A human-like device interaction method based on data analysis

Abstract

This invention relates to the field of interactive data processing, and more particularly to a human-like device interaction method based on data analysis. This invention constructs a time-based movement path for feature points using moving coordinates, determines the path-independent features of the movement path within each time domain segment, calculates relatively independent characterization values, and sets independent labels for the movement path within each time domain segment. Subsequently, a predetermined proportion of sub-time domain segments are extracted at the beginning of each time domain segment, and the fitting degree between the movement path corresponding to the sub-time domain segment and the sample movement path is analyzed. Based on the independent labels, the movement path is adaptively analyzed, and the movement path of the time domain segment is selectively analyzed or the reference width is adjusted as appropriate. This invention can reduce misjudgments of movement paths, save computing power, improve the speed of movement path recognition, and thus improve the smoothness and reliability of the interactive experience.

Description

Technical Field

[0001] This invention relates to the field of interactive data processing, and more particularly to a human-like device interaction method based on data analysis. Background Technology

[0002] With the development of computer technology and Internet of Things technology, people's reliance on computers and smart devices is deepening. Correspondingly, users' expectations for interactive experience are also rising. While voice interaction technology has emerged, gesture recognition interaction technology is also making great strides. Users interact with devices using gestures, which greatly improves the naturalness and humanization of the interaction.

[0003] For example, patent publication number CN102854983A discloses a human-computer interaction method based on gesture recognition. First, a gesture image video stream is captured by a camera and converted into image frames. Then, the shape and features of the gesture, as well as its location information, are extracted from the image using a specific algorithm, and classification criteria are established to recognize the gesture. Next, coordinates or action commands are mapped based on the gesture shape and location to obtain specific system commands. Finally, specific system actions are driven according to requirements to simulate system mouse events, thus enabling human-computer interaction. This gesture recognition-based human-computer interaction method can replace traditional mouse operation and can be widely applied in scenarios such as dining, shopping, entertainment activities, or large-screen presentations in conferences, enhancing human-computer interaction.

[0004] However, the following problems still exist in the existing technology.

[0005] In practical applications, due to the sensitivity of smart devices to motion capture, users' slight daily hand movements, such as the natural swing of the arm while walking or the normal displacement of the hand when operating other objects, may be misinterpreted by the system as valid gesture commands, thereby triggering unnecessary functions and causing accidental touches. Furthermore, if relatively complex analysis methods are used to analyze gesture commands, the recognition speed is low, which reduces the smoothness and reliability of the interactive experience. Summary of the Invention

[0006] To address this, the present invention provides a human-like device interaction method based on data analysis, which overcomes the problem that in the prior art, normal hand movements and other operations may be misjudged as valid gesture commands by the system, thereby triggering unnecessary function activation and causing accidental touches. Furthermore, if a relatively complex analysis method is used to analyze gesture commands, the recognition speed is low, which reduces the smoothness and reliability of the interactive experience.

[0007] To achieve the above objectives, the present invention provides a human-like device interaction method based on data analysis, comprising:

[0008] The control and interactive device continuously monitors the movement coordinates of feature points of targets within the area;

[0009] Based on the moving coordinates, a time-based movement path of feature points is constructed, and the path-independent features of the movement path within each time domain segment are determined, including:

[0010] Divide the movement path into several path segments;

[0011] Compare the path segment with the previous path segment to determine the direction and change the path segment.

[0012] The number of path segments whose direction changes is determined as the number of movement direction switching times;

[0013] Calculate the difference in the number of movement direction switching times between the current time segment and the adjacent time segment;

[0014] The average curvature of the moving path is calculated as the curvature feature;

[0015] Calculate the difference between the curvature feature corresponding to the movement path and the reference curvature feature;

[0016] Calculate the relatively independent characterization value for the movement path within the time domain segment based on the path independence characteristics, including:

[0017] The ratio of the difference in the number of movement direction switching times to the standard threshold of the difference in the number of movement direction switching times is the movement direction independent characterization factor;

[0018] The ratio of the difference quantity to the difference quantity threshold is calculated as an independent characterization factor for curvature features;

[0019] The weighted summation of the independent characterization factor of the direction of movement and the independent characterization factor of the curvature feature yields the relatively independent characterization value.

