Meal delivery robot and user interaction optimization method based on artificial intelligence

Abstract

The invention discloses a meal delivery robot and user interaction optimization method based on artificial intelligence. The method comprises the steps of data acquisition, data preprocessing, user ordering prediction, meal delivery time prediction, prediction model hyper-parameter optimization and bidirectional interaction establishment. The invention relates to the technical field of man-machine interaction data processing, in particular to a meal delivery robot and user interaction optimization method based on artificial intelligence, which innovatively provides a service mode that a meal delivery robot actively recommends meal to a user and provides meal delivery time, and solves the problem that a traditional meal delivery robot is single in service mode. A neuron weight matrix is obtained again by adopting a priority weight probability method to establish an ordering prediction model, so that the accuracy and robustness of an ordering model prediction result are improved; the algorithm for obtaining the optimal parameters of the prediction model is improved by adopting a convergence factor, a dynamic position and local search, so that the optimal parameter combination is obtained, and the accuracy of a model prediction result is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of human-computer interaction data processing, and in particular to an artificial intelligence-based method for optimizing interaction between a food delivery robot and a user. Background Art

[0002] In today's society, people have higher and higher requirements for the efficiency and quality of catering services. As an innovative service method, food delivery robots are faced with the challenge of how to better interact with users. However, in the existing methods of interaction between food delivery robots and users, there is a situation where dining services are carried out only by relying on user command recognition. However, this passive service method can no longer meet people's needs. In the existing model suitable for order prediction, there is an uneven distribution of neuron weights, which leads to the problem of unstable state and poor generalization ability of the order prediction model. In the existing model suitable for prediction, there is an inappropriate setting of built-in parameters, which leads to inaccurate model prediction results. Summary of the invention

[0003] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides an optimization method for the interaction between a food delivery robot and a user based on artificial intelligence. In view of the fact that the existing method for interacting with a food delivery robot and a user only relies on the recognition of user commands to provide dining services, however, this passive service mode has been unable to meet people's needs. This solution innovatively analyzes the historical interaction data between the food delivery robot and the user. By analyzing the data, the user's dining intention is inferred, so that the food delivery robot can actively recommend meals to the user and provide delivery time, thereby solving the problem that the traditional food delivery robot service mode is single and passive, and greatly improving the user experience and the initiative and intelligence of the food delivery robot service; in view of In the existing model suitable for ordering prediction, there is an uneven distribution of neuron weights, which leads to the problem of unstable state and poor generalization ability of the ordering prediction model. This scheme adopts the method of priority weight probability to re-acquire the neuron weight matrix to establish the ordering prediction model, so that the ordering prediction model is in a stable state during the training and prediction process, thereby improving the accuracy and robustness of the prediction results of the ordering model; in the existing model suitable for prediction, there is an inappropriate setting of built-in parameters, which leads to inaccurate model prediction results. This scheme improves the algorithm for obtaining the optimal parameters of the prediction model by adopting convergence factor, dynamic position and local search, so as to obtain the optimal parameter combination and improve the accuracy of the model prediction results.

[0004] The technical solution adopted by the present invention is as follows: The method for optimizing the interaction between a food delivery robot and a user based on artificial intelligence provided by the present invention comprises the following steps:

[0005] Step S1: data collection;

[0006] Step S2: data preprocessing;

[0007] Step S3: User order prediction;

[0008] Step S4: predicting meal delivery time;

[0009] Step S5: Prediction model hyperparameter optimization;

[0010] Step S6: Establish two-way interaction.

[0011] Furthermore, in step S1, the data collection is specifically to obtain the original data of human-computer interaction optimization by collecting from the user order system of the food delivery robot;

[0012] The human-computer interaction optimization raw data includes user information data, food delivery robot data, catering data and human-computer real-time interaction data;

[0013] The user information data includes historical data of user ordering, user personal preference information and historical human-computer interaction data;

[0014] The food delivery robot data includes food delivery robot load-bearing data, food delivery robot driving data, and food delivery robot positioning and navigation information data;

[0015] The catering data includes the name of the meal, description of the meal, picture of the meal, serving time of the meal and price of the meal;

[0016] The human-computer real-time interaction data includes voice data, text data, behavior data and image data.

[0017] Further, in step S2, the data preprocessing is specifically performing data preprocessing on the original data of human-computer interaction optimization through the numerical filtering, data normalization and feature selection to obtain a preliminary data set for human-computer interaction optimization, including the following steps:

[0018] Step S21: numerical filtering, specifically processing missing values, abnormal values ​​and duplicate values ​​in the data;

[0019] Step S22: data normalization, which is used to normalize the data, specifically, normalize the data set by using the minimum-maximum method;

[0020] Step S23: feature selection, specifically using an adaptive feature selection method to filter out important features related to the user and the food delivery robot from the original data of human-computer interaction optimization.

