is abnormal. Specifically, when both analysis values ​​of the temperature residual are higher than the set first abnormal power replenishment threshold, the warning light on the service robot is on to remind the staff to improve the observation of the charging status; when both analysis values ​​of the temperature residual are higher than the set second abnormal power replenishment threshold, indicating that the vehicle temperature is abnormal, the service robot will immediately stop the power replenishment to ensure the safety of the power replenishment and rescue process.

[0043] Embodiment 2 This embodiment discloses a control method for a vehicle-mounted power grid service robot.

[0044] like Figure 3 As shown, a control method for a vehicle-mounted power grid service robot includes: Step S1, receiving the location information and vehicle information of the rescue vehicle sent by the remote cloud server; Step S2, calculating the rescue distance between the current location and the location of the vehicle to be rescued; Step S3, obtaining the rescue distance of other service robots and comparing the length with the rescue distance of other service robots; determining whether the rescue conditions are met in combination with the vehicle information; and determining to perform the rescue work when the rescue conditions are met; Step S4: During the power replenishment rescue process, the vehicle temperature is predicted in real time based on the convolutional hybrid neural network. When the vehicle temperature is predicted to be abnormal, the power replenishment is stopped in time to ensure the safety of the power replenishment rescue process.

[0045] Based on the above method, the present invention can realize rapid power supply rescue for the rescued vehicle on the basis of ensuring the safety of power supply. The specific implementation of steps S1-S4 is further explained and illustrated below, specifically: Step S1: Receive the location information and vehicle information of the rescue vehicle sent by the remote cloud server.

[0046] The owner of the vehicle to be rescued can initiate a rescue request through the relevant application installed on the smart terminal (mobile phone, iPad, etc.). When initiating the rescue request, the owner also needs to add the location information and vehicle information of the vehicle to be rescued on the application; the vehicle information includes the model and the battery capacity of the model. After the owner confirms the sending, the application will send the location information and vehicle information of the vehicle to be rescued together with the rescue request to the remote cloud server. When receiving the rescue request, the remote cloud server will send the location information and vehicle information of the vehicle to be rescued to the controller of each service robot; the controller on each service robot exchanges data with the remote cloud server through the Internet of Things, that is, the controller on each service robot receives the location information and vehicle information of the vehicle to be rescued sent by the remote cloud server through the Internet of Things.

[0047] Step S2: Calculate the rescue distance between the current location and the location of the vehicle to be rescued.

[0048] The controller needs to rely on map software to calculate the rescue distance; first, select one of the existing map software as the connection object; then, connect the controller to the API interface provided by the selected map software through the process of registering a developer account, obtaining an API key, and configuring interface parameters; after the connection is established, the controller sends a request to the map software through the API interface to obtain the current position of the corresponding service robot and use the position of the vehicle to be rescued as the target position, so as to calculate the rescue distance. In this embodiment, the selected map software is Baidu Map. Since establishing a connection with the map software belongs to the existing technology, this embodiment will not be described in detail here.

[0049] Step S3, obtaining the rescue distance of other service robots and comparing the length with the rescue distance of other service robots; determining whether the rescue conditions are met in combination with the vehicle information; and determining to perform the rescue work when the rescue conditions are met.

[0050] When each service robot obtains the rescue distance of other service robots, if the total number of service robots used for power replenishment rescue is n, the controller on the service robot will sort and number itself and other service robots from 1 to n from near to far according to the comparison of the length of the rescue distance. First, service robot No. 1 (the service robot with the closest rescue distance) is used as the reserve service robot, and the controller on the reserve service robot determines whether the rescue conditions are met according to the current state of the reserve service robot; wherein the current state of the reserve service robot includes the action state and the existing state of the energy storage power. When the rescue conditions are not met, the next numbered service robot is selected as the new reserve service robot for re-judgment until a service robot that meets the rescue conditions is found. It should be noted that when the reserve service robot is selected, whether the reserve service robot meets the rescue conditions is only judged by the reserve service robot itself, and other service robots do not participate in the judgment of whether the conditions are met; and when the reserve service robot does not meet the rescue conditions, the next numbered service robot is judged as the new reserve service robot.

