Massage robot compliance control method with rigidity self-adaption function

Abstract

The invention discloses a compliance control method for a massage robot with a rigidity self-adaption function. The massage robot is composed of a six-degree-of-freedom mechanical arm, an end effector, a six-dimensional force sensor, a control host and other modules. The six-dimensional force sensor is installed at the tail end, the contact force of the robot and the surface of the human body can be collected in real time, tail end motion data obtained through the pose sensor is combined, the system executes a compliant control closed loop, and self-adaptive and safe massage operation is achieved. According to the method part, a force control closed loop can be constructed in real time, accurate, safe and smooth massage operation can be achieved through a force sensor installed at the tail end of the robot in combination with a smooth control algorithm based on displacement errors without depending on a visual system, meanwhile, massage area rigidity detection and self-adaptive control are achieved, and the user experience is effectively improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of robot control, and particularly relates to a massage robot compliant control method with stiffness self-adaptive function. BACKGROUND

[0002] With the wide application of intelligent robots in the fields of medical treatment, rehabilitation and life service, massage robots gradually become a kind of potential auxiliary equipment. Most of the existing massage robots rely on visual sensing technology for human body part recognition and path planning, and the system cost is high, the configuration is complex, and there are uncertainty problems such as occlusion and illumination change, which limit the universality of the equipment and the interactive convenience of the user. At the same time, the traditional control method is mostly open loop or semi-closed loop control, without real-time force feedback information, lacking real-time response ability to the applied force and user reaction, and easily causing excessive massage force or inaccuracy, causing damage to the user's body.

[0003] In addition, the current massage robot still lacks the detection and adaptive adjustment ability of the stiffness of the massaged area, not only cannot adjust the force according to the user's body muscle feedback, but also cannot intelligently adjust the force according to the soft and hard degree of different parts of the body, and there are problems of insufficient safety and comfort.

[0004] Therefore, it is necessary to propose a compliant control method without relying on a visual system, capable of establishing a closed-loop control system through force sensing and displacement feedback, and automatically detecting and adapting to local stiffness, to improve the intelligent level and human-computer interaction naturalness of the massage robot. SUMMARY

[0005] In order to solve the above difficulties and problems existing in the prior art, the present application innovatively proposes a massage robot compliant control method with stiffness self-adaptive function, which can real-time construct a force control closed loop, through a force sensor installed at the end of the robot, combined with a compliant control algorithm based on displacement error, to realize precise, safe and compliant massage operation without relying on a visual system, and realize stiffness detection and self-adaptive control of the massage area, effectively improving the user experience.

[0006] The purpose of the present application is to propose a massage robot compliant control method with stiffness self-adaptive function, and the specific scheme is as follows:

[0007] Step 1: force sensor installation and force data acquisition

[0008] A six-axis force sensor is installed at the end of a six-axis mechanical arm through a self-made flange, and is rigidly connected with an end effector through a self-made mechanism to ensure the accuracy of force signal transmission. During the execution of the massage task by the robot, the sensor real-time collects six-axis force data of the contact between the end of the robot and the human body as the basis for subsequent compliant control input;

[0009] Step 2: Pose information acquisition:

[0010] In the case of a robot with six degrees of freedom, the end effector is composed of position and attitude, and the attitude change is mapped to angular velocity using the Euler angle modeling method. To establish the mapping relationship, the end effector attitude rotation matrix needs to be analyzed, and the Euler angle velocity transformation model is introduced. The attitude adopts the ZYX Euler angle modeling method, which means rotating around Z first, then Y, and finally X;

[0011] Step 3: Building a compliant control model:

[0012] The compliant control algorithm is introduced, combining impedance control and admittance control mechanisms. The six-axis force sensor installed at the end of the robot collects real-time interaction force information when the robot contacts the human body. Based on the error between the desired pose and the actual pose of the end effector, the system dynamically calculates the required response behavior including displacement and attitude, and generates feedback control instructions accordingly to achieve compliant response effects that meet human-robot interaction requirements, and builds a complete force-position closed-loop control system.