[0021] Furthermore, in step S3, the user order prediction is specifically to predict the user's order by establishing an order prediction model, including the following steps:

[0022] Step S31: Establishing a meal order prediction model, specifically including the following steps:

[0023] Step S311: Initialize model parameters and set the model weight matrix, specifically including the following steps:

[0024] Step S3111: Initialize the weight matrix of the order prediction model. The formula used is as follows:

[0025] ;

[0026] In the formula, represents the weight matrix, i represents the index of the input neuron, x represents the index of the output neuron, and n represents the number of row and column elements of the weight matrix;

[0027] Step S3112: Calculate network strength using the following formula:

[0028] ;

[0029] In the formula, represents the network strength of the ith element at time t, represents the weight value of the element in the i-th row and the first column of the weight matrix, represents the network strength of the ith element at time t-1, represents the weight value of the element in the i-th row and t-1th column in the weight matrix, and t represents the number of iterations;

[0030] Step S3113: Calculate the priority weight probability, the formula used is as follows:

[0031] ;

[0032] In the formula, represents the priority weight probability, express The minimum value of express The sum of

[0033] Step S3114: updating the weight matrix. Specifically, the target node is obtained according to the priority weight probability through a random selection function, the column data corresponding to the target node is set as the sorted weight index column data, and the weight matrix is ​​updated;

[0034] Step S3115: Obtain the weight matrix. If the maximum number of iterations is reached, terminate the iteration and obtain the final weight matrix. Otherwise, continue to update the weight matrix.

[0035] Step S312: Forget gate calculation process, the formula used is as follows:

[0036] ;

[0037] In the formula, f t represents the output of the forget gate, X t Represents input data, H t-1 Indicates the hidden state at the previous moment, W xf and W hf represents the weight matrix corresponding to the forget gate, b f represents the bias parameter of the forget gate, represents the Sigmoid function, represents the Hadamard product;

[0038] Step S313: Input gate calculation process, the formula used is as follows:

[0039] ;

[0040] In the formula, i t represents the output of the input gate; W xi and W hi represents the corresponding input gate weight matrix; b i Represents the bias parameter of the input gate;

[0041] Step S314: Calculate new cell state candidate values , the formula used is as follows:

[0042] ;

[0043] Where W xc and W hc represents the corresponding cell unit state weight matrix; b c The bias parameter representing the cell state, It is the hyperbolic tangent function;

[0044] Step S315: Cell unit status update process, the formula used is as follows:

[0045] ;

[0046] In the formula, c t Indicates the current cell state, c t-1 It indicates the cell state at the last moment. Represents element-wise multiplication;

[0047] Step S316: Output gate calculation process, the formula used is as follows:

[0048] ;

[0049] In the formula, o t Represents the output of the output gate; Wxo and W ho It represents the corresponding output gate weight matrix; b o Represents the bias parameter of the output gate;

[0050] Step S317: hidden state calculation process, the formula used is as follows:

[0051] ;

[0052] In the formula, h t Indicates the hidden state at the current moment;

[0053] Step S318: Get the current prediction result , the formula used is as follows:

[0054] ;

[0055] Where b v Represents the bias parameter of the predicted value, and V represents the parameter matrix used to linearly transform the hidden state;

[0056] Step S32: training the order prediction model, specifically using the user information data in the human-computer interaction optimization preliminary data set to train the order prediction model to obtain a trained order prediction model;

[0057] Step S33: predicting user orders, specifically using the catering data in the preliminary data set for human-computer interaction optimization as input data for the trained order prediction model to obtain a predicted menu of the user's order.

[0058] Furthermore, in step S4, the delivery time prediction is specifically to predict the delivery time after ordering by establishing a delivery time prediction model, including the following steps:

[0059] Step S41: Establishing a meal delivery time prediction model, specifically including the following steps:

[0060] Step S411: Initialize the model structure, including the input layer, hidden layer and output layer;

[0061] Step S412: Update gate calculation, the formula used is as follows:

[0062] ;

[0063] In the formula, Represents the output of the update gate, which determines whether to update the current memory unit to a new memory unit; x t Indicates inputting data at the current moment, h t-1 Indicates the hidden state at the previous moment, W z represents the weight matrix of the update gate, bz represents the bias parameter of the update gate, Represents the Sigmoid function;

[0064] Step S413: reset gate calculation, the formula used is as follows:

[0065] ;

[0066] In the formula, Represents the output of the reset gate, which controls the influence of past memory at the current moment; W r represents the weight matrix of the reset gate, b r Represents the bias parameter of the reset gate;

[0067] Step S414: Memory unit calculation, the formula used is as follows:

[0068] ;

[0069] In the formula, Indicates the current state of the memory unit. represents the weight matrix of the current memory unit, b r Represents the bias parameter of the current memory unit, It is the hyperbolic tangent function;

[0070] Step S415: Hide status update, the formula used is as follows:

[0071] ;

[0072] In the formula, h t Indicates the hidden state at the current moment;

[0073] Step S416: Get the current prediction value ; The formula used is as follows:

[0074] ;

[0075] In the formula, represents the weight matrix of the output layer, represents the activation function, b y Represents the bias parameter of the output layer;

[0076] Step S42: training a meal delivery time prediction model, specifically, using the historical data of user ordering in the preliminary data set for human-computer interaction optimization and the catering data to train meal delivery time prediction, and obtaining a trained meal delivery time prediction model;

[0077] Step S43: predicting the meal delivery time, specifically, using the catering data in the preliminary data set of human-computer interaction optimization and the prediction results of the user's order as input data of the trained meal delivery time prediction model to obtain the meal delivery time prediction results.