[0051] There are two criteria for the preparatory service robot to meet the rescue conditions, and both criteria must be met at the same time for the preparatory service robot to be considered to meet the rescue conditions; specifically, the first criterion is that the working state of the preparatory service robot is idle, and the second criterion is that the current state of the energy storage of the preparatory service robot is not less than the first preset threshold of electric energy; wherein the first preset threshold of electric energy is higher than the battery capacity of the vehicle to be rescued. In this embodiment, the first preset threshold of electric energy is set to 1.5 times the battery capacity of the vehicle to be rescued, that is, only when the preparatory service robot is in an idle state and its remaining energy storage is not less than 1.5 times the battery capacity of the vehicle to be rescued, can it be considered to have the conditions for power-on rescue; when the rescue conditions are met, it is determined to perform the rescue work.

[0052] Step S4: During the power-on rescue process, the vehicle temperature is predicted in real time based on the convolutional hybrid neural network. When the vehicle temperature is predicted to be abnormal, the power-on rescue process is stopped in time to ensure the safety of the power-on rescue process. Specifically, this can be achieved through the following process: Step S4-1, obtaining charging data of the vehicle to be rescued and screening the main data.

[0053] The controller on the service robot obtains the charging data of the battery of the vehicle to be rescued before charging, which is stored in the battery management system of the vehicle to be rescued, through the CAN bus; the charging data that can be obtained includes the rated capacity, rated voltage, rated current, required voltage, required current, maximum allowable voltage, maximum allowable current and charging method (constant voltage charging, constant current charging), etc. of the battery.

[0054] Step S4-2: preprocess the data and divide it into historical and real-time data sets.

[0055] After acquiring the charging data, the controller on the service robot performs normalization on the charging data and divides the data sets into historical and real-time data sets; the data sets are mapped to [-1,1]. The normalization operation is calculated by the following formula: ; in, Represents the value of the charging data output after normalization, Indicates the actual value of the charging data of the rescue vehicle. and They respectively represent the maximum and minimum values ​​of the actual values ​​of the charging data of the rescue vehicle before normalization.

[0056] Step S4-3: Based on historical data, determine the convolutional hybrid neural network model parameters and train the corresponding model to determine the evaluation criteria.

[0057] Step S4-3-1, construct a convolutional neural network model, and use the convolutional neural network to extract features from the historical charging data of the rescue vehicle.

[0058] The convolutional neural network model consists of a self-attention layer, a convolution layer, a pooling layer, and a deep feature layer. The calculation formula of the convolutional neural network model is expressed as: ; in, represents the charging data sequence of the rescue vehicle obtained after convolution, represents the activation function, represents the convolution kernel, Indicates that the rescue vehicle is in Charging data at all times, Represents the bias coefficient of the convolution operation.

[0059] Step S4-3-2: construct a bidirectional gated recurrent unit BiGRU, train the model and output the trained historical parameters.

[0060] The bidirectional gated recurrent unit BiGRU consists of a front hidden layer and a back hidden layer, which can characterize the time series characteristics of the charging data of the rescue vehicle by combining the historical charging data (front hidden layer) and the future charging state (back hidden layer) of the rescue vehicle. The hidden layer of the bidirectional gated recurrent unit BiGRU is expressed as: ; in, Indicates single The computational process of the network, and Respectively represent single The state of the network's forward hidden layer and the state of the backward hidden layer, and Respectively represent single The previous state of the network’s forward hidden layer and the previous state of the backward hidden layer; and Respectively represent single The output weight of the previous state of the network's forward hidden layer and the output weight of the previous state of the backward hidden layer, The bias optimization parameter representing the previous state of the real-time hidden layer; represents the output weight of the current state of the real-time hidden layer, Represents the output weight of the previous state of the real-time hidden layer.

