Multi-index man-machine credibility evaluation method

Abstract

The invention discloses a multi-index man-machine trust degree evaluation method which is suitable for evaluating interaction between a user and a rehabilitation nursing robot. According to the method, data of three dimensions of subjective feeling, physiological status and biological feedback are integrated, and user data are collected and analyzed in real time through an electronic questionnaire, a heart rate monitoring belt, a sphygmomanometer, a brain blood oxygen monitor and a sound sensor, a touch sensor and a visual sensor. The comprehensive man-machine credibility is calculated by using a predetermined formula, and results are classified into high, medium and low levels. An evaluation result is presented in the form of numerical values, graphs and characters to form a comprehensive evaluation report. According to the method, real-time and comprehensive credibility feedback can be provided for design and operation of the robot, and optimization of user experience and improvement of acceptability of the robot are facilitated.

Description

Technical Field

[0001] The present invention relates to the technical field of health monitoring, and in particular to a multi-index human-machine trust evaluation method. Background Art

[0002] In modern society, with the development of science and technology, artificial intelligence and robotics have gradually penetrated into people's daily lives, especially in the fields of rehabilitation care, education, and service industries. As an important application form of artificial intelligence, robots are increasingly valued for their interactivity with humans and the effectiveness of task execution. In this interactive process, human-machine trust becomes a key factor, which directly affects users' willingness to use robots and the quality of interaction. Therefore, how to accurately evaluate and improve human-machine trust has become a hot topic and challenge in current research.

[0003] The concept of human-machine trust originates from the trust relationship between people, but its complexity lies in the unpredictability and behavioral transparency of machines. Studies have shown that human-machine trust is not only directly affected by machine performance, but also related to multiple factors such as the user's psychological state, emotional experience, cultural background, and previous interaction experience. In addition, the machine's appearance design, behavioral performance, and the friendliness of the interactive interface may affect the user's sense of trust. Due to the multidimensionality and complexity of these factors, traditional human-machine trust assessment methods often cannot fully reflect the true trust status, which limits the further popularization and application of artificial intelligence products.

[0004] At present, most human-machine trust assessment methods rely mainly on questionnaires or simple experimental observations. These methods often ignore the physiological reactions and biofeedback information of users in actual operations, making the assessment results lack real-time and accuracy. For example, questionnaires alone may not be able to capture the instantaneous physiological changes of users when interacting with robots, such as physiological signals such as accelerated heart rate and increased blood pressure, which are important indicators for assessing true trust.

[0005] In view of this, it is particularly important to develop a comprehensive multi-index human-robot trust evaluation method. This method should be able to collect and analyze the user's subjective feelings, physiological state, and biofeedback in real time during the interaction with the robot, and more accurately evaluate the human-robot trust by integrating these data. Specifically, subjective feelings can be quantified through an improved questionnaire system, the evaluation of physiological state should include the monitoring of heart rate, blood pressure, and cerebral blood oxygen level, and biofeedback involves the analysis of the user's sound, touch, and visual responses.

[0006] In addition, current evaluation systems often lack adaptive adjustment mechanisms. In practical applications, the user's trust may fluctuate due to the robot's different performance or changes in the external environment. Therefore, an ideal human-machine trust evaluation system should have the ability to learn and self-adjust, and be able to automatically optimize the evaluation model based on user feedback and historical interaction data.

[0007] In summary, the present invention provides a multi-indicator human-machine trust evaluation method, which can not only comprehensively analyze various types of data, but also enhance the accuracy and practicality of the evaluation through an adaptive adjustment mechanism, providing scientific and comprehensive trust feedback for robot designers and developers, thereby optimizing human-machine interaction design and improving user satisfaction and the practicality of the robot. Summary of the invention

[0008] The purpose of the present invention is to provide a multi-index human-machine trust evaluation method to solve the problems raised in the above background technology.

[0009] To achieve the above object, the present invention provides the following technical solution: a multi-index human-machine trust evaluation method, comprising the following steps:

[0010] Establish a human-machine trust evaluation index system that includes three dimensions: subjective feelings, physiological state, and biofeedback;

[0011] Using software automation tools, data on subjective feelings, including but not limited to emotional reactions and trust perception, are collected in real time during the user's interaction with the rehabilitation care robot, and a subjective trust score is obtained through automatic analysis by an electronic questionnaire system;

[0012] The user's physiological status data, such as heart rate, blood pressure and brain blood oxygen level, is monitored and recorded in real time through heart rate monitoring belts, blood pressure monitors and brain blood oxygen monitor sensor devices. The obtained data is analyzed by the processing server and the physiological trust score is calculated according to a predetermined formula;

[0013] Collect and analyze the user's auditory, tactile and visual feedback on the robot, including the voice response time recorded by the sound sensor, the touch intensity and duration measured by the pressure sensor, and the gaze track and duration recorded by the eye tracker, and calculate the biofeedback trust score according to a predetermined formula;

[0014] The overall human-machine trust is calculated comprehensively according to the following formula:

[0015] T total =α·T subjective +β·T physys +γ·T bio

[0016] Among them, T subjective ,Tphysys ,T bio They represent the trust obtained from subjective feelings, physiological state and biofeedback respectively, and α, β, and γ are their respective weight coefficients;

[0017] Rating is performed based on the calculated overall human-machine trust score, with the trust divided into three levels: high, medium, and low, and a comprehensive evaluation report is generated including numerical values, graphics, and text descriptions.

