Autism diagnosis and treatment auxiliary system based on traditional Chinese medicine dialectical physical examination and contact robot

Abstract

The application provides an autism diagnosis and treatment auxiliary system based on traditional Chinese medicine syndrome differentiation and a contact robot, relates to the field of medical data processing, and comprises the following modules: a data acquisition module for acquiring traditional Chinese medicine syndromes and standardized assessment results of autistic children; a parameter adjustment module for determining parameters of the contact robot, including weights of different contact positions; a record acquisition module for acquiring historical contact records of the autistic children and the contact robot; a feature extraction module for determining contact features of the autistic children based on the historical contact records of the autistic children and the contact robot and the weights of different contact positions of the contact robot; a type determination module for determining emotional types of the autistic children based on the contact features of the autistic children; and a scheme determination module for determining treatment schemes of the autistic children based on the contact features and the emotional types of the autistic children, which has the advantage of improving the individual targeting of diagnosis and treatment of the autistic children.

Description

Technical Field

[0001] This invention relates to the field of medical data processing, and in particular to an autism diagnosis and treatment assistance system based on traditional Chinese medicine diagnostic examination and contact robots. Background Technology

[0002] Autism spectrum disorder (ASD) is a global public health challenge, with core symptoms including social impairment, repetitive and stereotyped behaviors, and sensory abnormalities. The diagnosis and treatment of ASD has long faced challenges due to insufficient personalized intervention and a lack of dynamic assessment.

[0003] Traditional Chinese medicine (TCM) uses the theory of "syndrome differentiation and treatment" combined with the four diagnostic methods (inspection, auscultation and olfaction, inquiry, and palpation) to identify syndromes such as liver qi stagnation and heart-spleen deficiency in children, providing holistic guidance for intervention. Modern medicine relies on standardized assessment tools (such as the ABC scale and CARS scale) to quantify core symptoms such as social and behavioral issues. Traditional diagnosis and treatment, relying on manual observation and scale assessment, struggles to fully capture the differentiated responses of children to tactile stimuli, and tactile sensitivity or insensitivity directly affects intervention effectiveness. For example, some children exhibit strong resistance to contact with specific areas, which may be related to their emotional state or TCM syndromes (such as local meridian blockage caused by liver qi stagnation), but current technology lacks quantitative analysis methods. Contact robots can be used to collect data on children with autism. Through historical contact records accumulated from multiple rounds of interaction, the dynamic changes in children's tactile preferences can be revealed. However, the data collected by existing contact robots is superficial, simply counting the number of contacts or pressure values, without exploring the deeper connections between tactile responses and emotional states and syndrome characteristics, thus failing to provide effective diagnostic and treatment assistance.

[0004] Therefore, there is a need to provide an autism diagnosis and treatment assistance system based on traditional Chinese medicine diagnosis and physical examination and contact robots to improve the individualized treatment of children with autism. Summary of the Invention

[0005] This invention provides an autism diagnosis and treatment assistance system based on traditional Chinese medicine (TCM) diagnostic examination and a contact robot, comprising: a data acquisition module for acquiring TCM syndromes and standardized assessment results of autistic children; a parameter adjustment module for determining the parameters of the contact robot based on the TCM syndromes and standardized assessment results of autistic children, wherein the parameters of the contact robot include at least the weights of different contact positions; a record acquisition module for acquiring historical contact records between autistic children and the contact robot; a feature extraction module for determining the contact characteristics of autistic children based on the historical contact records and the weights of different contact positions of the contact robot; a type determination module for determining the emotional type of autistic children based on their contact characteristics; and a treatment plan determination module for determining the treatment plan for autistic children based on their contact characteristics and emotional type.

[0006] Furthermore, the parameter adjustment module is further used to: acquire the TCM syndromes, standardized assessment results, and historical contact records of autistic children from multiple tests; determine multiple autism types based on the TCM syndromes, standardized assessment results, and historical contact records of autistic children from multiple tests; determine the autism type of autistic children based on the TCM diagnosis results, standardized assessment results, and multiple autism types of autistic children; and determine the weights of different contact positions of the contact robot based on the autism type of autistic children.

[0007] Furthermore, the parameter adjustment module is further used to: for each tested autistic child, determine the contact frequency of different contact positions corresponding to the tested autistic child based on the historical contact records of the tested autistic child; for any two tested autistic children, determine multiple autism types based on the TCM syndromes, standardized assessment results, and corresponding contact frequencies of different contact positions of multiple tested autistic children; for each autism type, determine the weight of different contact positions of the contact robot corresponding to the autism type based on the contact frequencies of different contact positions corresponding to the tested autistic children included in the autism type.

[0008] Furthermore, the parameter adjustment module is further used to: calculate the matching degree between autistic children and each type of autism based on the TCM syndrome and standardized assessment results of autistic children; and determine the weight of different contact positions of the contact robot based on the matching degree between autistic children and each type of autism and the weight of different contact positions of the contact robot corresponding to each type of autism.

[0009] Furthermore, the feature extraction module is further used to: filter key contact records from the historical contact records between autistic children and the contact robot based on the weights of different contact positions of the contact robot; determine multiple key feature factors based on the historical contact records of autistic children from multiple tests; and construct a contact feature vector of autistic children based on the key contact records and multiple key feature factors.

[0010] Furthermore, the feature extraction module is further configured to: determine multiple feature factors, wherein the multiple feature factors include at least time-domain feature factors, stress distribution factors, and contact trajectory feature factors; for each autism type and feature factor, calculate the dispersion of the feature factor corresponding to the autism type based on the historical contact records of the autistic children tested for the autism type; for each autism type, determine the key feature factor corresponding to the autism type based on the dispersion of the feature factor corresponding to the autism type; and determine multiple key feature factors based on the matching degree between the autistic children and each autism type and the key feature factor corresponding to each autism type.