[0078] Furthermore, in step S5, the prediction model hyperparameter optimization specifically includes the following steps:

[0079] Step S51: Initializing parameters, specifically by constructing initial algorithm parameters;

[0080] The initial parameters of the algorithm include the population size N, the maximum number of iterations t max , search for the lower limit lb and upper limit ub;

[0081] Step S52: Initialize the population, including the following steps:

[0082] Step S521: Generate N chaotic sequence values, and initialize the random values ​​to y 0 ; The formula used is as follows:

[0083] ;

[0084] In the formula, represents the i-th chaotic sequence value, represents the i+1th chaotic sequence value;

[0085] Step S522: Chaotic sequence value conversion search individual; the formula is as follows:

[0086] ;

[0087] In the formula, represents the position of the i-th individual;

[0088] Step S53: Calculate the individual fitness value f in the population i , the performance of the order prediction model and the delivery time prediction model established based on the individual location is taken as the individual fitness value; the individuals are sorted from best to worst according to the fitness value, and the individual fa with the best fitness and its location X are defined a , the individual fb with the second best fitness and its position X b , the third best individual fc and its position X c , classify fa, fb and fc as hunting individuals;

[0089] Step S54: Obtain the hunting individual position; the formula used is as follows:

[0090] ;

[0091] ;

[0092] ;

[0093] ;

[0094] In the formula, represents the position of the i-th individual after moving to fa, represents the position of the i-th individual after moving to fb, represents the position of the i-th individual after moving to fc, represents the position of the i-th individual in the t-th generation population, , and represents the randomness parameter in the range of [0,1], and t represents the number of iterations;

[0095] Step S55: Individual dynamic location update; the formula used is as follows:

[0096] ;

[0097] ;

[0098] ;

[0099] ;

[0100] In the formula, , and Respectively represent the updated new positions of fa, fb and fc, , and Respectively, in descending order , and The new value of represents the minimum evolution factor, and represents a random value in the range [0,0.3], represents the position of the i-th individual in the t+1 generation population;

[0101] Step S56: local search; the formula used is as follows:

[0102] ;

[0103] ;

[0104] Where ω(t) represents the mutation factor, and Both represent random numbers in the range [0,1]. is a smoothing term;

[0105] Step S57: search determination, specifically, by constructing a search termination condition, searching and determining the optimal individual position, and obtaining the optimal individual position data setting;

[0106] The search termination conditions include threshold termination and iteration termination;

[0107] The threshold termination is specifically to set the fitness threshold, when the individual fitness value f i When it is above the fitness threshold, the hyperparameter search is completed;

[0108] The iteration termination specifically refers to terminating the iteration and obtaining the optimal individual position when the maximum number of iterations is reached;

[0109] The optimal individual position includes order prediction model parameters and delivery time prediction model parameters.

[0110] Furthermore, in step S6, the establishment of a two-way interaction, specifically data interaction between the food delivery robot and the user, specifically includes the following steps:

[0111] Step S61: Active service strategy, specifically, the food delivery robot actively recommends dishes to the user and informs the food arrival time based on the predicted menu and delivery time prediction results of the order;

[0112] Step S62: User selection and confirmation, specifically the user adjusts the selection in the recommended dish list according to personal preferences. After the dish selection and order confirmation, the food delivery robot confirms the order details according to the user's selection, including the dish and quantity, and informs the user of the food arrival time predicted by the delivery time prediction model.

[0113] The beneficial effects achieved by the present invention using the above scheme are as follows:

[0114] (1) In the existing methods of interaction between food delivery robots and users, there is a situation where dining services are provided only by relying on user command recognition. However, this passive service method can no longer meet people's needs. This solution innovatively analyzes the historical interaction data between the food delivery robot and the user. By analyzing the data, the user's dining intention can be inferred, so that the food delivery robot can actively recommend food to the user and provide delivery time, thus solving the problem of the traditional food delivery robot's single and passive service method, and greatly improving the user experience and the initiative and intelligence of the food delivery robot service.

[0115] (2) In view of the problem that the existing order prediction model has uneven distribution of neuron weights, which leads to unstable state and poor generalization ability of the order prediction model, this scheme adopts the method of priority weight probability to re-acquire the neuron weight matrix to establish the order prediction model, so that the order prediction model is in a stable state during the training and prediction process, thereby improving the accuracy and robustness of the order prediction results of the model.

