Manipulator contact force monitoring system based on Internet of Things
Through the Internet of Things-based robot contact force monitoring system, combined with sensors and intelligent algorithms, the contact force is monitored in real time and the grasping force is automatically adjusted, which solves the problem of adjusting the grasping force in the fields of precision operation and medical care, and improves the grasping force Stability and adaptability reduce costs.
Patent Information
- Application Number
- CN202411387843.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-07
- Publication Date
- 2024-11-12
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
In the fields of precision operation and medical, existing robots are difficult to monitor the micro-force state of the contact area in real time, resulting in the inability to accurately adjust the grasping force, affecting the stability and safety of the grasping. The traditional method has poor generalization ability and cannot be met. The demand for fragmented scenarios.
Using a robotic contact force monitoring system based on the Internet of Things, combined with sensors and intelligent algorithms, the distribution characteristics of contact force are monitored in real time through the local IoT platform and edge gateway, and the gripping force is automatically adjusted. Drive motors and torque sensors of multiple joints are included. and structural anti-slip pads, using the system force balance and moment balance equations to obtain contact pressure and tangential force, combined with artificial intelligence algorithms and large models, dynamically adjust the gripping force.
It realizes adaptive adjustment of the gripping force, improves the stability and adaptability of the gripping, and reduces damage to objects. It is especially suitable for precision manufacturing and medical scenarios, reduces costs and does not require additional tactile sensors.
Smart Images

Figure CN118927256A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to an intelligent monitoring device, in particular to a manipulator contact force monitoring system based on the Internet of Things. Background Art
[0002] In the precision operation and complex interaction scenarios of the robot, the monitoring of contact force is crucial, which directly affects the firmness of the grasping and the safety of the manipulated object. The contact area can usually be divided into a slip zone and an adhesion zone. The slip zone is usually at the edge of the contact, while the adhesion zone is close to the center. As the contact pressure increases, the slip zone gradually shrinks and the adhesion zone increases. By accurately monitoring the evolution of the contact force, especially the dynamic changes of the slip zone and the adhesion zone, it is ensured that the robot can automatically adjust the grasping force and reduce the damage to the object during the grasping process.
[0003] At present, there are also some technologies related to manipulator gripping force monitoring on the market. For example, Chinese patent CN201310640086.X discloses a medical intelligent clamping manipulator. Through the gripping finger mechanism and displacement sensor, it can detect the size of the object and realize automatic clamping, effectively avoiding damage to the medicine bottle caused by excessive gripping force and the medicine bottle falling off caused by insufficient gripping force. The manipulator in the project cannot monitor the microscopic force state of the contact area in real time. Since the glass medicine bottle is a rigid material, there is a large range between the minimum gripping force Fmin and the maximum gripping force Fmax of the manipulator. Therefore, when formulating the gripping force in the project, the technicians will choose a larger safety factor to give priority to ensuring that the medicine bottle is in a stable state. However, this method cannot meet the needs of the medical field, and its human-computer interaction characteristics are poor. For example, the amplitude of the gripping force has a great impact on the patient's experience. When a larger gripping force is used, the patient's skin surface will have problems such as redness, swelling and congestion; using a smaller gripping force will cause sticky slippage, causing unstable gripping problems.