[0011] Furthermore, the type determination module is further configured to: determine multiple candidate children with autism for each test based on the matching degree between the child with autism and each type of autism; for each candidate child with autism, construct a contact feature vector based on the candidate child's historical contact records, the weights of different contact positions of the contact robot, and multiple key feature factors; filter children with autism for similar tests based on the contact feature vectors of the children with autism and the contact feature vectors of each candidate child with autism; and determine the emotion type of the child with autism based on the emotion labels of the children with autism for similar tests.

[0012] Furthermore, the type determination module is further configured to: determine the weight of each key feature factor based on the matching degree between the autistic child and each autism type and the dispersion of each feature factor corresponding to each autism type; calculate the similarity between the autistic child and each candidate autistic child in the test based on the contact feature vector of the autistic child, the contact feature vector of each candidate autistic child in the test, and the weight of each key feature factor; and filter autistic children in similar tests based on the similarity between the autistic child and each candidate autistic child in the test.

[0013] Furthermore, the treatment plan determination module is further used to: construct a treatment plan prediction model; and, based on the TCM syndrome, standardized assessment results, contact characteristics, and emotional type of children with autism, determine a treatment plan for children with autism using the treatment plan prediction model, wherein the treatment plan includes at least a TCM massage plan.

[0014] Furthermore, the proposed treatment model includes a multimodal feature fusion module and a hierarchical decision-making mechanism. The multimodal feature fusion module is used to fuse the TCM syndromes, standardized assessment results, contact characteristics, and emotional types of children with autism to generate a fused feature vector. The hierarchical decision-making mechanism is used to determine the TCM treatment plan for children with autism through a decision tree.

[0015] Compared with existing technologies, the autism diagnosis and treatment assistance system based on traditional Chinese medicine diagnostic examination and contact robot provided by this invention has at least the following beneficial effects: 1. By integrating TCM diagnostic examination data with data from contact robots, the system can comprehensively and accurately grasp the condition of children with autism. TCM syndromes provide key information from the perspective of imbalances in the body's Qi, blood, Yin and Yang, and organ functions. Standardized assessment results reflect the severity of the condition through quantitative indicators such as social skills and language abilities. Contact characteristics recorded by the contact robot and the resulting emotion types supplement information from the behavioral interaction level. This multi-dimensional data corroborates and complements each other, allowing the system to gain a deep understanding of each child's uniqueness. The parameter adjustment module determines the robot's contact position weights based on this information, the feature extraction module constructs precise contact feature vectors, the type determination module accurately identifies the emotion type, and the final treatment plan module develops a treatment plan tailored to individual needs, especially TCM massage plans, such as spleen-tonifying massage for children with weak spleen and stomach, greatly improving diagnostic accuracy and laying the foundation for effective treatment.

[0016] 2. The data acquisition module automatically collects TCM syndromes and standardized assessment results, while the parameter adjustment module determines robot parameters based on big data and algorithms, eliminating the need for tedious manual analysis and judgment. The record acquisition module automatically acquires historical contact records, the feature extraction module quickly constructs contact feature vectors, the type determination module efficiently determines the emotion type, and the treatment plan determination module rapidly generates a treatment plan using a treatment plan prediction model. The entire process reduces human intervention, avoids errors caused by human factors, improves diagnostic and treatment efficiency, enables children with autism to receive timely and scientifically appropriate treatment, and also reduces the workload of medical staff.

[0017] 3. Traditional Chinese medicine (TCM) massage plans are formulated based on different TCM syndromes, such as massage techniques to tonify the spleen (earth element) and clear the liver (wood element). These techniques improve children's physical condition by regulating organ function and balancing Qi, blood, Yin, and Yang, thereby strengthening their constitution, enhancing immunity, and promoting growth and development. The multimodal feature fusion module in the plan prediction model integrates TCM syndromes with other data, providing a more scientific basis for TCM diagnosis and treatment. The hierarchical decision-making mechanism uses decision trees to determine the massage plan, combining TCM experience with modern technology. This allows TCM massage to play a greater role in the treatment of autism, providing more comprehensive and effective treatment options for children with autism and promoting the development of TCM in the field of special disease diagnosis and treatment. Attached Figure Description

[0018] This specification will be further described by way of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting; in these embodiments, the same reference numerals denote the same structures, wherein: Figure 1 This is a schematic diagram of a module of an autism diagnosis and treatment assistance system based on traditional Chinese medicine diagnostic examination and a contact robot, according to some embodiments of this specification; Figure 2This is a schematic diagram illustrating the process of preprocessing historical measured data from the upstream of the watershed where the forecast section is located, based on some embodiments of this specification. Figure 3 This is a schematic diagram of the structure of an electronic device according to some embodiments of this specification. Detailed Implementation

[0019] To more clearly illustrate the technical solutions of the embodiments in this specification, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are merely some examples or embodiments of this specification. For those skilled in the art, these drawings can be applied to other similar scenarios without creative effort. Unless obvious from the context or otherwise specified, the same reference numerals in the drawings represent the same structures or operations.

[0020] Figure 1 These are schematic diagrams of modules of an autism diagnosis and treatment assistance system based on traditional Chinese medicine diagnostic examination and a contact robot, as shown in some embodiments of this specification. Figure 1 As shown, the autism diagnosis and treatment assistance system based on traditional Chinese medicine syndrome differentiation and physical examination and contact robot may include a data acquisition module, a parameter adjustment module, a record acquisition module, a feature extraction module, a type determination module, and a treatment plan determination module.