[0116] (3) In the existing models suitable for prediction, there are cases where the built-in parameters are improperly set, resulting in inaccurate model prediction results. This scheme improves the algorithm for obtaining the optimal parameters of the prediction model by adopting convergence factor, dynamic position and local search, so as to obtain the optimal parameter combination and improve the accuracy of the model prediction results. BRIEF DESCRIPTION OF THE DRAWINGS

[0117] Figure 1 A schematic diagram of the flow of the method for optimizing the interaction between a food delivery robot and a user based on artificial intelligence provided by the present invention;

[0118] Figure 2 is a schematic flow chart of step S2;

[0119] Figure 3 is a schematic diagram of the process of step S3;

[0120] Figure 4 is a schematic flow chart of step S4;

[0121] Figure 5 is a schematic flow chart of step S5;

[0122] Figure 6 is a schematic diagram of the process of step S6;

[0123] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0124] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0125] In the description of the present invention, it is necessary to understand that terms such as “upper”, “lower”, “front”, “back”, “left”, “right”, “top”, “bottom”, “inside” and “outside” indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the referred system or element must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as a limitation on the present invention.

[0126] Example 1, see Figure 1 The present invention provides an artificial intelligence-based food delivery robot and user interaction optimization method, the method comprising the following steps:

[0127] Step S1: data collection;

[0128] Step S2: data preprocessing;

[0129] Step S3: User order prediction;

[0130] Step S4: predicting meal delivery time;

[0131] Step S5: Prediction model hyperparameter optimization;

[0132] Step S6: Establish two-way interaction.

[0133] Example 2, see Figure 1 ,This embodiment is based on the above embodiment. In step S1, the data collection is specifically to obtain the original data of human-computer interaction optimization by collecting from the user order system of the food delivery robot;

[0134] The human-computer interaction optimization raw data includes user information data, food delivery robot data, catering data and human-computer real-time interaction data;

[0135] The user information data includes historical data of user ordering, user personal preference information and historical human-computer interaction data;

[0136] The food delivery robot data includes food delivery robot load-bearing data, food delivery robot driving data, and food delivery robot positioning and navigation information data;

[0137] The catering data includes the name of the meal, description of the meal, picture of the meal, serving time of the meal and price of the meal;

[0138] The human-computer real-time interaction data includes voice data, text data, behavior data and image data.

[0139] Example 3, see Figure 1 and Figure 2This embodiment is based on the above embodiment. In step S2, the data preprocessing is specifically to preprocess the original data of human-computer interaction optimization by numerical filtering, data normalization and feature selection to obtain a preliminary data set for human-computer interaction optimization, including the following steps:

[0140] Step S21: numerical filtering, specifically processing missing values, abnormal values ​​and duplicate values ​​in the data;

[0141] Step S22: data normalization, which is used to normalize the data, specifically, normalize the data set by using the minimum-maximum method;

[0142] Step S23: feature selection, specifically using an adaptive feature selection method to filter out important features related to the user and the food delivery robot from the original data of human-computer interaction optimization.

[0143] Example 4, see Figure 1 and Figure 3 This embodiment is based on the above embodiment. In step S3, the user order prediction is specifically to predict the user's order by establishing an order prediction model, including the following steps:

[0144] Step S31: Establishing a meal order prediction model, specifically including the following steps:

[0145] Step S311: Initialize model parameters and set the model weight matrix, specifically including the following steps:

[0146] Step S3111: Initialize the weight matrix of the order prediction model. The formula used is as follows:

[0147] ;

[0148] In the formula, represents the weight matrix, i represents the index of the input neuron, x represents the index of the output neuron, and n represents the number of row and column elements of the weight matrix;

[0149] Step S3112: Calculate network strength using the following formula:

[0150] ;

[0151] In the formula, represents the network strength of the ith element at time t, represents the weight value of the element in the i-th row and the first column of the weight matrix, represents the network strength of the ith element at time t-1, represents the weight value of the element in the i-th row and t-1th column in the weight matrix, and t represents the number of iterations;

[0152] Step S3113: Calculate the priority weight probability, the formula used is as follows:

[0153] ;

[0154] In the formula, represents the priority weight probability, express The minimum value of express The sum of

[0155] Step S3114: updating the weight matrix. Specifically, the target node is obtained according to the priority weight probability through a random selection function, and the column data corresponding to the target node is set as the sorted weight index column data to update the weight matrix.

[0156] Step S3115: Obtain the weight matrix. If the maximum number of iterations is reached, terminate the iteration and obtain the final weight matrix. Otherwise, continue to update the weight matrix.