[0020] Based on the relatively independent characterization value, an independent label is set for the movement path within the time domain segment;

[0021] A predetermined proportion of sub-time domain segments are extracted from the beginning of a time domain segment. The fit between the movement path corresponding to the sub-time domain segment and the sample movement path is analyzed. Based on the independent labels, the movement path is analyzed to determine the interaction commands, including...

[0022] Determine whether to analyze the movement path of the remaining sub-time domain segments, and determine the corresponding interaction command based on the analysis results;

[0023] Alternatively, determine whether to adjust the reference width of the sample movement path, compare and analyze the adjusted sample path with the movement path, and determine the corresponding interaction command based on the analysis results;

[0024] The reference curvature feature is determined based on several historical movement paths of the target.

[0025] Furthermore, setting independent labels for the movement path within the time domain segment based on the relatively independent representation value includes:

[0026] If the relative independent characterization value is greater than or equal to the preset relative independent characterization standard threshold, then a relative independent label is set for the movement path within the time domain segment;

[0027] If the relative independence standard value is less than the preset relative independence characterization standard threshold, then a non-relative independence label is set for the mobile path within the time domain segment.

[0028] Furthermore, extracting a predetermined proportion of sub-time domain segments at the beginning of the time domain segment includes,

[0029] Construct a timeline and determine the initial time corresponding to the beginning of the time domain segment on the timeline;

[0030] The time point is shifted backward along the time axis by a predetermined proportion to obtain the shifted time point;

[0031] The time period from the initial time to the shifted time is defined as the sub-time domain segment;

[0032] The proportional length is determined based on the total length of the time domain segment on the time axis.

[0033] Furthermore, based on the independent tags, the movement path is analyzed to determine the interaction instructions, including:

[0034] If the movement path within the time domain segment is a non-relatively independent label, then it is determined whether to analyze the movement path of the remaining sub-time domain segments, and the corresponding interaction command is determined based on the analysis result.

[0035] If the movement path within the time domain segment is a relatively independent label, then it is determined whether to adjust the reference width of the sample movement path, compare and analyze the adjusted sample path with the movement path, and determine the corresponding interaction command based on the analysis results.

[0036] Furthermore, determining whether to analyze the movement paths of the remaining sub-time domain segments includes,

[0037] If the fit between the movement path corresponding to the sub-time domain segment and the sample movement path is greater than or equal to the predetermined fit threshold, it is determined that the movement paths of the remaining sub-time domain segments need to be analyzed.

[0038] If the fit between the movement path corresponding to the sub-time domain segment and the sample movement path is less than the predetermined fit threshold, it is determined that there is no need to analyze the movement paths of the remaining sub-time domain segments.

[0039] Further, it is determined whether to adjust the reference width of the sample movement path, where,

[0040] If the fit between the moving path corresponding to the sub-time domain segment and the sample moving path is greater than or equal to the predetermined fit threshold, it is determined that the reference width of the sample moving path needs to be adjusted.

[0041] If the fit between the moving path corresponding to the sub-time domain segment and the sample moving path is less than the predetermined fit threshold, it is determined that there is no need to adjust the reference width of the sample moving path.

[0042] Furthermore, the adjusted sample movement path is compared and analyzed with the original movement path, including:

[0043] The sample movement path is normalized to the movement path and placed in the same coordinate system;

[0044] Construct a set of points whose distance from the sample's movement path does not exceed the reference width;

[0045] In the same coordinate system, the movement path is moved, and the maximum proportion of the movement path falling into the set of points is determined. The maximum proportion is determined as the degree of overlap.

[0046] Furthermore, if the overlap is greater than or equal to a predetermined overlap threshold, the movement path is determined to be valid, and a corresponding interaction command is generated based on the association relationship.

[0047] The association between the movement path and the interaction command is pre-built.

