Robots, human-robot interaction method based on robot emotion hierarchy, and memory

Abstract

The application provides a robot, a human-computer interaction method based on a robot emotion level and a memory.The human-computer interaction method based on the robot emotion level takes the robot as a core, first summarizes emotion states that the robot should have, and then describes an emotion state conversion mechanism, so that the robot can better process input speech of a user in an actual interaction process, and the emotion interaction ability of the robot is stronger.

Description

Technical Field

[0001] This invention relates to the field of robot interaction, and more particularly to a robot, a human-computer interaction method based on robot emotional levels, and a memory. Background Technology

[0002] Robot emotional interaction is a crucial part of the robotics field. Robots determine a user's emotional state and provide appropriate responses by recognizing the user's emotional polarity. However, current robot emotional interaction systems often only consider the user's emotional state—that is, the robot responds based on the user's reply—lacking modeling of the robot's own emotional state. In other words, current robot emotional interaction systems cannot categorize the robot's emotional states, nor can they enable the robot to switch between different emotional states like a human. In short, current robot emotional interaction systems exhibit stiff emotional expressions, failing to enhance the user experience. Summary of the Invention

[0003] To address the shortcomings of existing technologies, such as stiff emotions and poor user experience during robot interaction, this invention proposes a robot, a human-computer interaction method based on robot emotional levels, and a memory.

[0004] This invention proposes a human-computer interaction method based on the robot's emotional level, which makes the robot more human-like during the interaction process.

[0005] This invention proposes a human-computer interaction method based on robot emotion levels, comprising the following steps:

[0006] St1, Set the emotion transition template {(X i Y i ); X i+1}, where i represents the sequence number of the emotion transition template, X i Y represents the robot's current emotional level. i This represents the sentiment level of the user information received by the robot; the current sentiment level is X. i The robot receives emotions at level Y. i The emotional level after user information jumps to X i+1 X i+1 This is denoted as the emotional level pointed to by the i-th emotional transition template;

[0007] Set up a chatbot script library, which stores scripts. Each script is associated with an emotional level and an applicable scenario. The applicable scenario is used to determine the logical coherence between the script used by the chatbot and the user information.

[0008] St2. During the interaction between the robot and the user, the robot's emotional state X is determined in real time before it receives user information. j The emotional level Y of the received user information j ; Get (X) j Y j The emotional level pointed to by the corresponding emotional transition template is used as the interactive emotional level; the robot extracts the interaction scenario based on the user information, and selects the corresponding interactive emotional level and interaction scenario from the robot's dialogue script library to reply to the user.

[0009] Preferably, St2 specifically includes the following steps:

[0010] St21, Set the robot's current emotional level to X. j j represents the number of human-computer interaction rounds, and the initial value of j is 0;

[0011] St22, the robot receives user information and analyzes the corresponding context as an interaction scenario; the robot analyzes the emotional level Y corresponding to the user information. j Robot Acquisition (X) j Y j The corresponding emotional transition template points to the emotional level as the interactive emotional level X. j+1 ;

[0012] St23, the robot selects a target message from its message library based on the level of emotional interaction and the context of the interaction, and then uses the target message to reply to the user.

[0013] St24, let j = j + 1, then return to step St21.

[0014] Preferably, before St1, there is a step St0: defining a set of emotional levels A and X that includes multiple emotional levels. i Y i and X i+1 All belong to A; the emotional level set A includes negative emotional level, neutral emotional level and positive emotional level.

[0015] Preferably, the emotional level set A contains multiple negative emotional levels with sequentially increasing emotional intensity and multiple positive emotional levels with sequentially increasing emotional intensity.

[0016] Preferably, A is defined as {Hp, ..., H2, H1, N, S1, S2, ..., Sq}, where H1, H2, ..., Hp are positive emotion levels with increasing emotional intensity, S1, S2, ..., Sq are negative emotion levels with increasing emotional intensity, and N is a neutral emotion level; let H = {H1, H2, ..., Hp}, and S = {S1, S2, ..., Sq};

[0017] The emotional jump template set in St1 satisfies the following constraint conditions:

[0018] When Y i ∈H, and X i ∈S∪N, then X i+1 =H1;

[0019] When Y i ∈H, and X i =Hr, then X i+1 =Hu; when r < p, then u = r + 1; when r = p, then u = r;

[0020] When Y i ∈S, and X i ∈H∪N, then X i+1 =S1;

[0021] When Y i ∈S, and X i =Sw, then X i+1 =Sg; when w < q, then g = w + 1; when w = q, then g = w;

[0022] When Y i =N, then X i+1 =N.

