Grabbing target hardness recognition and holding posture sample collection system and method

Abstract

The invention discloses a grabbed target hardness recognition and holding posture sample collection system. The system comprises an array type touch sensor, an inertial measurement unit, an algorithm controller and a pedal control module. The invention further discloses a grabbing target hardness recognition and holding posture sample collection method. The method comprises the steps that the array type touch sensor and the inertial measurement unit are fixed to the finger pulp and fingernail positions of the tail end of a finger; the algorithm controller and the pedal control module are fixed to the wrist position; the pedal control module is used for controlling the array type touch sensor and the inertial measurement unit to calibrate; the array type touch sensor collects the interaction force of the tail ends of the fingers in the grabbing process; the inertial measurement unit is used for collecting finger tail end pose data; and the algorithm controller records the interaction force and the posture of the finger tail end to obtain the hardness information of the grabbing target. The method does not involve model training, is small in calculation amount and small in data volume, can obtain accurate collected data, is convenient for transmission and later use, and has important guiding significance in the technical field of information collection.

Description

Technical Field

[0001] This invention relates to a system and method for identifying the hardness of a grasping target and collecting grip posture samples, belonging to the field of information acquisition technology. Background Technology

[0002] As a multi-degree-of-freedom end effector, the dexterous hand achieves autonomous grasping through imitation learning of grip posture samples, a convenient and common method. Simultaneously, real-time acquisition and transmission of grip posture data enables high-precision master-slave teleoperation and dexterous task teaching control. For the demands of dexterous and precise tasks in complex and variable scenarios, in addition to acquiring grip posture samples, identifying the hardness attribute of the grasped target helps adjust the expected grasping force, thereby significantly improving the grasping success rate of the dexterous hand. In recent years, digital sensor technology has developed rapidly and become increasingly advanced. Currently, the most common methods for hardness identification of targets include tactile sensing and visual image processing for hardness recognition. Furthermore, there are increasingly more methods available for acquiring grasping posture, with the most commonly used methods including fiber optic flexible sensors and inertial sensors.

[0003] Existing object attribute recognition methods based on dexterous hand tactile perception mainly suffer from the following problems:

[0004] (1) It is necessary to collect a large amount of feedback data from different types of objects and use them as learning samples to train a convolutional neural network model with multi-dimensional features in parallel. Although this method has good generalization, the accuracy of the model training in the early stage is highly dependent on the types of objects. When the types are not rich enough, the model error is large.

[0005] (2) The more objects in the learning sample, the more accurate the model training result, but the slower the model training speed. When the amount of sample data is too large or too small, the model training will not converge, resulting in the failure of model training.

[0006] Existing methods for real-time measurement of the hardness of multiple objects using visual-tactile sensors mainly suffer from the following problems:

[0007] (1) Visual-tactile sensing requires equipment such as cameras and light sources, which are large in size and not conducive to portable integration;

[0008] (2) The hardness model needs to be pre-trained. The accuracy of the model depends on the sample data of the pre-training, and there is a risk that the model training will not converge.

[0009] (3) Visual-touch sensing is easily affected by external light sources. When the contact surface is blocked during operation, it is impossible to obtain a relatively accurate result. Summary of the Invention

[0010] The purpose of this invention is to overcome the aforementioned shortcomings and provide a system and method for identifying the hardness of grasping targets and collecting grip posture samples, thus solving the technical problem of difficulty in identifying the hardness and grip posture of grasping targets. This invention does not involve model training, has low computational load and small data volume, and can obtain accurate collected data, facilitating transmission and subsequent use. It has significant guiding significance in the field of information acquisition technology.

[0011] To achieve the above-mentioned objectives, the present invention provides the following technical solution:

[0012] A system for identifying the hardness of a grasping target and collecting grip posture samples includes: an array of tactile sensors, an inertial measurement unit, an algorithm controller, and a pedal control module;

[0013] An array of tactile sensors is fixed at the fingertips to collect the interaction force at the fingertips during grasping and output the interaction force to the algorithm controller.