[0021] The data acquisition module is used to obtain the TCM syndromes and standardized assessment results of children with autism.

[0022] Specifically, the data acquisition module can obtain TCM syndromes and standardized assessment results of children with autism from external data sources.

[0023] Traditional Chinese medicine (TCM) collects multidimensional physical examination data of children through a "four diagnostic methods" system, combines it with the theory of the Six Channels Differentiation, and comprehensively analyzes their organ functions, the imbalance of Qi, blood, Yin, and Yang, ultimately deriving the TCM syndrome of children with autism. Specifically, the four diagnostic methods encompass the following dimensions: Visual diagnosis focuses on a child's outward manifestations, capturing internal pathological information through observation of features such as expression, complexion, tongue coating, and eye gaze. Children with autism often exhibit social interaction deficits such as eye avoidance and apathy, reflecting the pathogenesis of "lack of nourishment of the spirit"; a pale or sallow complexion may indicate liver stagnation and spleen deficiency or insufficient qi and blood; a thick, greasy tongue coating with obvious teeth marks is often related to internal phlegm and dampness; a dull or wandering gaze may be associated with disharmony between the heart and kidneys or phlegm obstructing the mind's orifices. For example, children with heart and liver fire syndrome may have a flushed complexion and a bright red tongue tip; those with phlegm obstructing the mind's orifices may have a white, greasy tongue coating and dull eyes.

[0024] Auscultation and olfaction are used to aid in diagnosis by listening to speech, sounds, and breathing patterns. Language disorders in children with autism often manifest as repetitive, stereotyped phrases, unusual vocalizations, or uncontrolled volume. Combined with shortness of breath, wheezing, or sighing, this may indicate disordered Qi flow or airway obstruction. For example, children with deficiency of both heart and spleen may have a weak voice and insufficient breath; those with deficiency of kidney essence may exhibit delayed language development accompanied by shallow breathing.

[0025] The consultation focuses on parental feedback, comprehensively collecting data on medical history, symptoms, and lifestyle habits. In Traditional Chinese Medicine (TCM), the core symptoms of autism (social impairment, narrow interests, repetitive behaviors) are often attributed to "disuse of the spirit" or "phlegm and blood stasis." Further investigation using the four diagnostic methods (physiognomy, clinical diagnosis, and TCM) can corroborate findings by inquiring about dietary preferences (e.g., a preference for sweet and greasy foods can generate phlegm and dampness), sleep quality (e.g., night crying and restlessness suggest excessive heart and liver fire), and bowel and bladder regularity (e.g., constipation is related to intestinal heat accumulation). For example, long-term picky eating may worsen spleen and stomach weakness, leading to deficiency of both heart and spleen; while excessive fascination with specific items may reflect phlegm obstructing the heart or liver stagnation transforming into fire.

[0026] Palpation is used to analyze the state of Qi and blood circulation through pulse diagnosis. The pulse of children with autism often presents abnormalities such as wiry and thin, slippery and rapid, or deep and slow: a wiry pulse indicates liver Qi stagnation, which is common in children with unstable emotions and irritability; a thin pulse indicates insufficient Qi and blood, which is common in those with developmental delays and low physical strength; a slippery and rapid pulse indicates internal disturbance of phlegm and heat, which is related to rigid behavior, impulsivity and irritability; a deep and slow pulse reflects kidney Yang deficiency, accompanied by delays in language and motor function.

[0027] Based on the data from the four diagnostic methods, a comprehensive analysis was conducted using the Six Channels Differentiation method: If the child exhibits irritability, insomnia, vivid dreams, a red tongue with a yellow coating, and a wiry and rapid pulse, combined with behaviors such as avoiding eye contact and repeatedly clapping, it can be diagnosed as a syndrome of excessive heart and liver fire; if there is delayed speech, slow reaction, a white and greasy tongue coating, a slippery pulse, accompanied by social withdrawal and narrow interests, it belongs to the syndrome of phlegm obstructing the heart orifices; if there is developmental delay, clumsy movements, aversion to cold, cold limbs, and a deep and slow pulse, it is mostly classified as a syndrome of kidney essence deficiency; while a sallow complexion, poor appetite, loose stools, a weak pulse, accompanied by stereotyped behaviors and sleep disorders, it conforms to the syndrome of deficiency of both heart and spleen.

[0028] After obtaining the TCM syndromes of children with autism, TCM practitioners can upload them to the data acquisition module.

[0029] We use a combination of internationally recognized assessment tools to obtain standardized assessment results for children with autism from multiple dimensions.

[0030] The PEP-3 (Personal Educational Assessment Scale for Children with Autism, Third Edition) is a comprehensive assessment tool covering four major areas: cognitive development, language expression and comprehension, behavioral performance, and life adaptation. It assesses children's cognitive processing abilities, including problem-solving, memory, and attention levels, through structured tasks (such as puzzles and imitation exercises). The language module distinguishes between expressive language (such as naming and asking questions) and receptive language (such as instruction comprehension and vocabulary recognition). Behavioral assessment focuses on stereotyped behaviors, social interaction, and emotion regulation. The self-care and adaptive behavior dimension examines daily skills such as dressing, eating, and toileting. The advantages of PEP-3 lie in its high ecological validity, its assessment scenarios closely resembling real-life situations, and its results being convertible to developmental age equivalents, directly reflecting the gap between children and their peers.