[0018] As a preferred technical solution of the present invention: the subjective feeling evaluation is carried out using an electronic questionnaire including ten questions, each question is designed based on the human-machine trust scale and the human-human trust scale, and the user's choice is analyzed by an algorithm to calculate the subjective trust score.

[0019] As a preferred technical solution of the present invention: the evaluation of the physiological state adopts at least three physiological monitoring devices to collect data, specifically a heart rate monitoring belt to measure heartbeat changes, a sphygmomanometer to measure blood pressure levels, and a brain blood oxygen monitor to measure brain blood oxygen saturation. The data are weighted and processed to synthesize a physiological trust score.

[0020] As a preferred technical solution of the present invention: the specific calculation formula of the physiological trust is as follows:

[0021] T phys =0.3·HR+0.4·BP+0.3·Hbo

[0022] Among them, HR represents the normalized heart rate change value, BP represents the normalized blood pressure value, Hbo represents the normalized cerebral blood oxygen value, and the weights of each item are determined according to pre-set standards.

[0023] As a preferred technical solution of the present invention: the biofeedback evaluation collects data through dedicated sensor devices, namely, a sound sensor for the response time of voice commands, a pressure sensor for recording touch intensity and time, and an eye tracker for tracking eye gaze points and duration.

[0024] As a preferred technical solution of the present invention: the calculation formula of the biofeedback trustworthiness is:

[0025] T bio =0.2·T audio +0.5·T tactile +0.3·T visual

[0026] Among them, T audio Score auditory feedback, T tactile Score the tactile feedback, T visual Visual feedback is scored, and weights are set based on experimental data and expert evaluation.

[0027] As a preferred technical solution of the present invention: the total score of the human-machine trust is calculated based on a comprehensive scoring formula, and the trust is graded into three levels: high, medium and low according to the scoring results for subsequent analysis and user understanding.

[0028] As a preferred technical solution of the present invention: the generation of the evaluation report includes numerical scores, graphical displays and detailed text descriptions, which can be displayed or printed in electronic format through the user interface so that users can choose the viewing method according to their personal preferences.

[0029] As a preferred technical solution of the present invention: the physiological status data are all encrypted and transmitted to the server through a secure network protocol to ensure the security and privacy protection of user data.

[0030] As a preferred technical solution of the present invention: the software automation tool includes an adaptive adjustment mechanism based on user feedback, which automatically adjusts weights and parameters according to user usage feedback to optimize the evaluation model and improve the trust of human-computer interaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 This is a flow chart of a multi-index human-machine trust evaluation method proposed by the present invention.

[0032] Beneficial effects:

[0033] Comprehensive evaluation framework: This method provides a comprehensive perspective for evaluating users' trust in robots by combining three different dimensions: subjective feelings, physiological state, and biofeedback. This multi-dimensional evaluation method is more comprehensive than a single subjective questionnaire or simple physiological measurement, and can more realistically reflect the user's trust state.

[0034] Real-time data collection and analysis: By leveraging modern sensors and analytical techniques, this method can collect and process user data in real time. This not only increases the timeliness of the evaluation, but also allows the robotic system to adjust its behavior based on immediate feedback, thereby improving the user's trust in real time.

[0035] Customized evaluation and application: The weight coefficients included in the evaluation model can be adjusted according to different application scenarios and user needs. This flexibility allows the evaluation method to be widely used in a variety of fields, from rehabilitation care to educational assistance, to optimize the human-computer interaction experience for specific applications.

[0036] Enhanced user experience and satisfaction: By comprehensively evaluating users’ subjective and physiological responses, robots can better understand and adapt to users’ needs and preferences, thereby providing more personalized interactions, enhancing user experience and improving user satisfaction.

[0037] Security and Privacy Protection: This approach includes measures to encrypt the collected data and transmit it through secure protocols, ensuring the security and privacy of user data, which is crucial to enhancing user trust in the use of robotics technology.