[0004] In order to obtain the optimal grasping force, traditional research methods generally use trial and error to determine the amplitude of the grasping force through experiments. However, this method has very poor generalization ability and cannot meet the grasping needs of fragmented scenarios. It also requires a lot of manpower and material resources. In order to solve the above problems, we have developed a manipulator contact force monitoring system based on the Internet of Things. Combining sensing technology, intelligent algorithms and local Internet of Things platforms, it monitors the distribution characteristics of contact force in real time, effectively distinguishes the changes in the slip zone and the adhesion zone, and effectively improves the grasping quality and human-machine compatibility. Summary of the invention
[0005] In view of the above problems, the present invention provides a manipulator contact force monitoring system based on the Internet of Things, which is aimed at the fields of precision grasping and medical monitoring. It obtains the microscopic force state of the contact surface through sensors and mechanical analysis, automatically adjusts the grasping force, and reduces the damage to the object during the grasping process. It is characterized by comprising: a manipulator, an edge gateway, a local Internet of Things platform and a local large model:
[0006] The manipulator comprises a plurality of joints, which are ,......, , each joint includes a drive motor, a sensor module and a structural anti-skid pad; the drive motor can control the state of the joint; the sensor module includes a torque sensor and an encoder, the torque sensor can dynamically measure the joint torque, and the encoder can obtain the position information of the joint; the structural anti-skid pad is made of Ecoflex material with an elastic modulus of 0.02-0.2Mpa, which has excellent deformation ability and biocompatibility and effectively reduces the peak value of the contact force; the contact force includes contact pressure and tangential force, which are obtained according to the system force balance and torque balance equations, specifically:
[0007]
[0008]
[0009] in For the The contact pressure at each contact point, For the Tangential force at each contact point; For the The spatial position vector of each joint; For the The torque of each joint;
[0010] The shear stress amplitude in the contact area is:
[0011] S1: When When the manipulator's clamping force is too small, the contact surface slides as a whole, and the shear stress distribution is:
[0012] < <
[0013] S2: When < When , the manipulator has a stable grasping state, and its shear stress is not enough to cause the object to slide as a whole. The total contact surface includes an adhesion area and sliding areas on both sides at a microscopic perspective; when When , the shear stress distribution in the slip zone is:
[0014]
[0015] when < < When , the shear stress distribution in the adhesion area is:
[0016]
[0017] in is the half width of the contact area, is the half width of the adhesion zone, is the friction coefficient; is the peak contact pressure in the contact area, specifically:
[0018]
[0019] is a correction coefficient that characterizes the influence of the mechanical properties of the material on the contact pressure peak; when the manipulator is in a stable grasping state, the shear stress in the adhesion area , Is the safety factor reflecting the firmness of the grip, where:
[0020]
[0021] The edge gateway includes an industrial computer and a communication module; the industrial computer has a 485 interface and can receive data transmitted by the torque sensor and encoder; the communication module uses 4G to send data to the local Internet of Things platform; the industrial computer can control the drive motor, adjust the amplitude of the contact pressure and shear stress, and ensure the stability of the robot gripping state;
[0022] The local IoT platform interacts bidirectionally with the edge gateway via the MQTT protocol; the local IoT platform can analyze, aggregate and permanently store the received data; the local IoT platform can issue instructions to control the working state of the drive motor through the edge gateway and dynamically adjust the grip amplitude; the IoT platform is integrated with a data visualization module that can display the torque amplitude, position information and force state of each joint of the manipulator;
[0023] The local large model can perform two-way data exchange with the local IoT platform. When the manipulator is working in a steady state, a gripping force increment is applied by driving the motor, and then the joint torque and encoder position data are fitted using an artificial intelligence algorithm to obtain the modulus and hardness information of the material. Finally, the type of material is inferred through the local large model, and a correction coefficient is formulated. and safety factor ; The coefficients are then passed to the local IoT platform and sent to the edge gateway through OTA upgrade, ultimately controlling the drive motor to achieve flexible grasping.
[0024] The advantages of the present invention are:
[0025] 1. Strong adaptability: The system can adaptively adjust the grasping force, use the large model to obtain the material properties of the object to be grasped, and then obtain the grasping force through the algorithm. The drive motor is controlled through the IoT platform and edge gateway to achieve flexible grasping in various scenarios.
[0026] 2. High grasping stability: Considering the microscopic force state of grasping, the minimum grasping force is adopted to ensure the stability and safety of the grasping process, and to reduce the damage to the object during the grasping process, especially suitable for the flexible grasping requirements of precision manufacturing and medical scenarios;
[0027] 3. Low cost: The system does not require additional tactile sensors. Through low-cost torque sensors, mechanism kinematic analysis, mechanical analysis and numerical simulation, it can obtain the microscopic force state of the contact area and obtain the optimal grasping force. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of the present invention, and do not constitute a limitation of the embodiments of the present invention.