[0031] The VB-MAPP (Language Behavior Milestone Assessment and Placement Program) focuses on assessing core language and social impairments. Based on Skinner's theory of language behavior, this tool breaks down language ability into key areas such as naming (naming objects), conversation (responding to others and maintaining communication), and listening skills (comprehension of instructions and differentiation of words). It quantifies a child's language development stages through graded tasks (such as from "pointing to objects" to "describing functions"). Furthermore, VB-MAPP assesses impaired behaviors (such as self-stimulation and aggression) and learning potential, providing a behavioral basis for intervention strategies. Its milestone-based design can precisely pinpoint a child's language weaknesses; for example, if a child lags behind in the "listener response" stage, instruction comprehension training needs to be strengthened.

[0032] By cross-validating the results of the two tools, an integrated assessment report is generated: if PEP-3 shows language comprehension is better than expression, and VB-MAPP confirms a lag in "listener skills," then the overall assessment is "receptive language to expressive language conversion disorder"; if the high-frequency stereotyped behaviors assessed in PEP-3 overlap with the self-stimulatory behaviors assessed in VB-MAPP, then priority should be given to intervening in the behavioral problems to improve learning efficiency. The final assessment results are presented in visual charts showing developmental age equivalents, language stage milestones, and the frequency of behavioral problems.

[0033] After obtaining standardized assessment results for children with autism, healthcare professionals can upload them to the data acquisition module.

[0034] The parameter adjustment module is used to determine the parameters of the contact robot based on the TCM syndromes and standardized assessment results of children with autism.

[0035] Among them, the parameters of contact robots include at least the weights of different contact positions.

[0036] In some embodiments, the parameter adjustment module is further configured to: Obtain TCM syndromes, standardized assessment results, and historical contact records of autistic children from multiple tests; Based on the TCM syndromes of autistic children from multiple tests, standardized assessment results, and historical contact records, multiple types of autism were identified. Based on the TCM diagnostic results and standardized assessment results of children with autism, as well as various autism types, the autism type of children with autism is determined, and based on the autism type of children with autism, the weights of different contact positions of the contact robot are determined.

[0037] Specifically, the contact robot, as a key execution end of the autism diagnosis and treatment assistance system, is designed to closely match the physiological characteristics of children and the meridian theory of traditional Chinese medicine. The robot is equipped with multiple sets of touch sensors distributed in core contact areas such as the forehead, face, shoulders, upper arm, forearm, palm, back of hand, thigh and calf (as shown in the figure), which can collect interactive data such as contact pressure, contact duration and reaction actions with autistic children in real time.

[0038] The data acquisition parameters for touch sensors must meet the following requirements: a sampling frequency of at least 500Hz to capture rapid touch events; tactile thresholds adjusted according to the sensitivity of children with ASD, especially low-threshold responses to emotional touches; and data format including basic parameters such as touch location, force, and duration. In the preprocessing stage, a bidirectional linear digital filter is used for noise reduction, removing outliers that deviate from the mean by more than three times the standard deviation, and the data is standardized using the z-score method to ensure that the data is distributed near zero. These preprocessing steps effectively improve the accuracy of subsequent feature extraction and the stability of model training.

[0039] The weighting of different contact locations refers to the priority assigned to various contact points (such as the palm, shoulder, and calf) of the contact robot when interacting with children with autism, based on traditional Chinese medicine syndromes and assessment results. It characterizes the contribution of tactile stimulation at each location to the development of the current treatment plan for children with autism; a higher weight indicates that the contact at that location is more crucial for achieving personalized treatment.

[0040] In some embodiments, the parameter adjustment module is further configured to: For each autistic child in the test, the contact frequency for different contact locations is determined based on the child's historical contact records. For any two autistic children, multiple autism types can be identified based on the TCM syndromes of autistic children from multiple tests, standardized assessment results, and the frequency of contact at different contact locations. For each type of autism, the weights of different contact positions of the contact robot corresponding to the autism type are determined based on the contact frequency of different contact positions corresponding to the autism children tested for that autism type.

[0041] Specifically, the historical contact records of the autistic children being tested can record every contact between the autistic children and the contact robot. For each contact location, the ratio of the total number of contacts between the autistic children and that location to the corresponding duration of the historical contact records can be calculated as the contact frequency for that location for the autistic children.

[0042] For any two autistic children who have undergone any two tests, the overall similarity between the two tests is calculated based on their TCM syndromes, standardized assessment results, and the frequency of contact at different contact locations. Then, using a clustering algorithm (e.g., K-means clustering), the autistic children from multiple tests are clustered according to their overall similarity, resulting in multiple clusters, each corresponding to a different autism type.

[0043] First, for TCM syndromes (such as heart and liver fire syndrome, phlegm obstructing the heart orifices syndrome, etc.), similarity is calculated through syndrome type matching or vector space model. Same syndromes are scored as high similarity, while different syndromes are assigned different weighted scores based on their characteristics (such as the degree of liver stagnation, the degree of phlegm dampness). Second, for the standardized assessment results of PEP-3 and VB-MAPP, developmental age equivalent difference or milestone stage distance is used for quantification; the closer the developmental levels in the assessment domains (cognition, language, etc.), the higher the score. Finally, for contact frequency data, the frequency distribution patterns of various contact locations (palms, shoulders, etc.) of autistic children in the two tests are compared using cosine similarity or Euclidean distance; the more similar the patterns, the higher the score. The three types of scores are weighted and summed according to preset weights (e.g., syndrome accounts for 40%, assessment results for 30%, and contact frequency for 30%) to obtain the comprehensive similarity value between the two tests for autistic children. This value reflects the overall similarity between the two in the multidimensional features of physiological, behavioral and tactile responses, providing a quantitative basis for subsequent clustering algorithms to classify autism types (such as tactile defensive type and language-delayed type), and ensuring that autistic children tested in the same cluster have highly similar characteristics.