[0157] Step S312: Forget gate calculation process, the formula used is as follows:

[0158] ;

[0159] In the formula, f t represents the output of the forget gate, X t Represents input data, H t-1 Indicates the hidden state at the previous moment, W xf and W hf represents the weight matrix corresponding to the forget gate, b f represents the bias parameter of the forget gate, represents the Sigmoid function, represents the Hadamard product;

[0160] Step S313: Input gate calculation process, the formula used is as follows:

[0161] ;

[0162] In the formula, i t represents the output of the input gate; W xi and W hi represents the corresponding input gate weight matrix; b i Represents the bias parameter of the input gate;

[0163] Step S314: Calculate new cell state candidate values , the formula used is as follows:

[0164] ;

[0165] Where W xc and W hc represents the corresponding cell unit state weight matrix; b c The bias parameter representing the cell state, It is the hyperbolic tangent function;

[0166] Step S315: Cell unit status update process, the formula used is as follows:

[0167] ;

[0168] In the formula, c t Indicates the current cell state, c t-1 It indicates the cell state at the last moment. represents element-wise multiplication;

[0169] Step S316: Output gate calculation process, the formula used is as follows:

[0170] ;

[0171] In the formula, o t Represents the output of the output gate; W xo and W ho It represents the corresponding output gate weight matrix; b o Represents the bias parameter of the output gate;

[0172] Step S317: hidden state calculation process, the formula used is as follows:

[0173] ;

[0174] In the formula, h t Indicates the hidden state at the current moment;

[0175] Step S318: Get the current prediction result , the formula used is as follows:

[0176] ;

[0177] Where b v Represents the bias parameter of the predicted value, and V represents the parameter matrix used to linearly transform the hidden state;

[0178] Step S32: training the order prediction model, specifically using the user information data in the human-computer interaction optimization preliminary data set to train the order prediction model to obtain a trained order prediction model;

[0179] Step S33: predicting user orders, specifically using the catering data in the preliminary data set for human-computer interaction optimization as input data for the trained order prediction model to obtain a predicted menu of the user's order.

[0180] By performing the above operations, in order to solve the problem that the existing order prediction model has uneven distribution of neuron weights, resulting in unstable state and poor generalization ability of the order prediction model, this scheme adopts the method of priority weight probability to re-acquire the neuron weight matrix to establish the order prediction model, so that the order prediction model is in a stable state during the training and prediction process, thereby improving the accuracy and robustness of the prediction results of the order prediction model.

[0181] Example 5, see Figure 1 and Figure 4 This embodiment is based on the above embodiment. In step S4, the meal delivery time prediction is specifically to predict the meal delivery time after ordering by establishing a meal delivery time prediction model, including the following steps:

[0182] Step S41: Establishing a meal delivery time prediction model, specifically including the following steps:

[0183] Step S411: Initialize the model structure, including the input layer, hidden layer and output layer;

[0184] Step S412: Update gate calculation, the formula used is as follows:

[0185] ;

[0186] In the formula, Represents the output of the update gate, which determines whether to update the current memory unit to a new memory unit; x t Indicates inputting data at the current moment, h t-1 Indicates the hidden state at the previous moment, W z represents the weight matrix of the update gate, b z represents the bias parameter of the update gate, Represents the Sigmoid function;

[0187] Step S413: reset gate calculation, the formula used is as follows:

[0188] ;

[0189] In the formula, Represents the output of the reset gate, which controls the influence of past memory at the current moment; W r represents the weight matrix of the reset gate, b r Represents the bias parameter of the reset gate;

[0190] Step S414: Memory unit calculation, the formula used is as follows:

[0191] ;

[0192] In the formula, Indicates the current state of the memory unit. represents the weight matrix of the current memory unit, b r Represents the bias parameter of the current memory unit, It is the hyperbolic tangent function;

[0193] Step S415: Hide status update, the formula used is as follows:

[0194] ;

[0195] In the formula, h t Indicates the hidden state at the current moment;

[0196] Step S416: Get the current prediction value ; The formula used is as follows:

[0197] ;

[0198] In the formula, represents the weight matrix of the output layer, represents the activation function, b y Represents the bias parameter of the output layer;

[0199] Step S42: training a meal delivery time prediction model, specifically, using the historical data of user ordering in the preliminary data set for human-computer interaction optimization and the catering data to train meal delivery time prediction, and obtaining a trained meal delivery time prediction model;

[0200] Step S43: predicting the meal delivery time, specifically, using the catering data in the preliminary data set of human-computer interaction optimization and the prediction results of the user's order as input data of the trained meal delivery time prediction model to obtain the meal delivery time prediction results.