[0029] Figure 1 : Microscopic morphology of the contact area during the robot grasping process.
[0030] Figure 2 : Schematic diagram of the three-joint manipulator structure.
[0031] Figure 3 : Distribution law of shear stress in contact area.
[0032] Figure 4 : Data transmission path of contact force monitoring system.
[0033] Figure 5 : Implementation framework for prediction of material modulus and hardness information.
[0034] Figure 6 :Local Internet of Things platform functional module.
[0035] Figure 7 :Deployment example of large AI model. DETAILED DESCRIPTION
[0036] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.
[0037] Embodiment 1
[0038] Attached Figure 1-2 A contact force monitoring system based on the Internet of Things manipulator is demonstrated. It is aimed at the fields of precision grasping and medical monitoring. It obtains the microscopic force state of the contact surface through sensors and mechanical analysis, automatically adjusts the grasping force, and reduces the damage to the object during the grasping process. It is characterized by including: a manipulator, an edge gateway, a local Internet of Things platform, and a local large model:
[0039] The manipulator comprises a plurality of joints, which are ,......, , each joint includes a drive motor, a sensor module and a structural anti-skid pad; the drive motor can control the state of the joint; the sensor module includes a torque sensor and an encoder, the torque sensor can dynamically measure the joint torque, and the encoder can obtain the position information of the joint; the structural anti-skid pad is made of Ecoflex material with an elastic modulus of 0.02-0.2Mpa, which has excellent deformation ability and biocompatibility and effectively reduces the peak value of the contact force; the contact force includes contact pressure and tangential force, which are obtained according to the system force balance and torque balance equations, specifically:
[0040]
[0041]
[0042] in For the The contact pressure at each contact point, For the Tangential force at each contact point; For the The spatial position vector of each joint; For the The torque of each joint;
[0043] Attached Figure 2 The schematic diagram of the three-shutdown manipulator structure is shown. The project substitutes the joint torque measured by the torque sensor and the joint position information obtained by the encoder into the above force balance and torque balance equations, and finally obtains the contact pressure and tangential force, as shown below:
[0044] The robot includes joints and , the length of the lever arm of each joint is , , ; By adjusting the drive motor, the manipulator can stably grasp objects; the joint torques measured by the torque sensor are: , , ; In steady state, the positions of the joints obtained through the encoder are , , When , the steps to obtain the contact force are:
[0045] S1: Calculation of contact pressure :
[0046] ,Right now , ,
[0047] S2: Establish static equilibrium equation: Under static equilibrium conditions, the system needs to satisfy the following equilibrium equation:
[0048] Horizontal (x-direction) force balance:
[0049] Force balance in the vertical direction (y direction):
[0050] Torque balance:
[0051] Solve the system of equations: , ,
[0052] Implementation Example 2
[0053] Attached Figure 3 The distribution law of shear stress in the contact area is shown. Studies have shown that the amplitude of shear stress in the contact area is closely related to the grasping state, specifically:
[0054] S1: When When the manipulator's clamping force is too small, the contact surface slides as a whole, and the shear stress distribution is:
[0055] < <
[0056] S2: When < When , the manipulator has a stable grasping state, and its shear stress is not enough to cause the object to slide as a whole. The total contact surface includes an adhesion area and sliding areas on both sides at a microscopic perspective; when When , the shear stress distribution in the slip zone is:
[0057]
[0058] when < < When , the shear stress distribution in the adhesion area is:
[0059]
[0060] in is the half width of the contact area, is the half width of the adhesion zone, is the friction coefficient; is the peak contact pressure in the contact area, specifically:
[0061]
[0062] is a correction coefficient that characterizes the influence of the mechanical properties of the material on the contact pressure peak; when the manipulator is in a stable grasping state, the shear stress in the adhesion area , Is the safety factor reflecting the firmness of the grip, where:
[0063]
[0064] Attached Figure 4 The data transmission path of the contact force monitoring system is demonstrated; the edge gateway includes an industrial computer and a communication module; the industrial computer has a 485 interface and can receive data transmitted by the torque sensor and encoder; the communication module uses 4G to send data to the local Internet of Things platform; the industrial computer can control the drive motor, adjust the amplitude of the contact pressure and shear stress, and ensure the stability of the robot gripping state;
[0065] Implementation Example 3
[0066] Attached Figure 5 The implementation framework of material modulus and hardness information prediction is demonstrated. When the manipulator is working in a steady state, a gripping force increment is applied by the drive motor, and then the joint torque and encoder position data are fitted using an artificial intelligence algorithm to obtain the material modulus and hardness information. The specific process is as follows:
[0067] S1: Data collection: Control the drive motor through the edge gateway to obtain joint torque and encoder position data at different grip forces;
[0068] S2: Data preprocessing: normalize the data, remove outliers and erroneous data, and randomly divide the data set into training set, test set and validation set;
[0069] S3: Selection of neural network model parameters: Select a suitable neural network structure and determine the number of input layer neurons, hidden layer neurons, and output layer neurons; in this example, what needs to be predicted is the model and hardness information of the material, so the number of output layer neurons is selected as 2;
[0070] S4: Model training and validation: Select Adam optimizer;
[0071] S5: Model deployment and iterative optimization.