[0044] For each autism type and each contact location, the mean contact frequency of the tested autistic children within the cluster corresponding to that autism type is calculated as the average contact frequency for that contact location corresponding to that autism type. The ratio of the average contact frequency of that contact location to that autism type to the sum of the average contact frequencies of all contact locations to that autism type is used as the weight of that contact location to that autism type. By integrating the weights of each contact location to that autism type, the weights of different contact locations on the contact robot corresponding to each autism type are obtained.

[0045] In some embodiments, the parameter adjustment module is further configured to: Based on the TCM syndromes and standardized assessment results of children with autism, the matching degree between children with autism and each type of autism was calculated; The weights of different contact positions of the contact robot are determined based on the matching degree between autistic children and each type of autism, as well as the weights of different contact positions of the contact robot corresponding to each type of autism.

[0046] Specifically, for each type of autism, the similarity between the TCM syndrome and standardized assessment results of autistic children and the TCM syndrome and standardized assessment results of autistic children corresponding to the cluster center of the cluster corresponding to the autism type can be calculated as the matching degree between autistic children and each type of autism. The higher the similarity, the higher the matching degree.

[0047] For each contact point, the weight of that contact point for the contact robot can be obtained by weighting and summing the weights of each autism type according to the matching degree.

[0048] For example, the weight of each contact position of a contact robot can be calculated using the following formula: in, Let be the weight of the i-th contact position of the contact robot. To determine the match between a child with autism and the j-th type of autism, Let the weight of the i-th contact location correspond to the j-th autism type. To determine the match between autistic children and the kth type of autism, The weight of the j-th autism type corresponding to the k-th contact location. The total number of autism types, This represents the total number of contact points.

[0049] Determining the weights of different contact points on a contact robot is crucial for adapting to the differentiated contact patterns of children with different types of autism when expressing emotions, thereby improving the accuracy of emotion recognition and the efficiency of intervention. For example, children with different symptoms or behavioral characteristics, such as "excessive heat" or "deficiency of kidney," exhibit significant preferences or avoidance in their responses to different contact points on the robot. For instance, children with "excessive heat" may express discomfort by avoiding facial contact when experiencing emotional fluctuations due to facial sensitivity, while children with "deficiency of kidney" may show anxiety or stereotyped behaviors in specific limb contact (such as repeated arm touch). If the robot treats all sensor data equally, it will not only introduce a large amount of irrelevant or interfering information (such as meaningless contact data in low-sensitivity areas), increasing the computational burden on the emotion recognition algorithm, but may also reduce the real-time performance and accuracy of recognition due to the submersion of key contact signals. By determining the weights of contact points and focusing on sensor points that are strongly correlated with the emotional expression of children with specific types, the data dimensionality can be reduced from all sensors to a key subset. This reduces redundant computation, improves system response speed, and allows for more sensitive capture of emotional changes based on core contact patterns, providing a reliable basis for subsequent personalized interventions.

[0050] The record acquisition module is used to acquire historical contact records between autistic children and contact robots.

[0051] Specifically, contact robots can record the characteristics of each contact between a child with autism and the contact robot (e.g., time point, pressure, etc.), forming a historical contact record between the child with autism and the contact robot.

[0052] The feature extraction module is used to determine the contact characteristics of children with autism based on their historical contact records with the contact robot and the weights of different contact positions of the contact robot.

[0053] In some embodiments, the feature extraction module is further configured to: Based on the weights of different contact positions of the contact robot, key contact records were selected from the historical contact records of autistic children and the contact robot. Based on the historical contact records of autistic children from multiple tests, several key characteristic factors were identified; Based on key contact records and multiple key feature factors of children with autism, a contact feature vector of children with autism is constructed.

[0054] Specifically, valid contact positions can be filtered based on the weight of each contact position of the contact robot. For example, contact positions with a weight greater than 0.1 can be considered valid contact positions. Invalid contact positions are deleted from the historical contact records between autistic children and the contact robot, while valid contact positions are retained as key contact records.

[0055] In some embodiments, the feature extraction module is further configured to: Multiple characteristic factors are identified, including at least time-domain characteristic factors (e.g., contact frequency, average force, maximum force, duration, etc.), pressure distribution factors (e.g., contact area, pressure center, pressure distribution symmetry, pressure gradient, etc.), and contact trajectory characteristic factors (e.g., trajectory length, movement speed, trajectory complexity, region preference, etc.). For each autism type and trait factor, the dispersion of the trait factor corresponding to the autism type is calculated based on the historical contact records of autistic children included in the tests for the autism type. For each autism type, the key feature factors corresponding to the autism type are determined based on the dispersion of the feature factors corresponding to the autism type; Based on the matching degree between autistic children and each autism type, and the key characteristic factors corresponding to each autism type, multiple key characteristic factors are identified.

[0056] Specifically, for each autism type and trait factor, based on the historical contact records of autistic children tested within the autism type, the value of the trait factor corresponding to the autism type tested within the autism type is determined, and its variance is calculated as the dispersion of the trait factor corresponding to the autism type.

[0057] For each autism type, based on the dispersion of each feature factor corresponding to the autism type from largest to smallest, a certain proportion (e.g., the top 30%) of the feature factors are selected as key feature factors.

[0058] Based on the matching degree between autistic children and each autism type, the autism type matched by the autistic child is determined. According to the key feature factors corresponding to the autism type matched by the autistic child, multiple key feature factors are determined. For example, the autism type with the highest matching degree is taken as the autism type matched by the autistic child, and the key feature factors corresponding to the autism type matched by the autistic child are taken as multiple key feature factors.