[0048] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention constructs a time-based movement path for feature points by moving coordinates, determines the path-independent features of the movement path within each time domain segment, calculates relatively independent characterization values, and sets independent labels for the movement path within the time domain segment. Subsequently, a predetermined proportion of sub-time domain segments are extracted at the beginning of the time domain segment, and the fitting degree between the movement path corresponding to the sub-time domain segment and the sample movement path is analyzed. Based on the independent labels, the movement path is adaptively analyzed, and the movement path of the time domain segment is selectively analyzed or the reference width is adjusted. This invention can reduce misjudgments of movement paths, save computing power, improve the recognition speed of movement paths, and thus improve the smoothness and reliability of the interactive experience.

[0049] In particular, this invention calculates relatively independent representation parameters based on path independence features. In reality, when a target makes a control gesture, it is usually an independent action that may differ from daily life or normal behavior. Therefore, this invention first considers whether the movement path within a time domain is relatively independent, and determines the path independence features to represent the above factors. Among them, the difference in the number of movement direction switching can represent the specificity or relative independence of the movement trajectory within the current time domain relative to other movement trajectories. At the same time, the difference quantity represents the specificity or relative independence of the current movement trajectory from normal actions. Thus, the independent representation parameters are calculated from two dimensions, which can represent the independence of the movement path relative to other movement paths. This provides data support for setting independent labels and performing adaptive analysis of movement paths based on independent labels, thereby reducing misjudgments of movement paths, saving computing power, improving the speed of movement path recognition, and improving the smoothness and reliability of the interactive experience.

[0050] In particular, this invention adaptively analyzes movement paths based on independent labels. In practice, for gesture control, the resulting movement paths are relatively independent and have certain specificity. Therefore, movement paths with relatively independent labels indicate that the target is making unconventional movements and has a higher tendency to perform gesture control. Consequently, this invention quickly determines the reference width of sample movement paths based on a certain proportion of movement paths. Under the premise of relatively matching path segments, due to the high tendency for gesture control, the reference width is adjusted to adaptively relax the judgment criteria for movement paths. This can improve recognition efficiency while ensuring reliability, thereby reducing misjudgments of movement paths, saving computing power, increasing the recognition speed of movement paths, and ultimately improving the smoothness and reliability of the interactive experience.

[0051] In particular, for movement paths with non-relatively independent labels, a certain proportion of the movement paths are analyzed to quickly determine whether to abandon the analysis of subsequent movement paths. In practice, movement paths with non-relatively independent labels represent movement paths with poor independence and a weaker tendency for the target to perform gesture control. Therefore, only a certain proportion of the path segments are analyzed, and the analysis of some movement paths is adaptively discarded. This can improve recognition efficiency while ensuring reliability, thereby reducing misjudgments of movement paths, saving computing power, improving the recognition speed of movement paths, and ultimately improving the smoothness and reliability of the interactive experience. Attached Figure Description

[0052] Figure 1 This is a schematic diagram illustrating the steps of a data analysis-based anthropomorphic device interaction method according to an embodiment of the invention.

[0053] Figure 2 A logic decision diagram for setting labels for a movement path according to an embodiment of the invention;

[0054] Figure 3 This is a logic block diagram illustrating how the movement path is analyzed and interactive instructions are determined based on the independent tags, according to an embodiment of the invention.

[0055] Figure 4 This is a logic diagram for determining whether a movement path is valid, as shown in an embodiment of the invention. Detailed Implementation

[0056] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0057] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0058] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0059] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0060] Please see Figure 1 The diagram illustrates the steps of a data-analysis-based anthropomorphic device interaction method according to an embodiment of the present invention. The data-analysis-based anthropomorphic device interaction method according to an embodiment of the present invention includes:

[0061] Step S1: Control the interactive device to continuously monitor the movement coordinates of feature points of the target within the area;

[0062] Step S2: Construct time-based movement paths for feature points based on the movement coordinates, and determine the path-independent features of the movement paths within each time segment, including:

[0063] Divide the movement path into several path segments;

[0064] Compare the path segment with the previous path segment to determine the direction and change the path segment.