[0013] Step 4. Actuator response and closed-loop control implementation:

[0014] According to the control target calculated in step 3, the controller generates the next period of end effector desired motion instructions, and controls the joint drive to execute the motion of the actuator, realizing complete force-position closed-loop control. This closed-loop does not rely on visual information and is completely based on force and pose feedback to achieve compliant interaction.

[0015] Step 5. Stiffness detection and dynamic response adjustment:

[0016] The system calculates the current contact area stiffness based on the interaction force F and the pose change Δx in each sampling period, based on the formula: k = F / Δx, where k is the stiffness of the human tissue. The stiffness result is used to adjust the controller stiffness and damping parameters in real time to adapt to the softness of the tissue and optimize the massage response characteristics.

[0017] Step 6. Introducing model reference adaptive control mechanism:

[0018] The model reference adaptive control strategy is introduced to optimize and dynamically adjust the parameters in the compliant control. This method can adaptively adjust the control parameters according to the error between the target model and the actual system, ensuring that the robot can maintain the desired dynamic response characteristics when facing different stiffness environments.

[0019] Step 7. Integration of voice control and intelligent interaction:

[0020] To improve the convenience and naturalness of human-computer interaction, so that users can remotely control or input high-level instructions to the massage task through natural language, the module can perform semantic analysis on the instructions issued by the user based on the trained voice recognition neural network model, and convert it into a standard instruction set executable by the robot control system. The system processes the voice signal through the voice large model;

[0021] Step 8: Control flow integration and overall scheduling;

[0022] The above control modules are integrated in the control host and run, and each hardware module communicates with the host through Ethernet. The entire control system can realize autonomous identification of the massage area, adaptive adjustment of the force, and natural language interaction, forming a complete closed-loop control process without the assistance of a visual system.

[0023] As a further improvement of the application, the step 2 Euler angle modeling is as follows:

[0024] Euler angle (θ, φ, ψ) , , Corresponding angular velocity The relationship between its derivative can be written as:

[0025] , ;

[0026] , ;

[0027] This matrix maps the Euler angle derivative To the angular velocity In the rigid body coordinate system, which is used for the angular velocity input of the attitude error feedback;

[0028] To facilitate the control system to deduce the Euler angle derivative from the angular velocity collected by the sensor, the inverse of the transformation matrix is often used in numerical control:

[0029] = .

[0030] As a further improvement of the application, the step 3 is as follows:

[0031] 3.1 Impedance model:

[0032] Simulate the dynamic behavior of a mass-spring-damper system to describe the target response state input of the end when subjected to external force. The input is the desired pose, and the output is the control force;

[0033] The general form of the impedance model is:

[0034] ;

[0035] wherein:

[0036] ;

[0037] ;

[0038] ;

[0039] The second derivative represents the error between the desired end-effector acceleration and the actual acceleration, the first derivative represents the error in velocity, and the zeroth derivative represents the static error in position and orientation, represents the desired position;

[0040] : are the desired mass, damping, and stiffness matrices, respectively;

[0041] : is the calculated force generated by the robot end-effector;

[0042] By adjusting the impedance parameters, including the stiffness , the response speed and compliance can be controlled, allowing the robot to adapt to different regions or environments;

[0043] 3.2 Admittance Control:

[0044] Unlike impedance control, admittance control takes external force as input and outputs the desired displacement change, which can be mathematically expressed as:

[0045] ;

[0046] : is the equivalent mass matrix set by the admittance model;

[0047] : is the equivalent damping matrix, used to control the smoothness of the response speed;

[0048] : is the equivalent stiffness matrix, used to set the displacement sensitivity of the end-effector to external force;

[0049] , , : are the end-effector acceleration, velocity, and displacement, respectively;