[0201] Example 6, see Figure 1 and Figure 5 This embodiment is based on the above embodiment. In step S5, the prediction model hyperparameter optimization specifically includes the following steps:

[0202] Step S51: Initializing parameters, specifically by constructing initial algorithm parameters;

[0203] The initial parameters of the algorithm include the population size N, the maximum number of iterations t max , search for the lower limit lb and the upper limit ub;

[0204] Step S52: Initialize the population, including the following steps:

[0205] Step S521: Generate N chaotic sequence values, and initialize the random values ​​to y 0 ; The formula used is as follows:

[0206] ;

[0207] In the formula, represents the i-th chaotic sequence value, represents the i+1th chaotic sequence value;

[0208] Step S522: Chaotic sequence value conversion search individual; the formula is as follows:

[0209] ;

[0210] In the formula, represents the position of the i-th individual;

[0211] Step S53: Calculate the individual fitness value f in the population i , the performance of the order prediction model and the delivery time prediction model established based on the individual location is taken as the individual fitness value; the individuals are sorted from best to worst according to the fitness value, and the individual fa with the best fitness and its location X are defined a , the individual fb with the second best fitness and its position X b , the third best individual fc and its position X c , classify fa, fb and fc as hunting individuals;

[0212] Step S54: Obtain the hunting individual position; the formula used is as follows:

[0213] ;

[0214] ;

[0215] ;

[0216] ;

[0217] In the formula, represents the position of the i-th individual after moving to fa, represents the position of the i-th individual after moving to fb, represents the position of the i-th individual after moving to fc, represents the position of the i-th individual in the t-th generation population, , and represents the randomness parameter in the range of [0,1], and t represents the number of iterations;

[0218] Step S55: Individual dynamic location update; the formula used is as follows:

[0219] ;

[0220] ;

[0221] ;

[0222] ;

[0223] In the formula, , and Respectively represent the updated new positions of fa, fb and fc, , and Respectively, in descending order , and The new value of represents the minimum evolution factor, and represents a random value in the range [0,0.3], represents the position of the i-th individual in the t+1 generation population;

[0224] Step S56: local search; the formula used is as follows:

[0225] ;

[0226] ;

[0227] Where ω(t) represents the mutation factor, and Both represent random numbers in the range [0,1]. is a smoothing term;

[0228] Step S57: search determination, specifically, by constructing a search termination condition, searching and determining the optimal individual position, and obtaining the optimal individual position data setting;

[0229] The search termination conditions include threshold termination and iteration termination;

[0230] The threshold termination is specifically to set the fitness threshold, when the individual fitness value f i When it is above the fitness threshold, the hyperparameter search is completed;

[0231] The iteration termination specifically refers to terminating the iteration and obtaining the optimal individual position when the maximum number of iterations is reached;

[0232] The optimal individual position includes order prediction model parameters and delivery time prediction model parameters.

[0233] By performing the above operations, it is found that there are inappropriate built-in parameter settings in the existing models suitable for prediction, which leads to inaccurate model prediction results. This solution improves the algorithm for obtaining the optimal parameters of the prediction model by adopting convergence factor, dynamic position and local search, so as to obtain the optimal parameter combination and improve the accuracy of model prediction results.

[0234] Embodiment 7, see Figure 1 and Figure 6 This embodiment is based on the above embodiment. In step S6, the establishment of a two-way interaction, specifically data interaction between the food delivery robot and the user, specifically includes the following steps:

[0235] Step S61: Active service strategy, specifically, the food delivery robot actively recommends dishes to the user and informs the food arrival time based on the predicted menu and delivery time prediction results of the order;

[0236] Step S62: User selection and confirmation. Specifically, the user can adjust the selection in the recommended dish list according to personal preferences. After the dish selection and order confirmation, the food delivery robot confirms the order details according to the user's selection, including the dish and quantity, and informs the user of the food arrival time predicted by the delivery time prediction model.

[0237] By performing the above operations, in view of the existing method of interaction between food delivery robots and users, there is a situation in which dining services are provided by relying solely on user command recognition. However, this passive service method can no longer meet people's needs. This solution innovatively analyzes the historical interaction data between the food delivery robots and users. By analyzing the data, the user's dining intention is inferred, so that the food delivery robot can actively recommend meals to the user and provide delivery time, thereby solving the problem of the traditional food delivery robot's single and passive service method, and greatly improving the user experience and the initiative and intelligence of the food delivery robot service.

[0238] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0239] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

[0240] The present invention and its embodiments are described above, and such description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design a structure and embodiment similar to the technical solution without creativity, which should belong to the protection scope of the present invention.

Claims

1. An optimization method for interaction between a food delivery robot and a user based on artificial intelligence, characterized in that: The method comprises the following steps: Step S1: Data collection, by collecting user information data, food delivery robot data, catering data and human-machine real-time interaction data from the food delivery robot user order system, to obtain the original data for human-machine interaction optimization; Step S2: data preprocessing, by performing data preprocessing on the original data, specifically performing data preprocessing on the human-computer interaction optimization original data, to obtain a preliminary data set for human-computer interaction optimization; Step S3: predicting the user's order, specifically, re-obtaining the neuron weight matrix by adopting the priority weight probability method to establish an order prediction model to predict the user's order, and obtain a predicted menu of the user's order; Step S4: predicting the meal delivery time, specifically, predicting the meal delivery time after ordering by establishing a meal delivery time prediction model to obtain a meal delivery time prediction result; Step S5: Prediction model hyperparameter optimization, by using convergence factor, dynamic position and local search to improve the algorithm for obtaining the optimal model parameters, so as to obtain the optimal model parameter combination; Step S6: Establish two-way interaction, specifically data interaction between the food delivery robot and the user. The food delivery robot actively recommends food and informs the user of the food arrival time. The user makes a new selection based on the recommended food. When the food changes, the delivery time also changes.