[0065] The number of path segments whose direction changes is determined as the number of movement direction switching times;

[0066] Calculate the difference in the number of movement direction switching times between the current time segment and the adjacent time segment;

[0067] The average curvature of the moving path is calculated as the curvature feature;

[0068] Calculate the difference between the curvature feature corresponding to the movement path and the reference curvature feature;

[0069] Step S3, calculate the relatively independent characterization value for the movement path within the time domain segment based on the path independence characteristics, including:

[0070] The ratio of the difference in the number of movement direction switching times to the standard threshold of the difference in the number of movement direction switching times is the movement direction independent characterization factor;

[0071] The ratio of the difference quantity to the difference quantity threshold is calculated as an independent characterization factor for curvature features;

[0072] The weighted summation of the independent characterization factor of the direction of movement and the independent characterization factor of the curvature feature yields the relatively independent characterization value.

[0073] Based on the relatively independent characterization value, an independent label is set for the movement path within the time domain segment;

[0074] Step S4: Extract a predetermined proportion of sub-time domain segments from the beginning of the time domain segment, analyze the fitting degree between the movement path corresponding to the sub-time domain segment and the sample movement path, analyze the movement path based on the independent labels, and determine the interaction instructions, including...

[0075] Determine whether to analyze the movement path of the remaining sub-time domain segments, and determine the corresponding interaction command based on the analysis results;

[0076] Alternatively, determine whether to adjust the reference width of the sample movement path, compare and analyze the adjusted sample path with the movement path, and determine the corresponding interaction command based on the analysis results;

[0077] The reference curvature feature is determined based on several historical movement paths of the target.

[0078] Specifically, there are no restrictions on the form of the interactive device. The interactive device can be a photographic device that can acquire target image information, or it can be other forms, as long as it can obtain the movement coordinates of the target feature points. This will not be elaborated further.

[0079] Specifically, in practice, the center of the hand is used as a feature point. This means that the target image can be obtained, the center of the hand can be determined from the target image, and then the coordinates of the movement of the center of the hand can be determined.

[0080] Specifically, the sample movement path can be a pre-stored movement path used as a reference. Those skilled in the art can pre-build the association between the sample movement path and the interaction command so that when the target makes a corresponding movement path, the interaction device can generate the corresponding interaction command.

[0081] In some possible implementations, the interactive device can be connected to other devices that need to be controlled, and then the interactive device can generate interactive commands to control the corresponding other devices.

[0082] For example, in a factory, an interactive device is connected to a processing device that needs to be controlled. A sample movement path is generated in advance. When the interactive device detects that the actual movement path generated by the target is valid, the interactive device generates an interactive command to control the processing device to start and execute the corresponding function.

[0083] Specifically, the reference curvature feature is determined based on several historical movement paths of the target, wherein several historical movement paths of the target are obtained in advance, the curvature feature is determined, and the mean value of the curvature feature is calculated as the reference curvature feature.

[0084] It is understandable that when dividing path segments, the corresponding time-domain segments of each path segment should be the same.

[0085] It is understandable that the mean curvature is the average of the curvatures corresponding to each point on the path segment.

[0086] The standard threshold for the difference in the number of movement direction switching times and the threshold for the difference amount are preset. Specifically, several historical movement paths of the target are obtained in advance, the difference in the number of movement direction switching times corresponding to each historical movement path is calculated, the mean of the difference in the number of movement direction switching times is calculated, and 1.15 times the mean is taken as the standard threshold for the difference in the number of movement direction switching times.

[0087] Solve for the difference amount corresponding to each historical movement path, calculate the mean of the difference amount, and take 1.2 times the mean as the difference amount threshold.

[0088] In practice, the weight of the independent characterization factor for the direction of movement is 0.65, and the weight of the independent characterization factor for curvature features is 0.35.

[0089] This invention calculates relatively independent representation parameters based on path independence features. In reality, when a target makes a control gesture, it is usually an independent action that may differ from daily life or normal behavior. Therefore, this invention first considers whether the movement path within a time domain is relatively independent, determining the path independence features to represent the aforementioned factors. Among these, the difference in the number of movement direction switching can represent the specificity or relative independence of the movement trajectory within the current time domain relative to other movement trajectories. Simultaneously, the difference quantity represents the specificity or relative independence of the current movement trajectory from normal actions. Thus, by calculating the independent representation parameters from two dimensions, the independence of the movement path relative to other movement paths can be represented. This provides data support for subsequent setting of independent labels and adaptive analysis of movement paths based on these labels, thereby reducing misjudgments of movement paths, saving computing power, improving the speed of movement path recognition, and ultimately improving the smoothness and reliability of the interactive experience.