[0050] : ;

[0051] 3.3 Dynamics Model Compensation:

[0052] The controller converts the end-effector force to joint space torque through Jacobian matrix, combined with the robot dynamics model for compensation, which is as follows:

[0053] ;

[0054] : control torque in joint space

[0055] : joint angle

[0056] : joint angular velocity vector

[0057] : joint angular acceleration vector

[0058] : inertia matrix in joint space, describing the effect of mass distribution on motion

[0059] : Coriolis / centrifugal force matrix, describing the dynamics effect when velocity changes

[0060] : gravity term, representing the torque generated at each joint due to gravity

[0061] Torque-based control implementation:

[0062] : transpose of Jacobian matrix from joint to Cartesian space, used to convert force in Cartesian space to torque in joint space

[0063] As a further improvement of the present application, the step 6 is specifically as follows:

[0064] 6.1. Model reference structure setting

[0065] ;

[0066] where , , are the reference mass, damping and stiffness matrices respectively, is the desired output trajectory. The actual system state is X, and the state error of the two is defined as:

[0067] -X

[0068] 6.2. Lyapunov function construction and stability analysis

[0069] A Lyapunov function of the following form is constructed to analyze the stability of the system and derive the parameter update law:

[0070] ;

[0071] wherein : is the impedance stiffness estimation error;

[0072] >0: is the adjustment gain, affecting the parameter convergence speed;

[0073] : is the desired system mass matrix;

[0074] Differentiating V with respect to time and combining with the system dynamics model, the stable adaptive rate is obtained;

[0075] 6.3. Adaptive rate design;

[0076] According to the actual detection of external force and the pose error , the adaptive parameter adjustment rate is designed as follows;

[0077] Stiffness adaptive law:

[0078]

[0079] Damping adaptive rate:

[0080]

[0081] wherein , >0: is the adjustment gain parameter;

[0082] f() is the speed error response function;

[0083] All adjustment rates are provided with physical constraint range[ ] to prevent stiffness from being too large or oscillation from occurring.

[0084] As a further improvement of the present application, the step 7 is specifically as follows:

[0085] 7.1 Voice acquisition and feature extraction;

[0086] The user inputs natural language instructions through the microphone, and the system acquires voice data and extracts mel spectrum, MFCC or other high-dimensional feature vectors as model input;

[0087] 7.2 Instruction recognition and intent analysis;

[0088] The system maps the voice information into a standard semantic structure through the trained voice recognition model, and further converts it into a predefined operation label through the intent recognition module;

[0089] 7.3 Control instruction generation and integration;

[0090] The conversion of the voice instruction into a control instruction is encapsulated as a standard control function, which is embedded into the current compliant control flow. Specifically, the controller receives the control function instruction, the controller updates the function parameters, the controller receives the instruction and passes it to the actuator, the actuator executes, and feedback is obtained until the user says OK or the flow is stopped.

[0091] 7.4 Control flow integration;

[0092] After the voice-converted control instruction is encapsulated, it is embedded into the compliant control flow and runs cooperatively with the end force feedback, stiffness detection, and adaptive adjustment module to ensure that the system behavior remains dynamically stable and interactive compliant under the influence of language input.

[0093] Compared with the prior art, the present application realizes significant innovation in structure design and control strategy. First, a complete force control closed-loop system based on compliant control is constructed, which enables the massage robot to perceive and respond to external interaction force in real time through the six-axis force sensor and pose error feedback mechanism, thereby realizing safe and precise force control operation. Second, by introducing the stiffness detection mechanism and model reference adaptive control method, the system can automatically adjust the control parameters based on the softness of the contact site, making the massage force of the robot more adaptive to the actual physiological characteristics of the human body, effectively improving the user's comfort and personalized experience. In addition, the system completely eliminates the dependence on traditional visual recognition modules, avoiding recognition errors caused by obstacles, lighting, and other problems, which not only greatly reduces the system cost and structural complexity, but also simplifies the use of the device. Further, the present application integrates a large model control interface based on voice recognition and semantic understanding, allowing users to input remote instructions or interact in real time through natural language, effectively enhancing the naturalness, intelligence, and convenience of human-computer interaction, and comprehensively improving the practical value and technical competitiveness of the massage robot in medical, rehabilitation, and home application scenarios. BRIEF DESCRIPTION OF DRAWINGS