[0059] Based on key contact records and multiple key feature factors of children with autism, a contact feature vector of children with autism is constructed, wherein the contact feature vector of children with autism can include the value corresponding to each key feature factor.

[0060] Historical contact records from contact robots contain comprehensive sensor data. However, different types of children exhibit significant differences in sensitivity to contact locations, resulting in a large amount of redundant information unrelated to emotional expression or behavioral characteristics within the complete data set. By introducing contact location weights, the system can accurately filter out valid contact location records with weights exceeding a threshold, eliminating invalid data and directly reducing the dimensionality and computational load of subsequent processing. Secondly, in the feature extraction stage, the system uses dispersion analysis to select key feature factors with high discriminative power for specific autism types from multiple dimensions such as time domain, pressure distribution, and contact trajectory, avoiding resource waste and noise interference caused by full-factor calculation. This two-stage dimensionality reduction strategy—first filtering invalid contact location records through weights, and then optimizing feature factor selection through dispersion—not only makes the contact feature vectors more focused on core information strongly correlated with child type, improving the real-time performance of emotion recognition and behavioral analysis, but also reduces the risk of model overfitting due to redundant data.

[0061] The type determination module is used to determine the emotional type of children with autism based on their contact characteristics.

[0062] In some embodiments, the type determination module is further configured to: Based on the matching degree between autistic children and each type of autism, multiple candidate tests for autistic children are identified; For each candidate autistic child in the test, a contact feature vector is constructed based on the candidate autistic child's historical contact records, the weights of different contact positions of the contact robot, and multiple key feature factors. Based on the contact feature vectors of children with autism and the contact feature vectors of children with autism for each candidate test, children with autism who are similar to the test are selected. Based on the emotion labels of autistic children using similar tests, the emotion types of autistic children are determined.

[0063] Specifically, children with autism who are matched with clusters corresponding to the autism type can be considered as candidates for the test.

[0064] The contact feature vector of a child with autism in the candidate test can include the value of each key feature factor extracted from the historical contact records of the child with autism in the candidate test.

[0065] The cosine similarity between the contact feature vector of a child with autism and the contact feature vector of a child with autism for each candidate test can be calculated. Children with autism who test with a cosine similarity greater than a cosine similarity threshold (e.g., 60%) are considered to be children with autism who test similar to each other.

[0066] Based on the emotion labels of autistic children from similar tests, the percentage of each emotion type was calculated, and the emotion with the highest percentage was identified as the emotion type of autistic children.

[0067] In some embodiments, the type determination module is further configured to: The weight of each key feature factor is determined based on the matching degree between autistic children and each type of autism and the dispersion of each feature factor corresponding to each type of autism. Based on the contact feature vectors of children with autism, the contact feature vectors of children with autism in each candidate test, and the weights of each key feature factor, the similarity between children with autism and children with autism in each candidate test is calculated. Children with autism are selected based on their similarity to children with autism on each candidate test.

[0068] Specifically, for each key feature factor, the matching degree between autistic children and each type of autism is used as a weight. The dispersion of the feature factor corresponding to each type of autism is weighted and summed to obtain the comprehensive dispersion of the key feature factor. The ratio of the comprehensive dispersion of the key feature factor to the sum of the comprehensive dispersions of each key feature factor is used as the weight of the key feature factor.

[0069] Based on the contact feature vectors of the autistic child and the contact feature vectors of the candidate autistic children in the test, the absolute value of the difference between the autistic child and the candidate autistic children in the test for each key feature factor is calculated. The absolute values ​​of the difference between the autistic child and the candidate autistic children in the test for each key feature factor are then weighted and summed according to the weight of each key feature factor to obtain the comprehensive distance between the autistic child and the candidate autistic children in the test. The similarity between the autistic child and each candidate autistic child in the test is calculated based on the comprehensive distance, where the larger the comprehensive distance, the smaller the similarity.

[0070] Children with autism can be screened based on their similarity to children with autism on each candidate test. For example, children with autism on candidate tests can be ranked from highest to lowest similarity, and a certain percentage (e.g., the top 30%) of children with autism on candidate tests can be selected as children with similar autism.

[0071] Candidate children are screened based on matching scores, and key feature factors are extracted from historical contact records to construct contact feature vectors. Cosine similarity calculations are used to quickly identify similar children, and then emotion types are determined based on the proportion of emotion tags, providing foundational data support for personalized intervention. A dynamic adjustment mechanism for feature factor weights is introduced. By combining autism type matching scores and dispersion to calculate the weights of each factor, the similarity assessment more closely reflects actual behavioral differences. Feature factors with high overall dispersion are given higher weights, while factors with strong stability are given lower weights, thereby improving the accuracy of similarity judgment. Furthermore, the absolute value of weighted differences is used to calculate the overall distance and convert it into a similarity value, avoiding the one-sidedness of a single indicator and making the screening of autistic children in similarity tests more consistent with the distribution patterns of behavioral characteristics. Through dynamic weight allocation and multi-factor collaborative analysis, the error problem caused by feature equalization in traditional methods is effectively solved, especially when dealing with the highly heterogeneous autism population, enabling more accurate capture of core differences in behavioral patterns. Ultimately, the analysis of the proportion of emotion labels of autistic children based on similar tests not only reduced the random interference of emotion labels of autistic children based on single similar tests, but also strengthened the reliability of emotion type judgment through the commonality of group behavior.

[0072] The treatment plan determination module is used to determine the treatment plan for children with autism based on their contact characteristics and emotional types.