[0072] Implementation Example 4
[0073] Attached Figure 6 The functional modules of the local IoT platform are demonstrated. The IoT platform draws on the advantages of Alibaba Cloud and Tencent Cloud IoT platforms during its development. It is the foundation for building smart factories and deploying artificial intelligence algorithms. The project combines years of engineering implementation experience and can meet the needs of industrial robots, smart factories, aerospace, and electric power energy.
[0074] The local IoT platform interacts bidirectionally with the edge gateway via the MQTT protocol; the local IoT platform can analyze, aggregate and permanently store the received data; the local IoT platform can issue instructions to control the working state of the drive motor through the edge gateway and dynamically adjust the grip amplitude; the IoT platform is integrated with a data visualization module, which can display the torque amplitude, position information and force state of each joint of the manipulator; the force state of the contact point includes: contact pressure and shear stress;
[0075] The left side of the IoT platform is the navigation bar, which includes the following functional modules:
[0076] 1. Home page;
[0077] 2. Equipment management: products, equipment, and groups;
[0078] 3. Operation and maintenance monitoring: OTA upgrade, remote configuration, online debugging, and log service;
[0079] 4. System management: user management, role management, menu list, interface management, and log management.
[0080] The IoT platform is written in go + python, and uses MySQL and TDengine databases, where MySQL is used to store relational data and TDengine is used to store time series data. The IoT platform uses MinIO as an object storage server to store unstructured data such as videos and files.
[0081] Implementation Example 5
[0082] The local large model can exchange data with the local IoT platform in a two-way manner. When the manipulator is working in a steady state, it applies a gripping force increment through the drive motor. Then, the artificial intelligence algorithm is used to fit the joint torque and encoder position data to obtain the modulus and hardness information of the material. Finally, the type of material is inferred through the local large model, and the correction coefficient is formulated. and safety factor ; The coefficients are then passed to the local IoT platform and sent to the edge gateway through OTA upgrade, ultimately controlling the drive motor to achieve flexible grasping.
[0083] Figure 7 The local AI big model fine-tuning, quantization, and deployment examples are shown. In recent years, artificial intelligence technology has achieved rapid development, and AI big models have gradually become one of the core technologies in various industries, with broad application potential in natural language processing, time series data processing, and medical fields. The local big model can predict the type of material based on the modulus and hardness information of the material, and then make the optimal grasping strategy. The deployment process of the AI big model is as follows:
[0084] S1: Dataset construction: Establish a standard material property database, including material types, elastic models and structural stiffness, and establish data labels through manual annotation, specifically: correction coefficient , Safety factor ;
[0085] S2: Model fine-tuning: fine-tune the AI large model based on PEFT and perform quantitative operations;
[0086] S3: Local AI large model deployment: Use NVIDIA RTX 4090 graphics card, install LM Studio software locally to load the model;
[0087] S4: Parameter adjustment and API service of AI big model: adjust the batch size, GPU load and Temperature parameters, open the API, and integrate the AI big model with the local IoT platform. The reference code for the IoT platform to call the AI big model is:
[0088] from openai import OpenAI
[0089] client = OpenAI(base_url="http: / / localhost:7500 / v1", api_key="lm-studio")
[0090] def get_embedding(text, model="model-identifier"):
[0091] text = text.replace("\n", " ")
[0092] return client.embeddings.create(input = [text], model=model).data[0].embedding.