[0094] Figure 1 : Schematic diagram of the overall structure of the massage robot;

[0095] Figure 2 : Compliant control system block diagram;

[0096] Figure 3 : Adaptive control flowchart;

[0097] Figure 4 : User interaction and response logic flowchart. DETAILED DESCRIPTION

[0098] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0099] The massage robot system of the present invention is as follows Figure 1 As shown, the system consists of a six-degree-of-freedom robotic arm, an end effector, a six-dimensional force sensor, and a control host. The six-dimensional force sensor installed at the end collects real-time contact force between the robot and the human body surface. Combined with the end motion data obtained by the posture sensor, the system executes a compliant control closed loop, achieving adaptive and safe massage operations.

[0100] After the robot enters the mission mode, the control system Figure 2 As shown in Figure 1, the impedance-admittance control structure is used to adjust the posture and force. The control system uses the end force information F and displacement error , construct a compliant control model and obtain the target dynamic behavior through the following sub-expression:

[0101]

[0102]

[0103]

[0104] in : are the desired mass, damping and stiffness matrices respectively, The equivalent mass matrix specified for the admittance model, is the equivalent damping matrix, which is used to control the response smoothing speed. is the equivalent stiffness matrix, which is used to set the displacement sensitivity of the end to the external force. , , : are the acceleration, velocity and displacement of the terminal, respectively. The d in the lower right corner indicates the desired position. Based on the above control objectives, the system generates the desired terminal position for the next cycle. The instructions are sent to the lower computer to drive the actuator to achieve precise movement.

[0105] During the continuous contact between the robot and the human body, the system obtains the interaction force F and the end displacement error in real time according to each sampling cycle. , calculate the stiffness value of the current contact area:

[0106]

[0107] The stiffness estimation will be used to dynamically adjust the control model parameters. When the muscle in the detection area is relatively stiff, the system automatically increases the damping coefficient and reduces the stiffness coefficient to avoid the discomfort caused by the violent response of the control system; when it is detected that the tissue is relatively soft, the stiffness and control accuracy are appropriately improved, and the massage depth and effectiveness are improved.

[0108] To enhance the response capability of the system in complex environments, the model reference adaptive control (MRAC) method is further introduced, as shown in Figure 3 The target reference model is:

[0109]

[0110] The actual system state is X, the error is -X, and the Lyapunov function is constructed

[0111]

[0112] According to the stability analysis, the adaptive law is as follows

[0113] Stiffness update law:

[0114]

[0115] Damping update rate:

[0116]

[0117] wherein, is the adjustment gain, f() is the speed error response function, and all updated parameters are limited in the physical safety range to avoid control instability caused by excessively high stiffness or response overshoot.

[0118] Finally, the system dynamically adjusts the control gain according to the estimated stiffness, so that the massage robot can form an adaptive compliant adjustment strategy for different human body regions during execution. Then the user can interact with the robot according to the trained voice module, as shown in Figure 4 Because the voice module has been embedded into the compliant control process, the personalized interaction of "massage where, adapt where" can be finally realized.

[0119] The entire control process does not depend on external vision systems at all. The user only needs to guide the mechanical arm to the target massage area or use drag, voice instructions, teaching points and other interactive methods to start control. The mechanical arm can independently plan the massage strategy and execute it on the basis of force feedback and stiffness judgment, completing an intelligent massage process with compliance, safety and high interaction.