[0038] Providing decision support: By generating a comprehensive evaluation report including numerical scores, graphical displays, and text descriptions, this method not only helps users understand the quality of their interaction with the robot, but also provides practical data support for robot designers and operators to help them optimize product design and services. Description of the drawings:

[0040] Figure 1 This is a flowchart of a multi-index human-machine trust evaluation method proposed by the present invention. DETAILED DESCRIPTION

[0041] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] See also Figure 1 ,

[0043] The present invention provides a multi-index human-machine trust evaluation method, which aims to evaluate the user's trust in the robot by integrating the user's subjective feelings, physiological state and biofeedback data. This method is particularly suitable for rehabilitation and nursing robots, and can help improve the design of robots and increase user acceptance and satisfaction.

[0044] Example 1: Rehabilitation Nursing Application

[0045] In a rehabilitation center, a rehabilitation care robot is used to assist elderly people with limited mobility in their daily activities. In order to evaluate the elderly's trust in the robot, the following steps were used:

[0046] Subjective evaluation: This is done through an electronic questionnaire consisting of ten questions, each of which is designed based on the human-machine trust scale and the human-human trust scale. Each choice made by the user is analyzed by an algorithm to calculate a subjective trust score.

[0047] Physiological status assessment: Use a heart rate monitor to measure heart rate changes, a sphygmomanometer to measure blood pressure levels, and a cerebral blood oxygen monitor to measure brain blood oxygen saturation. These data are transmitted to the processing server in real time and the physiological trust score is calculated according to a preset formula.

[0048] Biofeedback evaluation: The user’s response to the robot is collected through sound sensors, pressure sensors, and eye trackers. The evaluation includes the response time to voice commands, touch intensity and duration, and eye gaze point and duration.

[0049] Comprehensive trust evaluation: T total =α·T subjective +β·T physys +γ·T bio

[0050] Among them, T subjective ,T physys ,T bio They represent the trust obtained from subjective feelings, physiological state and biofeedback respectively, and α, β, γ are their respective weight coefficients.

[0051] Comprehensively calculate the overall human-machine trust.

[0052] Result analysis and report generation: Rating is performed based on the calculated overall human-machine trust score, and a comprehensive evaluation report including numerical values, graphics, and text descriptions is generated.

[0053] Example 2: Educational Assistance Robot

[0054] In one school, an educational assistive robot was used to help students learn science experiments. To assess students’ trust in the robot, the following steps were performed:

[0055] Subjective experience evaluation: Students filled out an electronic questionnaire after interacting with the robot. The questionnaire was designed taking into account the students’ age and educational background.

[0056] Physiological status assessment: During the experimental activities, students' heart rate and blood pressure were monitored through wearable devices, and their brain blood oxygen levels were measured through non-invasive sensors.

[0057] Biofeedback assessment: Analyze students’ vocal responses, tactile interactions, and visual attention while operating the robot.

[0058] Comprehensive trust evaluation: Based on the collected data, the comprehensive trust is calculated using the adjusted weight factors.

[0059] Result analysis and reporting: Based on the evaluation results, detailed reports are provided to teachers and school management to help them understand the effectiveness of students’ interaction with the robot.

[0060] Example 3: Application of medical assistance robots in hospitals

[0061] In hospital environments, medical assistive robots are used to help patients with rehabilitation training and daily care. The human-machine trust assessment of such robots includes the following steps:

[0062] Subjective experience evaluation: Develop a questionnaire specifically for medical rehabilitation scenarios to collect patients' feedback on the robot's performance, interaction methods, and helpfulness. The questionnaire is distributed through the hospital's internal network electronic system and integrated into the patient's daily feedback program.

[0063] Physiological status assessment: During the patient's interaction with the robot, the physiological monitoring system embedded in the rehabilitation equipment collects the patient's heart rate, blood pressure, and brain blood oxygen level in real time. These data are transmitted to the hospital's data processing center through a secure medical network for analysis.

[0064] Biofeedback assessment: Utilizing sensor technology on the rehabilitation robot to record the patient’s tactile response to the robot (e.g., grip strength) during physical rehabilitation, as well as analyzing the patient’s visual gaze and reaction time through a camera system installed in the rehabilitation room.

[0065] Comprehensive trust evaluation: Based on the collected subjective feelings, physiological status and biofeedback data, the comprehensive human-machine trust is calculated through advanced data analysis software. The evaluation model adjusts weights and parameters according to the specific needs of the medical scenario.

[0066] Result analysis and report generation: Generate detailed trust assessment reports, including statistical charts, patient feedback and suggested improvement measures, for reference by the medical team to optimize robot design and improve patient satisfaction.

[0067] Example 4: Application of service robots in retail environments

[0068] In retail environments, service robots are used to assist customers in selecting products and providing information query services. The specific steps of trust evaluation are as follows:

[0069] Subjective experience evaluation: Develop a concise electronic questionnaire that is integrated into interactive terminals and mobile applications in shopping malls to ask customers about their views on the robot’s helpfulness, interactive experience, and satisfaction.