[0090] Specifically, please refer to Figure 2 As shown, it is a logical decision diagram for identifying the current model data label category according to an embodiment of the invention. Setting independent labels based on the relatively independent representation value of the movement path within the time domain segment includes:

[0091] If the relative independent characterization value is greater than or equal to the preset relative independent characterization standard threshold, then a relative independent label is set for the movement path within the time domain segment;

[0092] If the relative independence standard value is less than the preset relative independence characterization standard threshold, then a non-relative independence label is set for the mobile path within the time domain segment.

[0093] In practice, the threshold for the relatively independent characterization standard is selected within the range [1.2, 1.5].

[0094] Specifically, extracting a predetermined proportion of sub-time domain segments at the beginning of a time domain segment includes,

[0095] Construct a timeline and determine the initial time corresponding to the beginning of the time domain segment on the timeline;

[0096] The time point is shifted backward along the time axis by a predetermined proportion to obtain the shifted time point;

[0097] The time period from the initial time to the shifted time is defined as the sub-time domain segment;

[0098] The proportional length is determined based on the total length of the time domain segment on the time axis.

[0099] In practice, the predetermined percentage is selected within the range of 20% to 40%.

[0100] The proportional length is the product of the total length and the predetermined proportion.

[0101] Specifically, please refer to Figure 3 The diagram shown is a logical block diagram of an embodiment of the invention for analyzing the movement path and determining the interaction instructions based on the independent tags. The analysis of the movement path and determination of the interaction instructions based on the independent tags includes...

[0102] If the movement path within the time domain segment is a non-relatively independent label, then it is determined whether to analyze the movement path of the remaining sub-time domain segments, and the corresponding interaction command is determined based on the analysis result.

[0103] If the movement path within the time domain segment is a relatively independent label, then it is determined whether to adjust the reference width of the sample movement path, compare and analyze the adjusted sample path with the movement path, and determine the corresponding interaction command based on the analysis results.

[0104] Specifically, there are no restrictions on the form of the labels. It can be understood that the labels are virtual labels, and their purpose is to distinguish movement paths. Those skilled in the art can choose the form of the data labels themselves.

[0105] Specifically, determining whether to analyze the movement paths of the remaining sub-time domain segments includes,

[0106] If the fit between the movement path corresponding to the sub-time domain segment and the sample movement path is greater than or equal to the predetermined fit threshold, it is determined that the movement paths of the remaining sub-time domain segments need to be analyzed.

[0107] If the fit between the movement path corresponding to the sub-time domain segment and the sample movement path is less than the predetermined fit threshold, it is determined that there is no need to analyze the movement paths of the remaining sub-time domain segments.

[0108] Specifically, determining whether to adjust the reference width of the sample movement path, where,

[0109] If the fit between the moving path corresponding to the sub-time domain segment and the sample moving path is greater than or equal to the predetermined fit threshold, it is determined that the reference width of the sample moving path needs to be adjusted.

[0110] If the fit between the moving path corresponding to the sub-time domain segment and the sample moving path is less than the predetermined fit threshold, it is determined that there is no need to adjust the reference width of the sample moving path.

[0111] Specifically, when comparing the movement path within a sub-time domain segment with the sample movement path, it is necessary to compare it with the corresponding part of the sample movement path. A predetermined proportion of the sample movement path can be extracted in the time domain dimension for comparison, which will not be elaborated further here.

[0112] The method for obtaining the goodness of fit is not limited in practice. For example, the cosine similarity between two movement paths can be calculated, and the cosine similarity can be defined as the goodness of fit to characterize the degree of similarity between the two movement paths. This will not be elaborated further.

[0113] In implementation, the fit threshold is predetermined, where,

[0114] Those skilled in the art can predetermine a sample movement path, mimic the sample movement path to perform corresponding actions, monitor several actual movement paths generated, and determine the average goodness of fit among the movement paths.

[0115] The product of the mean goodness of fit and the accuracy coefficient is determined as the goodness of fit threshold, and the accuracy coefficient is selected within the interval [0.75, 0.85].