[0120] The above shows and describes the basic principles and main features of the present application and the advantages of the present application. Those skilled in the art should understand that the present application is not limited to the above embodiments, and the above embodiments and descriptions in the specification are only to illustrate the principles of the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the claimed present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. A compliant control method for a massage robot with a stiffness adaptive function, characterized in that: The specific plan is as follows: Step 1: Force sensor installation and force data acquisition: The six-dimensional force sensor is installed on the end of the six-degree-of-freedom robotic arm through a self-made flange, and is rigidly connected to the end effector through a self-made mechanism to ensure the accuracy of force signal transmission. When the robot performs massage tasks, the sensor collects six-dimensional force data of the contact between the end of the robot and the human body in real time, which serves as the input basis for subsequent compliant control. Step 2: Get pose information: When the robotic arm has six degrees of freedom, the end is composed of position and posture. Using the Euler angle modeling method, a clear mapping relationship is established between posture changes and angular velocity. To establish this mapping relationship, it is necessary to analyze the end posture rotation matrix and introduce the Euler angular velocity transformation model. The posture is modeled using the Euler angle ZYX method, which means first rotating around Z, then around Y, and finally around X. Step 3: Construct a compliant control model: A compliant control algorithm is introduced, combining impedance control with admittance control mechanisms. A six-dimensional force sensor mounted on the end-mounted robot collects real-time information on the interaction force generated when the robot contacts the human body. Based on the error between the desired and actual end-mounted postures, the system dynamically calculates the required response behavior, including displacement and posture, and generates feedback control instructions accordingly, achieving a compliant response that meets the requirements of human-machine interaction and building a complete force-position closed-loop control system. Step 4. Actuator response and closed-loop control implementation; Based on the control target calculated in step 3, the controller generates the desired motion command at the end of the next cycle and controls the actuator to execute the motion through the joint driver, thus achieving a complete force-posture closed-loop control. This closed-loop does not rely on visual information and achieves compliant interaction based entirely on force and posture feedback. Step 5. Stiffness detection and dynamic response adjustment: During each sampling period, the system calculates the stiffness of the current contact area based on the interaction force F and the posture change Δx. This is based on the formula: k = F / Δx, where k is the stiffness of human tissue. The stiffness result is used to adjust the controller stiffness and damping parameters in real time to adapt to the softness and hardness of the tissue and optimize the massage response characteristics. Step 6. Introduce model reference adaptive control mechanism: A model-referenced adaptive control strategy is introduced to perform online optimization and dynamic adjustment of compliant control parameters. This method can adaptively adjust control parameters based on the error between the target model and the actual system, ensuring that the robot maintains the desired dynamic response characteristics when facing different stiffness environments. Step 7. Integrate voice control and intelligent interaction; To improve the convenience and naturalness of human-computer interaction, allowing users to remotely control massage tasks or input high-level commands through natural language, this module, based on a trained voice recognition neural network model, can semantically parse user commands and convert them into a standard instruction set executable by the robot control system. The system then processes the voice signal through a large voice model. Step 8: Control process integration and overall scheduling; All of the above control modules are integrated into the control host and run. Each hardware module communicates with the host via Ethernet. The entire control system can realize autonomous identification of massage areas, adaptive adjustment of force, and natural language interaction, forming a complete closed-loop control process that does not require the assistance of a visual system.