[0061] Step S4-3-3: Set evaluation criteria.

[0062] This embodiment uses the root mean square error as the evaluation standard. Specifically, the formula of the root mean square error is expressed as: ; in, represents the root mean square error value obtained, Indicates the number of charging temperature data obtained from the historical charging data of the vehicle to be rescued, Indicates the actual measured value of the charging temperature of the rescue vehicle. Indicates the predicted value of the corresponding charging temperature. If the root mean square error value is smaller, it means that the model training result is better.

[0063] Step S4-3-4: Send the historical parameters of the BiGRU training output into the sliding window, and continuously analyze the prediction residual.

[0064] This embodiment adopts the sliding window method, and specifies a length of L The sliding window of length L The principle of window sliding processing is: every time the sliding window outputs a sampling point, a new prediction value will be input into this window at the back end, and new windows with the same sequence length will be generated. This cycle repeats until all continuous sub-time series data are obtained. Finally, the window is covered with the true value, and the prediction residual is continuously analyzed. Specifically, when the length of the sliding window is L When , the calculation formula of the residual mean and residual standard deviation under the sliding window can be expressed as follows: ; in, and They represent the residual mean and temperature residual standard deviation of the charging temperature of the rescued vehicle, Indicates the first i The residual error of each sampling point.

[0065] Step S4-4: Analyze the residual of the normal charging data of the rescue vehicle based on the sliding window, set the first abnormal charging threshold and the second abnormal charging threshold, and judge the charging state according to the real-time data to determine whether the vehicle temperature is abnormal. This can be achieved by the following steps: Step S4-4-1, design reference warning threshold, namely: ; in, Indicates the warning upper limit of the residual mean value of the charging temperature of the rescue vehicle. Indicates the warning upper limit of the temperature residual standard deviation of the charging temperature of the rescue vehicle; and They represent the warning coefficients of the residual mean of the charging temperature of the rescued vehicle and the temperature residual standard deviation, represents the maximum absolute value of the residual mean of the charging temperature of the rescue vehicle, Indicates the maximum value of the residual standard deviation of the charging temperature of the rescue vehicle.

[0066] Step S4-4-2, setting a first abnormal power replenishment threshold and a second abnormal power replenishment threshold, the calculation formulas of the first abnormal power replenishment threshold and the second abnormal power replenishment threshold are as follows: ; in, Indicates the first abnormal power replenishment threshold, Indicates the second abnormal power replenishment threshold; and Specifically, in order to highlight that the mean temperature residual of the charging temperature of the rescued vehicle is more important than the standard deviation of the temperature residual, in this embodiment, the percentage coefficients are set and 70% and 80% respectively. It should be noted that the percentage coefficient and It can be set according to actual conditions, and this embodiment does not impose any specific limitation on this.

[0067] Step S4-4-3: Determine the charging status based on real-time data to determine whether the vehicle temperature is abnormal.

[0068] When both analysis values ​​of temperature residual are higher than the set first abnormal power replenishment threshold, that is, the mean and standard deviation of temperature residual of the rescue vehicle are higher than the set first abnormal power replenishment threshold When the charging status is low, the warning light on the service robot lights up to remind the staff to pay more attention to the charging status.

[0069] When both analysis values ​​of the temperature residual are higher than the set second abnormal power replenishment threshold, that is, the temperature residual mean and residual standard deviation of the vehicle to be rescued are higher than the set second abnormal power replenishment threshold When the temperature of the vehicle is abnormal, the service robot will immediately stop charging to ensure the safety of the charging and rescue process.

[0070] Based on the above process, the present invention can realize rapid power replenishment and rescue for the vehicle to be rescued on the basis of ensuring the safety of power replenishment.

[0071] Those skilled in the art should understand that the modules or steps of the present invention described above can be implemented by a general-purpose computer device, or alternatively, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. The present invention is not limited to any specific combination of hardware and software.