[0073] In some embodiments, the scheme determination module is further configured to: Construct a predictive model for the proposed solution; Using a predictive model, treatment plans for children with autism are determined based on their TCM syndromes, standardized assessment results, contact characteristics, and emotional types. These plans include at least one TCM massage therapy.

[0074] In some embodiments, the treatment plan prediction model includes a multimodal feature fusion module and a hierarchical decision mechanism. The multimodal feature fusion module is used to fuse the TCM syndromes, standardized assessment results, contact characteristics and emotion types of children with autism to generate a fused feature vector. The hierarchical decision mechanism is used to determine the TCM treatment plan for children with autism through a decision tree.

[0075] Specifically, the multimodal feature fusion module comprises a data preprocessing layer and a feature fusion layer. The data preprocessing layer preprocesses the TCM syndromes, standardized assessment results, contact characteristics, and emotional types of children with autism, for example, by using natural language processing techniques to transform them into quantifiable numerical features. The feature fusion layer employs a multimodal fusion method from deep learning to deeply integrate the four types of preprocessed features. First, a fully connected layer performs preliminary mapping on each type of feature, extracting their respective high-order feature representations. Then, an attention mechanism is used to assign different weights to different features, highlighting features more critical to determining the treatment plan.

[0076] The hierarchical decision-making mechanism uses a fused feature vector as input and a decision tree to determine TCM treatment plans. Its structure includes a decision tree construction layer and a plan generation layer. The decision tree construction layer builds the decision tree based on extensive clinical data and TCM expertise. Using machine learning decision tree algorithms, such as ID3, C4.5, or CART, features from the fused feature vector are used as splitting attributes. The optimal splitting point is selected based on metrics such as information gain, gain ratio, or Gini coefficient, progressively generating nodes and branches of the decision tree. For example, the presence or absence of spleen and stomach weakness in TCM syndromes can be used as the splitting attribute of the root node. If spleen and stomach weakness is detected, the node proceeds to the left subtree for further evaluation of other features; if spleen and stomach weakness is detected, the node proceeds to the right subtree. Through continuous splitting, a logically rigorous and hierarchically distinct decision tree is formed.

[0077] The solution generation layer takes the fused feature vectors as input to the constructed decision tree. Starting from the root node, it traverses the decision tree layer by layer according to the values ​​of each feature in the vectors until it reaches a leaf node. Each leaf node corresponds to a traditional Chinese medicine massage solution. For example, if the leaf node displays "tonify spleen earth," then the thumb pad is massaged with a circular motion to nourish the spleen and stomach and strengthen the body. If it displays "clear liver wood," then the index finger pad is massaged with a straight motion to clear liver heat and relieve symptoms such as irritability. Through this hierarchical and progressive decision-making, the individual differences and complex conditions of children are fully considered to formulate personalized and precise traditional Chinese medicine massage solutions, helping children with autism to recover better.

[0078] Figure 3 These are schematic diagrams of the electronic device according to some embodiments of this specification. It should be noted that... Figure 3 The electronic device shown is merely an example and should not be construed as limiting the functionality or scope of use of the embodiments described in this application. The electronic device is used to operate the aforementioned autism diagnosis and treatment assistance system based on traditional Chinese medicine diagnostic examination and a contact robot.

[0079] like Figure 3As shown, the computer system includes a Central Processing Unit (CPU), which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) or loaded from storage into random access memory (RAM), such as executing the methods described in the above embodiments. Various programs and data required for system operation are also stored in the RAM. The CPU 501, ROM 502, and RAM 503 are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0080] The following components are connected to the I / O interface: input sections including a keyboard, mouse, etc.; output sections 507 including cathode ray tubes (CRTs), liquid crystal displays (LCDs), and speakers, etc.; storage sections including hard disks, etc.; and communication sections including network interface cards such as LAN (Local Area Network) cards and modems, etc. The communication sections perform communication processing via networks such as the Internet. Drives are also connected to the I / O interface as needed. Removable media, such as disks, optical disks, magneto-optical disks, semiconductor memories, etc., are installed on the drive as needed so that computer programs read from them can be installed into the storage section as needed.

[0081] Specifically, according to embodiments of this application, the various modules of the system can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing a computer program for executing a computer program corresponding to an autism diagnosis and treatment assistance system based on traditional Chinese medicine diagnostic examination and contact robots. In such embodiments, the computer program can be downloaded and installed from a network via a communication component, and / or installed from a removable medium. When the computer program is executed by a central processing unit (CPU), it performs various functions defined in the system of this application.

[0082] It should be noted that the computer-readable medium shown in the embodiments of this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying a computer-readable computer program. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wired, etc., or any suitable combination thereof.

[0083] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0084] The units described in the embodiments of this application can be implemented in software or hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the specific unit itself.

[0085] Another aspect of this application provides a computer-readable storage medium storing a computer program thereon, which, when executed by a computer's processor, causes the computer to perform the method as described above. This computer-readable storage medium may be included in the electronic device described in the above embodiments, or it may exist independently and not assembled into the electronic device.

[0086] Finally, it should be understood that the embodiments described in this specification are merely illustrative of the principles of the embodiments described herein. Other variations may also fall within the scope of this specification. Therefore, alternative configurations of the embodiments described herein are intended to be illustrative rather than limiting, and should be considered consistent with the teachings of this specification. Accordingly, the embodiments described herein are not limited to those explicitly introduced and described herein.