[0116] Specifically, comparing and analyzing the adjusted sample movement path with the original movement path includes:

[0117] The sample movement path is normalized to the movement path and placed in the same coordinate system;

[0118] Construct a set of points whose distance from the sample's movement path does not exceed the reference width;

[0119] In the same coordinate system, the movement path is moved, and the maximum proportion of the movement path falling into the set of points is determined. The maximum proportion is determined as the degree of overlap.

[0120] It is understandable that when performing normalization, the sample movement path and the movement path can be scaled to the same ratio, which will not be elaborated further.

[0121] It is understandable that after constructing a point set based on the sample movement path, the area represented by the point set is a strip-shaped area along the sample movement path.

[0122] It is understandable that, by continuously changing the position of the movement path in the coordinate system, the proportion of the set of points into which the movement path falls can be calculated, and thus the maximum proportion can be obtained, which will not be elaborated further.

[0123] It is understandable that when calculating the proportion of points that the movement path falls into the set of points, the movement path can be sampled to obtain a set of points for each path path, and it can be determined whether each point of the movement path falls into the set of points constructed based on the sample movement path. The proportion is based on the ratio of the number of points in the set of points that fall into the movement path to the total number of points in the corresponding set and the midpoint.

[0124] This invention adaptively analyzes movement paths based on their independent labels. In practice, movement paths formed by gesture control are relatively independent and specific from other movement paths. Therefore, movement paths with relatively independent labels indicate that the target is making unconventional movements and has a higher tendency to perform gesture control. This invention quickly determines the reference width of sample movement paths based on a certain proportion of movement paths. Furthermore, under the premise of relatively matching path segments, due to the high tendency for gesture control, the reference width is adjusted to adaptively relax the judgment criteria for movement paths. This improves recognition efficiency while ensuring reliability, reduces misjudgments of movement paths, saves computing power, increases the speed of movement path recognition, and ultimately improves the smoothness and reliability of the interactive experience.

[0125] Specifically, please refer to Figure 4 As shown, it is a logic determination diagram for determining whether a movement path is valid according to an embodiment of the invention. If the overlap is greater than or equal to a predetermined overlap threshold, the movement path is determined to be valid, and a corresponding interactive instruction is generated based on the association relationship.

[0126] The association between the movement path and the interaction command is pre-built.

[0127] The overlap threshold is also predetermined.

[0128] Those skilled in the art can predetermine a sample movement path, mimic the sample movement path to perform corresponding actions, monitor several actual movement paths generated, and determine the average overlap between the actual movement path and the sample movement path.

[0129] The product of the mean similarity and the accuracy coefficient is determined as the overlap threshold, and the accuracy coefficient is selected within the interval [0.75, 0.85].

[0130] For movement paths with non-relatively independent labels, a certain proportion of these movement paths are analyzed to quickly determine whether to abandon the analysis of subsequent movement paths. In practice, movement paths with non-relatively independent labels represent movement paths with poor independence and a weaker tendency for the target to perform gesture control. Therefore, only a certain proportion of the path segments are analyzed, and the analysis of some movement paths is adaptively discarded. This can improve recognition efficiency while ensuring reliability, thereby reducing misjudgments of movement paths, saving computing power, increasing the speed of movement path recognition, and ultimately improving the smoothness and reliability of the interactive experience.

[0131] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A human-like device interaction method based on data analysis, characterized in that, include: The control and interactive device continuously monitors the movement coordinates of feature points of targets within the area; Based on the moving coordinates, a time-based movement path of feature points is constructed, and the path-independent features of the movement path within each time domain segment are determined, including: dividing the movement path into several path segments; comparing each path segment with the previous path segment to determine the direction-changing path segments; determining the number of direction-changing path segments as the movement direction switching count; calculating the difference in the movement direction switching count between the current time domain segment and the adjacent time domain segment; solving for the average curvature of the movement path as the curvature feature; and calculating the difference between the curvature feature corresponding to the movement path and the reference curvature feature. The calculation of the relatively independent characterization value for the movement path within the time domain segment based on the path-independent characteristics includes: solving the ratio of the difference in the number of movement direction switching times to the standard threshold of the difference in the number of movement direction switching times as the movement direction independent characterization factor; solving the ratio of the difference amount to the difference amount threshold as the curvature feature independent characterization factor; and weighted summing the movement direction independent characterization factor and the curvature feature independent characterization factor to obtain the relatively independent characterization value. Based on the relatively independent characterization value, an independent label is set for the movement path within the time domain segment; A predetermined proportion of sub-time domain segments are extracted from the beginning of a time domain segment. The fitting degree between the movement path corresponding to the sub-time domain segment and the sample movement path is analyzed. The movement path is analyzed based on the independent label to determine the interaction command, including: determining whether to analyze the movement path of the remaining sub-time domain segments and determining the corresponding interaction command based on the analysis result; or, determining whether to adjust the reference width of the sample movement path, comparing and analyzing the adjusted sample path with the movement path, and determining the corresponding interaction command based on the analysis result. The reference curvature feature is determined based on several historical movement paths of the target.