[0014] An inertial measurement unit is fixed at the fingernail position at the fingertip to collect fingertip pose data and output the fingertip pose data to the algorithm controller;

[0015] The algorithm controller is fixed at the wrist position and is used to record the interaction force of the fingertips transmitted by the array of tactile sensors and the fingertips pose data transmitted by the inertial measurement unit. The hardness information of the grasping target is obtained based on the interaction force of the fingertips.

[0016] The pedal control module is fixed at the wrist position and is used to control the array-type tactile sensor and inertial measurement unit to perform calibration before data acquisition, and to control the algorithm controller to stop data processing when data acquisition is abnormal.

[0017] Furthermore, the array-type tactile sensor and inertial measurement unit are fixed to the fingertip and nail position at the end of the finger through a fixed structure;

[0018] The fixed structure includes an inertial measurement unit fixed position located above and an array-type tactile sensor fixed position located below. The array-type tactile sensor and the inertial measurement unit are fixed to the array-type tactile sensor fixed position and the inertial measurement unit fixed position respectively by screws or adhesive.

[0019] The fixation structure is secured to the fingertip with a strap.

[0020] Furthermore, the array-type tactile sensor includes no fewer than 5 × 10 tactile units;

[0021] The interactive force output by the array-type tactile sensor includes the contact force of each tactile unit in the x, y, and z directions; the contact force of each tactile unit is synthesized by the algorithm controller into the resultant force of the entire array-type tactile sensor.

[0022] Furthermore, fingers and wrists include the fingers and wrists of the human hand or dexterous hand.

[0023] A method for identifying the hardness of a grasping target and acquiring grip posture samples, implemented using the aforementioned system for identifying the hardness of a grasping target and acquiring grip posture samples, includes:

[0024] S1 fixes the array-type tactile sensor and the inertial measurement unit to the fingertip and nail position at the fingertips, respectively; and fixes the algorithm controller and pedal control module to the wrist position.

[0025] S2 uses a pedal control module to control an array of tactile sensors and an inertial measurement unit for calibration;

[0026] The S3 array-type tactile sensor collects the interaction force of the fingertips during the grasping process and outputs the interaction force to the algorithm controller; the inertial measurement unit collects the fingertips pose data and outputs the fingertips pose data to the algorithm controller; the algorithm controller records the interaction force of the fingertips transmitted by the array-type tactile sensor and the fingertips pose data transmitted by the inertial measurement unit, and obtains the hardness information of the grasping target based on the interaction force of the fingertips.

[0027] When data acquisition is abnormal, the pedal control module controls the algorithm controller to stop data processing.

[0028] Furthermore, step S2, which uses the pedal control module to control the array of tactile sensors and the inertial measurement unit for calibration, includes the following methods:

[0029] When the fingertip is not in contact with the grasping target, the pedal control module is triggered to enter the array tactile sensor calibration stage, so that the contact force of each tactile unit in the array tactile sensor and the resultant force of the entire array tactile sensor are both 0.

[0030] With the fingers extended, the pedal control module is triggered to enter the inertial measurement unit calibration phase, obtaining the initial pose matrix of the inertial measurement unit relative to the algorithm controller.

[0031] Furthermore, in the inertial measurement unit (IMU) calibration process, the methods for obtaining the initial pose matrix of the IMU relative to the algorithm controller include:

[0032] The coordinate system of the inertial measurement unit at the fingertip is defined as the b system, the coordinate system of the inertial measurement unit on the algorithm controller at the wrist is defined as the p system, and the geodetic coordinate system is defined as the t system.

[0033] During the inertial measurement unit (IMU) calibration process, after a period of stillness of at least 3 seconds, the number of sampling points is recorded as N. This yields the attitude information of each of the n sampling points from two IMUs: the fingertip IMU (b-system) and the IMU on the algorithm controller at the wrist (p-system). and in, These are the pitch angle, roll angle, and yaw angle of the fingertip inertial measurement unit, respectively. These are the pitch angle, roll angle, and yaw angle of the inertial measurement unit on the algorithm controller at the wrist, respectively, n = 1, 2, ... N (sampling points);

[0034] Calculate the average value of each. and The attitude transition matrix is ​​given, and the attitude transition matrix corresponding to the average attitude is calculated respectively. and Thus, the initial pose matrix of the b-frame relative to the p-frame can be calculated. in, Let be the attitude transition matrix of the b-frame relative to the t-frame. Let be the attitude transfer matrix of the p-frame relative to the t-frame, and let x, y, and z be the position deviations of the fingertip inertial measurement unit relative to the inertial measurement unit on the algorithm controller at the wrist.