2. The massage robot compliance control method with stiffness adaptive function according to claim 1 is characterized in that: The Euler angle modeling in step 2 is specifically as follows: Euler angles ( , , The corresponding angular velocity The relationship between it and its derivative can be written as: , ； , ； This matrix converts the Euler angle derivatives Mapped to the angular velocity in the rigid body coordinate system , used to control the angular velocity input of attitude error feedback; In order to facilitate the control system to infer the Euler angle derivative from the angular velocity collected by the sensor, the inverse of the transformation matrix is ​​often used in numerical control: = 。 3. The massage robot compliance control method with stiffness adaptive function according to claim 1, characterized in that: The step 3 is specifically as follows: 3.1 Impedance model: Simulate the dynamic behavior of the mass-spring-damper system to describe the target response state input of the end when subjected to external force. The input is the desired posture and the output is the control force. The general form of the impedance model is; ； in: ； ； ； The second-order derivative is the error between the desired acceleration and the actual acceleration of the terminal. The first-order derivative represents the error in velocity. The zero-order derivative represents the static error between the terminal position and attitude. represents the desired position; : are the desired mass, damping and stiffness matrices respectively; : The calculated force generated by the robot end; By adjusting the impedance parameters, including stiffness , to achieve the regulation of response speed and compliance, so that the robot has sufficient adaptability in different areas or environments; 3.2 Admittance control: Different from impedance control, admittance control takes external force as input and outputs the desired displacement change. Its mathematical form is: ； : Equivalent mass matrix set by the admittance model; : Equivalent damping matrix, used to control the response smoothing speed; : Equivalent stiffness matrix, used to set the displacement sensitivity of the end to external forces; , , : are the terminal acceleration, velocity and displacement respectively; ： ； 3.

3. Dynamic model compensation: The controller converts the end force into joint space torque through the Jacobian matrix and compensates it in combination with the robot dynamics model. The dynamics model is as follows: ； : Control torque in joint space : joint angle; : joint angular velocity vector; : joint angular acceleration vector; : The inertia matrix of the joint space, which describes the influence of the robot's mass distribution on the motion; : Coriolis force / centrifugal force matrix, describing the dynamic effects of speed changes; : Gravity term, which represents the torque generated by gravity on each joint; Torque-based control implementation: ; : The transpose of the Jacobian matrix of the joint to Cartesian space, used to convert the force in Cartesian space to the moment in joint space.

4. The massage robot compliance control method with stiffness adaptive function according to claim 1, characterized in that: The step 6 is specifically as follows: 6.

1. Model reference structure setting; ； in 、 、 are the reference mass, damping and stiffness matrices respectively, is the expected output trajectory. The actual system state is X, and the state error between the two is defined as: -X； 6.

2. Construction and Stability Analysis of Lyapunov Functions; Construct the following Lyapunov function to analyze the stability of the system and derive the parameter update law: ； in : is the impedance stiffness estimation error; >0: adjust the gain, affecting the parameter convergence speed; : is the expected system mass matrix; Taking the time derivative of V and combining it with the system dynamics model can yield a stable adaptive rate. 6.

3. Adaptive rate design; According to the actual external force detected and pose error , design the following adaptive parameter adjustment rate; Stiffness adaptation law: ； Damping adaptation rate: ； in , >0: for adjusting the gain parameter; f() velocity error response function; All regulation rates have physical constraints [ ], to prevent excessive stiffness or oscillation.

5. The massage robot compliance control method with stiffness adaptive function according to claim 1, characterized in that: The step 7 is specifically as follows: 7.1 Voice collection and feature extraction; The user inputs natural language commands through the microphone, the system collects the speech data, and extracts the Mel-spectrogram, MFCC, or other high-dimensional feature vectors as model input; 7.2 Command recognition and intent analysis; The system uses a trained speech recognition model to map speech information into a standard semantic structure, and further converts it into predefined action labels through the intent recognition module; 7.3 Control instruction generation and integration; The aforementioned voice commands are converted into control commands, encapsulated as standard control functions, and embedded into the current flexible control process. Specifically, the controller receives the control function command, updates the function parameters, passes the received command to the actuator, and the actuator executes it and receives feedback until the user says "OK" or stops to end the process. 7.4 Control process integration; The voice-converted control instructions are encapsulated and embedded in the compliant control process, working in conjunction with the terminal force feedback, stiffness detection, and adaptive adjustment modules to ensure that the system behavior remains dynamically stable and interactively compliant under the influence of language input.

Images