2. The anthropomorphic device interaction method based on data analysis according to claim 1, characterized in that, Setting independent labels for movement paths within the time domain segment based on the relatively independent characterization values ​​includes: If the relative independent characterization value is greater than or equal to the preset relative independent characterization standard threshold, then a relative independent label is set for the movement path within the time domain segment; If the relative independent characterization value is less than the preset relative independent characterization standard threshold, then a non-relatively independent label is set for the mobile path within the time domain segment.

3. The anthropomorphic device interaction method based on data analysis according to claim 2, characterized in that, Extracting a predetermined proportion of sub-time domain segments from the beginning of a time domain segment includes: Construct a timeline and determine the initial time corresponding to the beginning of the time domain segment on the timeline; The time point is shifted backward along the time axis by a predetermined proportion to obtain the shifted time point; The time period from the initial time to the shifted time is defined as the sub-time domain segment; The proportional length is determined based on the total length of the time domain segment on the time axis.

4. The anthropomorphic device interaction method based on data analysis according to claim 3, characterized in that, Based on the independent tags, the movement path is analyzed to determine the interaction commands, including: If the movement path within the time domain segment is a non-relatively independent label, then it is determined whether to analyze the movement path of the remaining sub-time domain segments, and the corresponding interaction command is determined based on the analysis result. If the movement path within the time domain segment is a relatively independent label, then it is determined whether to adjust the reference width of the sample movement path, compare and analyze the adjusted sample path with the movement path, and determine the corresponding interaction command based on the analysis results.

5. The anthropomorphic device interaction method based on data analysis according to claim 1, characterized in that, Determining whether to analyze the movement paths of the remaining sub-time domain segments includes: If the fit between the movement path corresponding to the sub-time domain segment and the sample movement path is greater than or equal to the predetermined fit threshold, it is determined that the movement paths of the remaining sub-time domain segments need to be analyzed. If the fit between the movement path corresponding to the sub-time domain segment and the sample movement path is less than the predetermined fit threshold, it is determined that there is no need to analyze the movement paths of the remaining sub-time domain segments.

6. The anthropomorphic device interaction method based on data analysis according to claim 1, characterized in that, Determine whether to adjust the reference width of the sample movement path, where, If the fit between the moving path corresponding to the sub-time domain segment and the sample moving path is greater than or equal to the predetermined fit threshold, it is determined that the reference width of the sample moving path needs to be adjusted. If the fit between the moving path corresponding to the sub-time domain segment and the sample moving path is less than the predetermined fit threshold, it is determined that there is no need to adjust the reference width of the sample moving path.

7. The anthropomorphic device interaction method based on data analysis according to claim 1, characterized in that, The comparative analysis of the adjusted sample movement path and the movement path includes: The sample movement path is normalized to the movement path and placed in the same coordinate system; Construct a set of points whose distance from the sample's movement path does not exceed the reference width; In the same coordinate system, the movement path is moved, and the maximum proportion of the movement path falling into the set of points is determined. The maximum proportion is determined as the degree of overlap.

8. The anthropomorphic device interaction method based on data analysis according to claim 7, characterized in that, If the overlap is greater than or equal to a predetermined overlap threshold, the movement path is determined to be valid, and a corresponding interaction command is generated based on the association relationship. The association between the movement path and the interaction command is pre-built.

Images