2. The method for optimizing the interaction between a food delivery robot and a user based on artificial intelligence according to claim 1, characterized in that: In step S3, the user order prediction is specifically to predict the user's order by establishing an order prediction model, including the following steps: Step S31: Establishing a meal order prediction model, specifically including the following steps: Step S311: Initialize model parameters and set the model weight matrix, specifically including the following steps: Step S3111: Initialize the weight matrix of the order prediction model. The formula used is as follows: ； In the formula, represents the weight matrix, i represents the index of the input neuron, x represents the index of the output neuron, and n represents the number of row and column elements of the weight matrix; Step S3112: Calculate network strength using the following formula: ； In the formula, represents the network strength of the ith element at time t, represents the weight value of the element in the i-th row and the first column of the weight matrix, represents the network strength of the ith element at time t-1, represents the weight value of the element in the i-th row and t-1th column in the weight matrix, and t represents the number of iterations; Step S3113: Calculate the priority weight probability, the formula used is as follows: ； In the formula, represents the priority weight probability, express The minimum value of express The sum of Step S3114: updating the weight matrix. Specifically, the target node is obtained according to the priority weight probability through a random selection function, the column data corresponding to the target node is set as the sorted weight index column data, and the weight matrix is ​​updated; Step S3115: Obtain the weight matrix. If the maximum number of iterations is reached, terminate the iteration and obtain the final weight matrix. Otherwise, continue to update the weight matrix. Step S312: Forget gate calculation process, the formula used is as follows: ； In the formula, f t represents the output of the forget gate, X t Represents input data, H t-1 Indicates the hidden state at the previous moment, W xf and W hf represents the weight matrix corresponding to the forget gate, b f represents the bias parameter of the forget gate, represents the Sigmoid function, represents the Hadamard product; Step S313: Input gate calculation process, the formula used is as follows: ； In the formula, i t represents the output of the input gate; W xi and W hi represents the corresponding input gate weight matrix; b i Represents the bias parameter of the input gate; Step S314: Calculate new cell state candidate values , the formula used is as follows: ； Where W xc and W hc represents the corresponding cell unit state weight matrix; b c The bias parameter representing the cell state, It is the hyperbolic tangent function; Step S315: Cell unit status update process, the formula used is as follows: ； In the formula, c t Indicates the current cell state, c t-1 It indicates the cell state at the last moment. Represents element-wise multiplication; Step S316: Output gate calculation process, the formula used is as follows: ； In the formula, o t Represents the output of the output gate; W xo and W ho It represents the corresponding output gate weight matrix; b o Represents the bias parameter of the output gate; Step S317: hidden state calculation process, the formula used is as follows: ； In the formula, h t Indicates the hidden state at the current moment; Step S318: Get the current prediction result , the formula used is as follows: ； Where b v Represents the bias parameter of the predicted value, and V represents the parameter matrix used to linearly transform the hidden state; Step S32: training the order prediction model, specifically using the user information data in the human-computer interaction optimization preliminary data set to train the order prediction model to obtain a trained order prediction model; Step S33: predicting user orders, specifically using the catering data in the preliminary data set for human-computer interaction optimization as input data for the trained order prediction model to obtain a predicted menu of the user's order.

3. The method for optimizing the interaction between a food delivery robot and a user based on artificial intelligence according to claim 2, characterized in that: In step S4, the delivery time prediction is specifically to predict the delivery time after ordering by establishing a delivery time prediction model, including the following steps: Step S41: Establishing a meal delivery time prediction model, specifically including the following steps: Step S411: Initialize the model structure, including the input layer, hidden layer and output layer; Step S412: Update gate calculation, the formula used is as follows: ； In the formula, Represents the output of the update gate, which determines whether to update the current memory unit to a new memory unit; x t Indicates inputting data at the current moment, h t-1 Indicates the hidden state at the previous moment, W z represents the weight matrix of the update gate, b z represents the bias parameter of the update gate, Represents the Sigmoid function; Step S413: reset gate calculation, the formula used is as follows: ； In the formula, Represents the output of the reset gate, which controls the influence of past memory at the current moment; W r represents the weight matrix of the reset gate, b r Represents the bias parameter of the reset gate; Step S414: Memory unit calculation, the formula used is as follows: ； In the formula, Indicates the current state of the memory unit. represents the weight matrix of the current memory unit, b r Represents the bias parameter of the current memory unit, It is the hyperbolic tangent function; Step S415: Hide status update, the formula used is as follows: ； In the formula, h t Indicates the hidden state at the current moment; Step S416: Get the current prediction value ; The formula used is as follows: ； In the formula, represents the weight matrix of the output layer, represents the activation function, b y Represents the bias parameter of the output layer; Step S42: training a meal delivery time prediction model, specifically, using the historical data of user ordering in the preliminary data set for human-computer interaction optimization and the catering data to train meal delivery time prediction, and obtaining a trained meal delivery time prediction model; Step S43: predicting the meal delivery time, specifically, using the catering data in the preliminary data set of human-computer interaction optimization and the prediction results of the user's order as input data of the trained meal delivery time prediction model to obtain the meal delivery time prediction results.