[0023] Preferably, the number of negative emotional levels and positive emotional levels included in the emotional level set A is equal.

[0024] Preferably, X0 in St21 is the emotional level after the most recent interaction with the user.

[0025] Preferably, X0 in St21 is the neutral emotional level N.

[0026] The present invention proposes a robot with richer emotional expressions.

[0027] A robot proposed by the present invention adopts the described human - robot interaction method based on the robot's emotional levels when interacting with the user.

[0028] The present invention proposes a memory that can be directly used for the transformation of existing robots to achieve the emotional evolution of robots.

[0029] A memory proposed by the present invention stores a computer program, and when the computer program is executed, it implements the described human - robot interaction method based on the robot's emotional levels.

[0030] The advantages of the present invention are:

[0031] This invention proposes a human-computer interaction method based on robot emotional levels. Taking the robot as the core, it first summarizes the emotional states that the robot should possess, and then describes the emotional state transition mechanism. This invention enables the robot to better process user input during actual interaction, thereby making the robot's emotional interaction capabilities stronger.

[0032] In this invention, the robot's emotional state is divided into positive, neutral, and negative. During human-computer interaction, for each scenario, the robot can select different words according to the emotional level, which can not only ensure that the robot's reply to user information is logical, but also realize the robot's emotional expression, thereby improving the user experience.

[0033] This invention enables robots to share emotions with humans, resulting in a more complete human-computer interaction.

[0034] In this invention, the robot's emotions are divided into multiple levels, making the robot's emotional levels richer and the human-computer interaction more flexible.

[0035] In this invention, by constraining the emotional levels and emotional transition templates, the human-computer interaction method has a basis for implementation, thus providing a foundation for the widespread application of this invention.

[0036] The robot provided by this invention adopts the above-mentioned human-computer interaction method based on the robot's emotional level. This robot can interact with users more flexibly, has wider applicability, and is conducive to improving the user experience.

[0037] The present invention provides a memory that serves as a carrier for the aforementioned human-computer interaction method based on robot emotion levels, facilitating the modification of existing robots. Attached Figure Description

[0038] Figure 1 Here is a flowchart of a human-computer interaction method based on robot emotion levels;

[0039] Figure 2 This is a flowchart of the robot interaction process;

[0040] Figure 3 This is a schematic diagram of some of the emotional shifts in Example 1;

[0041] Figure 4 This is a schematic diagram of another part of the emotional shift in Example 1. Detailed Implementation

[0042] Example 1

[0043] In this embodiment, the emotional levels are set as follows:

[0044] Positive emotional levels H1 and H2; emotional intensity H2 > H1; that is, the emotional intensity of H2 is more intense than that of H1.

[0045] Negative emotional levels S1 and S2; emotional intensity S2 > S1; that is, the emotional intensity of S2 is more intense than that of S1.

[0046] Neutral emotional level N.

[0047] In this embodiment, positive refers to positive emotions such as happiness and joy, negative refers to negative emotions such as anger, sadness, and boredom, and neutral refers to no emotional fluctuation.

[0048] In this embodiment, the emotion transition template {(X i Y i ); X i+1 The following constraints apply to the settings of}:

[0049] When Y i ∈H, and X i If ∈S∪N, then X i+1 =H1;

[0050] When Y i ∈H, and X i =Hr, then X i+1 =Hu; when r<2, then u=r+1; when r=2, then u=r;

[0051] When Y i ∈S, and X i If ∈H∪N, then X i+1 =S1;

[0052] When Y i ∈S, and X i =Sw, then X i+1 =Sg; when w<2, then g=w+1; when w=2, then g=w;

[0053] When Y i =N, then X i+1 =N.

[0054] like Figure 3 , Figure 4 As shown in the table below, the specific emotion transition templates in this embodiment are as follows:

[0055] Table 1: Statistics of Emotional Jump Templates

[0056]

[0057]

[0058] In this embodiment, each dialogue in the robot dialogue library is associated with an emotional level and an applicable scenario. The scenario is used to determine whether the dialogue used by the robot is logically consistent with the user information, so as to avoid the robot's response being irrelevant.

[0059] Table 2. Examples of Phrases Related to Positive Emotional Levels

[0060]

[0061] Table 3. Examples of Phrases Related to Negative Emotional Levels

[0062]

[0063] In this embodiment, it is assumed that a user consults the robot about business procedures, and the emotional level of the robot during the business explanation process is N.