[0035] Furthermore, in step S3, the algorithm controller calculates the pose matrix of the fingertip relative to the wrist coordinate system, specifically including:

[0036] Based on the output data of the inertial measurement units at the fingertips and the inertial measurement units on the algorithm controller at the wrist, the attitude transfer matrices of the two inertial measurement units relative to the geodetic coordinate system t are obtained, respectively. and The position information of the inertial measurement unit at the fingertip and the inertial measurement unit on the algorithm controller at the wrist are respectively {x b ,y b ,z b} and {x p ,y p ,z p}, thus obtaining the pose matrix of the fingertips relative to the wrist coordinate system.

[0037] Furthermore, in step S3, the method by which the algorithm controller obtains the gripping target hardness information based on the interaction force at the fingertips includes:

[0038] Calculate the resultant force of the entire array of tactile sensors;

[0039] When |FF t When |≤β, the contact factor T = 0 is defined at each moment; where F t β is the characteristic force, determined based on the required gripping force of the target object made of different materials; β is the custom viewing window size.

[0040] The output force of each tactile unit in the array tactile sensor is examined sequentially at each moment, and the resultant force of each tactile unit at each moment is calculated.

[0041] When the resultant force of each tactile unit at any given moment is greater than or equal to F min When T = T + 1; where F min This is the minimum force value that the tactile unit can be sensitive to;

[0042] The hardness factor at that moment

[0043] Compared with the prior art, the present invention has at least one of the following advantages:

[0044] (1) This invention uses an array of tactile sensors to collect gripping force and identify hardness attributes. By calculating the hardness factor during the gradual increase of gripping force, hardness information is identified. It does not involve model training, has a small amount of computation and data, and is easy to transmit and use later.

[0045] (2) The present invention designs a fingertip fixed structure to fix the inertial sensor and the array tactile sensor. It can be installed at the fingertip or dexterous hand end as needed, without involving the wearing of gloves, with good compatibility. Different hand sizes or dexterous hand sizes do not affect the measurement results, and it is applicable to both left and right hands.

[0046] (3) The present invention directly outputs the pose matrix in Cartesian space, and under heterogeneous conditions, it is not necessary to perform forward kinematics calculation of Cartesian space pose information based on joint angle information.

[0047] (4) The present invention allows for selection of the number of fingers to be worn according to needs, and is compatible with dexterous hands of different configurations such as two fingers, three fingers, and five fingers. Attached Figure Description

[0048] Figure 1 The diagram shows the sensor layout and system composition of this invention; (a) is the back of the left hand when wearing the device, and (b) is the palm of the left hand when wearing the device.

[0049] Figure 2 This is a schematic diagram of the fixed connection structure of the present invention;

[0050] Figure 3 This is a flowchart of a method for identifying the hardness of a grasping target and collecting grip posture samples according to the present invention. Detailed Implementation

[0051] The features and advantages of the present invention will become clearer and more apparent from the following detailed description.

[0052] The term “exemplary” as used herein means “serving as an example, embodiment, or illustration.” Any embodiment illustrated herein as “exemplary” is not necessarily to be construed as superior to or better than other embodiments. Although various aspects of embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless specifically indicated otherwise.

[0053] This invention proposes a system and method for identifying the hardness of a grasping target and collecting grip posture samples. The invention designs a fingertip fixation structure to attach an inertial sensor and an array of tactile sensors to the fingertips of a human hand or dexterous hand where Cartesian spatial pose information needs to be measured. This allows for the acquisition of Cartesian spatial pose information, and the array of tactile sensors is used to collect gripping force and target hardness attribute information. Subsequently, a controller matches, stores, and transmits the gripping pose and force. This system and method are suitable for measuring the Cartesian spatial pose information of human hands or dexterous hands worn on the fingertips. It can be used for imitation learning sample collection in autonomous grasping by multi-finger dexterous hands, or for teaching and controlling bomb disposal robots, surgical robots, and space robots in remote master-slave teleoperation modes.