4. The method for optimizing the interaction between a food delivery robot and a user based on artificial intelligence according to claim 1, characterized in that: In step S5, the prediction model hyperparameter optimization specifically includes the following steps: Step S51: Initializing parameters, specifically by constructing initial algorithm parameters; The initial parameters of the algorithm include the population size N, the maximum number of iterations t max , search for the lower limit lb and upper limit ub; Step S52: Initialize the population, including the following steps: Step S521: Generate N chaotic sequence values, and initialize the random value to y0; the formula used is as follows: ； In the formula, represents the i-th chaotic sequence value, represents the i+1th chaotic sequence value; Step S522: Chaotic sequence value conversion search individual; the formula is as follows: ； In the formula, represents the position of the i-th individual; Step S53: Calculate the individual fitness value f in the population i , the performance of the order prediction model and the delivery time prediction model established based on the individual location is taken as the individual fitness value; the individuals are sorted from best to worst according to the fitness value, and the individual fa with the best fitness and its location X are defined a , the individual fb with the second best fitness and its position X b , the third best individual fc and its position X c , classify fa, fb and fc as hunting individuals; Step S54: Obtain the hunting individual position; the formula used is as follows: ； ； ； ； In the formula, represents the position of the i-th individual after moving to fa, represents the position of the i-th individual after moving to fb, represents the position of the i-th individual after moving to fc, represents the position of the i-th individual in the t-th generation population, , and represents the randomness parameter in the range of [0,1], and t represents the number of iterations; Step S55: Individual dynamic location update; the formula used is as follows: ； ； ； ； In the formula, , and Respectively represent the updated new positions of fa, fb and fc, , and Respectively, in descending order , and The new value of represents the minimum evolution factor, and represents a random value in the range [0,0.3], represents the position of the i-th individual in the t+1 generation population; Step S56: local search; the formula used is as follows: ； ； Where ω(t) represents the mutation factor, and Both represent random numbers in the range [0,1]. is a smoothing term; Step S57: search determination, specifically, by constructing a search termination condition, searching and determining the optimal individual position, and obtaining the optimal individual position data setting; The search termination conditions include threshold termination and iteration termination; The threshold termination is specifically to set the fitness threshold, when the individual fitness value f i When it is above the fitness threshold, the hyperparameter search is completed; The iteration termination specifically refers to terminating the iteration and obtaining the optimal individual position when the maximum number of iterations is reached; The optimal individual position includes order prediction model parameters and delivery time prediction model parameters.

5. The method for optimizing the interaction between a food delivery robot and a user based on artificial intelligence according to claim 1, characterized in that: In step S6, the establishment of a two-way interaction, specifically data interaction between the food delivery robot and the user, specifically includes the following steps: Step S61: Active service strategy, specifically, the food delivery robot actively recommends dishes to the user and informs the food arrival time based on the predicted menu and delivery time prediction results of the order; Step S62: User selection and confirmation, specifically the user adjusts the selection in the recommended dish list according to personal preferences. After the dish selection and order confirmation, the food delivery robot confirms the order details according to the user's selection, including the dish and quantity, and informs the user of the food arrival time predicted by the delivery time prediction model.

6. The method for optimizing the interaction between a food delivery robot and a user based on artificial intelligence according to claim 1, characterized in that: In step S1, the data collection is specifically to obtain the original data for human-computer interaction optimization by collecting from the user order system of the food delivery robot; the original data for human-computer interaction optimization includes user information data, food delivery robot data, catering data and real-time human-computer interaction data; the user information data includes historical data of user ordering, user personal preference information and historical human-computer interaction data; the food delivery robot data includes load-bearing data of the food delivery robot, driving data of the food delivery robot and positioning and navigation information data of the food delivery robot; the catering data includes meal name, meal description, meal picture, meal delivery time and meal price; the real-time human-computer interaction data includes voice data, text data, behavior data and image data.

7. The method for optimizing the interaction between a food delivery robot and a user based on artificial intelligence according to claim 1, characterized in that: In step S2, the data preprocessing is specifically to preprocess the original data of human-computer interaction optimization by numerical filtering, data normalization and feature selection to obtain a preliminary data set for human-computer interaction optimization, including the following steps: Step S21: numerical filtering, specifically processing missing values, abnormal values ​​and duplicate values ​​in the data; Step S22: data normalization, which is used to normalize the data, specifically, normalize the data set by using the minimum-maximum method; Step S23: feature selection, specifically using an adaptive feature selection method to filter out important features related to the user and the food delivery robot from the original data of human-computer interaction optimization.

Images