[0072] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A control system for a vehicle-mounted power grid service robot, characterized in that: include: Remote cloud servers and multiple service robots; Each of the service robots is provided with a controller with the same function; each of the service robots realizes data interaction with the remote cloud server and the controllers on other service robots through the controller, specifically: The remote cloud server is used to receive the rescue request, location information and vehicle information of the vehicle to be rescued; when receiving the rescue request, the remote cloud server sends the location information and vehicle information of the vehicle to be rescued to the controller of each service robot; The controller of each service robot performs calculation operations to calculate the distance between the current position of the corresponding service robot and the position of the vehicle to be rescued, that is, the rescue distance; then, it exchanges data with the controllers on other service robots to obtain the rescue distances of other service robots; Compare the rescue distance with other service robots and determine whether the rescue conditions are met in combination with vehicle information; when the rescue conditions are met, determine to perform the rescue work; During the power replenishment and rescue process, the vehicle temperature is predicted in real time based on the convolutional hybrid neural network. When the vehicle temperature is predicted to be abnormal, the power replenishment is stopped in time to ensure the safety of the power replenishment and rescue process.

2. The control system of a vehicle-mounted power grid service robot according to claim 1, characterized in that: The vehicle information includes the vehicle model and the battery capacity of the vehicle model.

3. The control system of a vehicle-mounted power grid service robot according to claim 1, characterized in that: Compare the length of the rescue distance with that of other service robots. Specifically: if the number of service robots is n, the controller sorts and numbers the corresponding service robots and other service robots from 1 to n from near to far according to the comparison of the length of the rescue distance; first, service robot No. 1 is used as the standby service robot, and the controller on the standby service robot determines whether the rescue condition is met according to the current state of the standby service robot; when the rescue condition is not met, select the service robot with the next number as the new standby service robot for re-judgment.

4. The control system of the vehicle-mounted power grid service robot according to claim 3, characterized in that: The current state of the service preparation robot includes the action state and the current state of the energy storage power.

5. A control system for a vehicle-mounted power grid service robot as described in any one of claims 3-4, characterized in that: There are two standards for the standby service robot to meet the rescue conditions, and both standards must be met at the same time; among them, the first is that the working state of the standby service robot is idle, and the second is that the current state of the energy storage power of the standby service robot is not lower than the first preset threshold of electric energy.

6. The control system of the vehicle-mounted power grid service robot according to claim 5, characterized in that: The first electric energy preset threshold is higher than the battery capacity of the vehicle to be rescued.

7. The control system of a vehicle-mounted power grid service robot according to claim 1, characterized in that: The convolutional hybrid neural network consists of two parts: a convolutional neural network and a bidirectional gated recurrent unit.

8. The control system of the vehicle-mounted power grid service robot according to claim 7, characterized in that: The convolutional neural network includes a self-attention layer, a convolution layer, a pooling layer and a deep feature layer; the bidirectional gated recurrent unit includes a front hidden layer and a rear hidden layer.

9. The control system of the vehicle-mounted power grid service robot according to claim 7, characterized in that: During the power replenishment rescue process, a sliding window is used to analyze the temperature residual of the collected vehicle charging data and set the first abnormal power replenishment threshold and the second abnormal power replenishment threshold; the analysis value of the temperature residual is compared with the first abnormal power replenishment threshold and the second abnormal power replenishment threshold to determine whether the vehicle temperature is abnormal.

10. A control method for a vehicle-mounted power grid service robot, characterized in that: include: Receive the location information and vehicle information of the rescue vehicle sent by the remote cloud server; Calculate the rescue distance between the current location and the location of the vehicle to be rescued; Obtain the rescue distance of other service robots and compare the length with the rescue distance of other service robots; determine whether the rescue conditions are met in combination with the vehicle information; when the rescue conditions are met, determine to perform the rescue work; During the power replenishment and rescue process, the vehicle temperature is predicted in real time based on the convolutional hybrid neural network. When the vehicle temperature is predicted to be abnormal, the power replenishment is stopped in time to ensure the safety of the power replenishment and rescue process.

Images