[0070] Physiological state assessment: During the interaction with customers, the environmental perception technology equipped by the service robot can perceive the basic physiological reactions of customers, such as estimating emotional states through facial recognition technology.

[0071] Biofeedback Assessment: Analyzes the customer’s vocal arousal and eye movement rate when interacting with the robot to assess customer engagement and interest.

[0072] Comprehensive trust assessment: Use adaptive algorithms to comprehensively analyze data collected from questionnaires, physiological and biofeedback, and adjust the trust assessment model in real time to match different customer groups and situations.

[0073] Results Analysis and Report Generation: Provides retail management teams with detailed reports including customer feedback, trust ratings, and feedback on service robot performance.

[0074] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0075] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A multi-index human-machine trust evaluation method, characterized in that: The steps include: Establish a human-machine trust evaluation index system that includes three dimensions: subjective feelings, physiological state, and biofeedback; Using software automation tools, data on subjective feelings, including but not limited to emotional reactions and trust perception, are collected in real time during the user's interaction with the rehabilitation care robot, and a subjective trust score is obtained through automatic analysis by an electronic questionnaire system; The user's physiological status data, such as heart rate, blood pressure and brain blood oxygen level, is monitored and recorded in real time through heart rate monitoring belts, blood pressure monitors and brain blood oxygen monitor sensor devices. The obtained data is analyzed by the processing server and the physiological trust score is calculated according to a predetermined formula; Collect and analyze the user's auditory, tactile and visual feedback on the robot, including the voice response time recorded by the sound sensor, the touch intensity and duration measured by the pressure sensor, and the gaze track and duration recorded by the eye tracker, and calculate the biofeedback trust score according to a predetermined formula; The overall human-machine trust is calculated comprehensively according to the following formula: T total =α·T subjective +β·T phys +γ·T bio Among them, T subjective ,T phys ,T bio They represent the trust obtained from subjective feelings, physiological state and biofeedback respectively, and α, β, and γ are their respective weight coefficients; Rating is performed based on the calculated overall human-machine trust score, with the trust divided into three levels: high, medium, and low, and a comprehensive evaluation report is generated including numerical values, graphics, and text descriptions.

2. A multi-index human-machine trust evaluation method according to claim 1, characterized in that: The subjective feelings are evaluated using an electronic questionnaire including ten questions, each question is designed based on a human-machine trust scale and a human-human trust scale, and the user's choices are analyzed by an algorithm to calculate a subjective trust score.

3. A multi-index human-machine trust evaluation method according to claim 1, characterized in that: The assessment of the physiological state uses at least three physiological monitoring devices to collect data, specifically a heart rate monitor to measure heartbeat changes, a sphygmomanometer to measure blood pressure levels, and a brain blood oxygen monitor to measure brain blood oxygen saturation. The data is weighted and processed to synthesize a physiological confidence score.

4. A multi-index human-machine trust evaluation method according to claim 1, characterized in that: The specific calculation formula of the physiological trust is as follows: T phys =0.3·HR+0.4·BP+0.3·Hbo Among them, HR represents the normalized heart rate change value, BP represents the normalized blood pressure value, Hbo represents the normalized cerebral blood oxygen value, and the weights of each item are determined according to pre-set standards.

5. The multi-index human-machine trust evaluation method according to claim 1 is characterized in that: The biofeedback assessment collects data through dedicated sensor devices, including sound sensors for response time to voice commands, pressure sensors for recording touch intensity and duration, and eye trackers for tracking eye gaze point and duration.

6. A multi-index human-machine trust evaluation method according to claim 1, characterized in that: The calculation formula of the biofeedback trust is: T bio =0.2·T audio +0.5·T tactile +0.3·T visual Among them, T audio Score auditory feedback, T tactile Score the tactile feedback, T visual Visual feedback is scored, and weights are set based on experimental data and expert evaluation.

7. A multi-index human-machine trust evaluation method according to claim 1, characterized in that: The total score of the human-machine trust is calculated based on a comprehensive scoring formula, and the trust is graded into three levels: high, medium, and low according to the scoring results for subsequent analysis and user understanding.

8. The multi-index human-machine trust evaluation method according to claim 1 is characterized in that: The evaluation report includes numerical scores, graphical displays, and detailed text descriptions, and can be displayed or printed in electronic format through a user interface so that users can choose a viewing method based on their personal preferences.

9. A multi-index human-machine trust evaluation method according to claim 1, characterized in that: The physiological status data are encrypted and transmitted to the server via a secure network protocol to ensure the security and privacy of user data.

10. The multi-index human-machine trust evaluation method according to claim 1, characterized in that: The software automation tool includes an adaptive adjustment mechanism based on user feedback, which automatically adjusts weights and parameters according to user usage feedback to optimize the evaluation model and improve the trust of human-computer interaction.

Images