[0093] print(get_embedding("The correction factor λ is 0.97."))
[0094] S5: Data delivery: Correction coefficient and safety factor It is transmitted to the local IoT platform and then sent to the edge gateway through OTA upgrade, which ultimately controls the drive motor to achieve flexible grasping.
[0095] The specific implementation method described above provides a detailed description of the objectives, technical solutions and beneficial effects of the present invention. It should be understood that the above is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. A manipulator contact force monitoring system based on the Internet of Things, aimed at the fields of precision grasping and medical monitoring, obtains the microscopic force state of the contact surface through sensors and mechanical analysis, automatically adjusts the grasping force, and reduces the damage to the object during the grasping process. It is characterized by: Includes: Robot, edge gateway, local IoT platform and local large model: The manipulator comprises a plurality of joints, which are ,......, , each joint includes a driving motor, a sensor module and a structural anti-skid pad; the driving motor can control the state of the joint; The sensing module includes a torque sensor and an encoder. The torque sensor can dynamically measure the joint torque, and the encoder can obtain the position information of the joint. The structural anti-slip pad is made of Ecoflex material with an elastic modulus of 0.02-0.2Mpa, which has excellent deformation ability and biocompatibility and effectively reduces the peak value of the contact force. The contact force includes contact pressure and tangential force, which are obtained according to the system force balance and torque balance equations, specifically: in For the The contact pressure at each contact point, For the Tangential force at each contact point; For the The spatial position vector of each joint; For the The torque of each joint; The shear stress amplitude in the contact area is: S1: When When the manipulator's clamping force is too small, the contact surface slides as a whole, and the shear stress distribution is: < < S2: When < When , the manipulator has a stable grasping state, and its shear stress is not enough to cause the object to slide as a whole. The total contact surface includes an adhesion area and sliding areas on both sides at a microscopic perspective; when When , the shear stress distribution in the slip zone is: when < < When , the shear stress distribution in the adhesion area is: in is the half width of the contact area, is the half width of the adhesion zone, is the friction coefficient; is the peak contact pressure in the contact area, specifically: is a correction coefficient that characterizes the influence of the mechanical properties of the material on the contact pressure peak; when the manipulator is in a stable grasping state, the shear stress in the adhesion area , Is the safety factor reflecting the firmness of the grip, where: The edge gateway includes an industrial computer and a communication module; the industrial computer has a 485 interface and can receive data transmitted by the torque sensor and encoder; the communication module uses 4G to send data to the local Internet of Things platform; the industrial computer can control the drive motor, adjust the amplitude of the contact pressure and shear stress, and ensure the stability of the robot gripping state; The local IoT platform interacts bidirectionally with the edge gateway via the MQTT protocol; the local IoT platform can analyze, aggregate and permanently store the received data; the local IoT platform can issue instructions to control the working state of the drive motor through the edge gateway and dynamically adjust the grip amplitude; the IoT platform is integrated with a data visualization module that can display the torque amplitude, position information and force state of each joint of the manipulator; The local large model can perform two-way data exchange with the local IoT platform. When the manipulator is working in a steady state, a gripping force increment is applied by driving the motor, and then the joint torque and encoder position data are fitted using an artificial intelligence algorithm to obtain the modulus and hardness information of the material. Finally, the type of material is inferred through the local large model, and a correction coefficient is formulated. and safety factor ; The coefficients are then passed to the local IoT platform and sent to the edge gateway through OTA upgrade, ultimately controlling the drive motor to achieve flexible grasping.
Citation Information
Cited By
Logistics robot control method and system based on artificial intelligence
CN120244964A
An artificial intelligence-based logistics robot control method and system
CN120244964B