[0064] Then the robot receives the user message "You explained it in great detail"; the robot parses the interaction scenario corresponding to the user message as "the user is praising the robot's work", the emotional level of the user message is positive emotional level, the robot's emotional level jumps to H1; the robot replies "I will continue to work hard".

[0065] The robot receives the user message again: "You are the most attentive customer service representative I have ever encountered." The robot analyzes the interaction scenario corresponding to the user message as "the user is praising the robot's work." The emotional level of the user message is positive, and the robot's emotional level jumps to H2. The robot replies: "Haha, I'm so happy to be praised by you. Thank you."

[0066] In this embodiment, for each scenario, the robot can select different words based on the emotional level, which can not only ensure that the robot's reply to user information is logical, but also realize the richness of the robot's emotions, thereby improving the user experience.

[0067] In practice, the script may include text, voice, and facial expressions.

[0068] The above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A human-computer interaction method based on robot emotional levels, characterized in that, Includes the following steps: St1, Set the emotion transition template {(X i Y i ); X i+1 }, where i represents the sequence number of the emotion transition template, X i Y represents the robot's current emotional level. i This represents the sentiment level of the user information received by the robot; the current sentiment level is X. i The robot receives emotions at level Y. i The emotional level after user information jumps to X i+1 X i+1 This is denoted as the emotional level pointed to by the i-th emotional transition template; Set up a chatbot script library, which stores scripts. Each script is associated with an emotional level and an applicable scenario. The applicable scenario is used to determine the logical coherence between the script used by the chatbot and the user information. St2. During the interaction between the robot and the user, the robot's emotional state X is determined in real time before it receives user information. j The emotional level Y of the received user information j ;j represents the human-computer interaction round; obtain (X j Y j The emotional level pointed to by the corresponding emotional transition template is used as the interactive emotional level; the robot extracts the interaction scenario based on the user information, and the robot selects the corresponding interactive emotional level and interaction scenario from the robot's dialogue script library to reply to the user. Before St1, there is a step St0: defining a set of sentiment levels A and X that contains multiple sentiment levels. i Y i and X i+1 All belong to A; the emotional level set A includes negative emotional levels, neutral emotional levels, and positive emotional levels; The set of emotional levels A contains multiple negative emotional levels with increasing emotional intensity and multiple positive emotional levels with increasing emotional intensity. Define A = {Hp, ..., H2, H1, N, S1, S2, ..., Sq}, where H1, H2, ..., Hp represent the positive emotional levels with increasing intensity, S1, S2, ..., Sq represent the negative emotional levels with increasing intensity, and N represents the neutral emotional level; let H = {H1, H2, ..., Hp} and S = {S1, S2, ..., Sq}. The emotion transition template set in St1 satisfies the following constraints: When Y i ∈H, and X i If ∈S∪N, then X i+1 =H1; When Y i ∈H, and X i = Hr, then X i+1 = Hu; when r < p, then u = r + 1; when r = p, then u = r; When Y i ∈S, and X i If ∈H∪N, then X i+1 =S1; When Y i ∈ S, and X i = Sw, then X i+1 = Sg; when w < q, then g = w + 1; when w = q, then g = w; When Y i =N, then X i+1 =N.

2. The human-computer interaction method based on robot emotion levels as described in claim 1, characterized in that, St2 specifically includes the following steps: St21, Set the robot's current emotional level to X. j j represents the number of human-computer interaction rounds, and the initial value of j is 0; St22, the robot receives user information and analyzes the corresponding context as an interaction scenario; the robot analyzes the emotional level Y corresponding to the user information. j Robot Acquisition (X) j Y j The corresponding emotional transition template points to the emotional level as the interactive emotional level X. j+1 ; St23, the robot selects a target message from its message library based on the level of emotional interaction and the context of the interaction, and then uses the target message to reply to the user. St24, let j = j + 1, then return to step St21.

3. The human-computer interaction method based on robot emotion levels as described in claim 2, characterized in that, The set of emotional levels A contains an equal number of negative emotional levels and an equal number of positive emotional levels.

4. The human-computer interaction method based on robot emotion levels as described in claim 2, characterized in that, In St21, X0 represents the emotional level following the most recent interaction with the user.

5. The human-computer interaction method based on robot emotion levels as described in claim 2, characterized in that, In St21, X0 represents the neutral emotional level N.

6. A robot, characterized in that, The human-computer interaction method based on robot emotion level as described in any one of claims 1 to 5 is used when interacting with users.

7. A memory, characterized in that, The system contains a computer program that, when executed, implements the human-computer interaction method based on robot emotion levels as described in any one of claims 1 to 5.

Images