Claims

1. An autism diagnosis and treatment assistance system based on traditional Chinese medicine diagnostic examination and contact robots, characterized in that, include: The data acquisition module is used to acquire the TCM syndromes and standardized assessment results of children with autism; The parameter adjustment module is used to determine the parameters of the contact robot based on the TCM syndrome and standardized assessment results of children with autism. The parameters of the contact robot include at least the weights of different contact positions. The record acquisition module is used to acquire historical contact records between autistic children and contact robots; The feature extraction module is used to determine the contact characteristics of children with autism based on their historical contact records with the contact robot and the weights of different contact positions of the contact robot. The type determination module is used to determine the emotion type of children with autism based on their contact characteristics; The treatment plan determination module is used to determine the treatment plan for children with autism based on their contact characteristics and emotional types.

2. The autism diagnosis and treatment assistance system based on traditional Chinese medicine diagnostic examination and contact robot as described in claim 1, characterized in that, The parameter adjustment module is further used for: Obtain TCM syndromes, standardized assessment results, and historical contact records of autistic children from multiple tests; Based on the TCM syndromes of autistic children from multiple tests, standardized assessment results, and historical contact records, multiple types of autism were identified. Based on the TCM diagnostic results and standardized assessment results of children with autism, as well as various autism types, the autism type of children with autism is determined, and based on the autism type of children with autism, the weights of different contact positions of the contact robot are determined.

3. The autism diagnosis and treatment assistance system based on traditional Chinese medicine diagnostic examination and contact robot according to claim 2, characterized in that, The parameter adjustment module is further used for: For each autistic child in the test, the contact frequency for different contact locations is determined based on the child's historical contact records. For any two autistic children, multiple autism types can be identified based on the TCM syndromes of autistic children from multiple tests, standardized assessment results, and the frequency of contact at different contact locations. For each type of autism, the weights of different contact positions of the contact robot corresponding to the autism type are determined based on the contact frequency of different contact positions corresponding to the autism children tested for that autism type.

4. The autism diagnosis and treatment assistance system based on traditional Chinese medicine diagnostic examination and contact robot according to claim 3, characterized in that, The parameter adjustment module is further used for: Based on the TCM syndromes and standardized assessment results of children with autism, the matching degree between children with autism and each type of autism was calculated; The weights of different contact positions of the contact robot are determined based on the matching degree between autistic children and each type of autism, as well as the weights of different contact positions of the contact robot corresponding to each type of autism.

5. The autism diagnosis and treatment assistance system based on traditional Chinese medicine diagnostic examination and contact robot according to claim 4, characterized in that, The feature extraction module is further used for: Based on the weights of different contact positions of the contact robot, key contact records were selected from the historical contact records of autistic children and the contact robot. Based on the historical contact records of autistic children from multiple tests, several key characteristic factors were identified; Based on key contact records and multiple key feature factors of children with autism, a contact feature vector of children with autism is constructed.

6. The autism diagnosis and treatment assistance system based on traditional Chinese medicine syndrome differentiation and physical examination and a contact robot according to claim 5, characterized in that, The feature extraction module is further used for: Multiple characteristic factors are identified, including at least a time-domain characteristic factor, a pressure distribution factor, and a contact trajectory characteristic factor. For each autism type and trait factor, the dispersion of the trait factor corresponding to the autism type is calculated based on the historical contact records of autistic children included in the tests for the autism type. For each autism type, the key feature factors corresponding to the autism type are determined based on the dispersion of the feature factors corresponding to the autism type; Based on the matching degree between autistic children and each autism type, and the key characteristic factors corresponding to each autism type, multiple key characteristic factors are identified.

7. The autism diagnosis and treatment assistance system based on traditional Chinese medicine syndrome differentiation and physical examination and a contact robot according to claim 6, characterized in that, The type determination module is further used for: Based on the matching degree between autistic children and each type of autism, multiple candidate tests for autistic children are identified; For each candidate autistic child in the test, a contact feature vector is constructed based on the candidate autistic child's historical contact records, the weights of different contact positions of the contact robot, and multiple key feature factors. Based on the contact feature vectors of children with autism and the contact feature vectors of children with autism for each candidate test, children with autism who are similar to the test are selected. Based on the emotion labels of autistic children using similar tests, the emotion types of autistic children are determined.

8. The autism diagnosis and treatment assistance system based on traditional Chinese medicine syndrome differentiation and physical examination and a contact robot according to claim 7, characterized in that, The type determination module is further used for: The weight of each key feature factor is determined based on the matching degree between autistic children and each type of autism and the dispersion of each feature factor corresponding to each type of autism. Based on the contact feature vectors of children with autism, the contact feature vectors of children with autism in each candidate test, and the weights of each key feature factor, the similarity between children with autism and children with autism in each candidate test is calculated. Children with autism are selected based on their similarity to children with autism on each candidate test.

9. The autism diagnosis and treatment assistance system based on traditional Chinese medicine syndrome differentiation and physical examination and a contact robot according to any one of claims 1-8, characterized in that, The scheme determination module is further used for: Construct a predictive model for the proposed solution; Using a predictive model, treatment plans for children with autism are determined based on their TCM syndromes, standardized assessment results, contact characteristics, and emotional types. These plans include at least one TCM massage therapy.

10. The autism diagnosis and treatment assistance system based on traditional Chinese medicine syndrome differentiation and physical examination and a contact robot according to claim 9, characterized in that, The proposed prediction model includes a multimodal feature fusion module and a hierarchical decision-making mechanism. The multimodal feature fusion module is used to integrate the TCM syndromes, standardized assessment results, contact characteristics, and emotional types of children with autism to generate a fused feature vector. The hierarchical decision-making mechanism is used to determine the TCM treatment plan for children with autism through a decision tree.

Images