[0054] Example:

[0055] like Figure 1 As shown, the sensor layout of the target hardness recognition and grip posture sample acquisition system of the present invention is as follows, taking the left hand with 3 fingers as an example, but 1 to 5 fingers can be worn as needed:

[0056] A system for recognizing the hardness of a grasping target and collecting grip posture samples includes at least one fixed structure, one array-type tactile sensor 1, two inertial measurement units 2, one algorithm controller 3, one wireless transmission module 4, and one pedal control module 5. The two inertial measurement units are located on the algorithm controller at the fingernail position and the wrist position, respectively.

[0057] Among them, the fixed structure is as follows Figure 2As shown, the device comprises four parts: an inertial measurement unit fixing position 61, an array-type tactile sensor fixing position 62, a strap fixing position 63, and a finger fixing position 64. At least three through holes are provided on the sensor mounting surface to facilitate sensor installation. These holes are used to fix the array-type tactile sensor and the inertial measurement unit to the fingertips of the person being measured using the straps 64. One array-type tactile sensor and one inertial measurement unit are respectively glued or screwed to the fingertip and nail positions of the fingertip fixing structure. The inertial measurement unit collects fingertip pose data, while the array-type tactile sensor is responsible for sensing the interaction force during grasping and identifying the hardness attribute of the target object through a hardness recognition algorithm. The wireless transmission module, the inertial measurement unit, and the algorithm controller are fixedly connected and secured to the wrist position of the person being measured using the straps. The data acquisition lines of the array-type tactile sensor and the inertial measurement unit are connected to the algorithm controller, as are the data lines of the pedal control module. The algorithm controller is responsible for recording, processing, storing, and outputting data from the inertial measurement unit and the array of tactile sensors. The hardness recognition algorithm is implemented in the algorithm controller, which ultimately transmits information such as fingertip pose data, contact force at each end, and target hardness attributes. The wireless transmission module is responsible for transmitting the results processed by the algorithm controller wirelessly to the host computer or dexterous hand controller.

[0058] An array-type tactile sensor includes no fewer than 5 × 10 tactile units and can output the contact force of no fewer than 50 tactile units. The combined force of the entire array of tactile sensors {F x ,F y ,F z}

[0059] like Figure 3 As shown, a method and steps for identifying the hardness of a grasping target and collecting grip posture samples are as follows:

[0060] Step (1) Secure the fixed structure to the fingertips of the human hand or dexterous hand using straps, ensuring that the array-type tactile sensor is at the fingertip and the inertial measurement unit is at the nail; secure the combination of the algorithm controller, pedal control module, and wireless transmission module to the wrist of the human hand or dexterous hand using straps; measure and record the position {x,y,z} of the inertial measurement unit at each fingertip relative to the inertial measurement unit on the algorithm controller.

[0061] Step (II) Perform initial calibration. With the wearer's fingertips not touching the pedal, press the pedal twice consecutively within 1 second to complete the zero calibration of the array-type tactile sensor. The contact force of no less than 50 tactile units should be measured. The combined force of the entire array of tactile sensors {F x ,F y ,Fz All values ​​are 0; then, with fingers extended, the human hand or dexterous hand presses the pedal three times in quick succession within one second to enter the inertial measurement unit calibration phase. The system remains stationary for at least three seconds to complete the initial calibration of the system.

[0062] The inertial measurement unit (IMU) calibration process will be introduced using an IMU at the tip of any finger and an IMU on the algorithm controller at the wrist as examples:

[0063] The coordinate system of the fingertip inertial measurement unit (IMU) is defined as the b-frame, the coordinate system of the IMU on the wrist algorithm controller is defined as the p-frame, and the geodetic coordinate system is defined as the t-frame. After a period of stillness of at least 3 seconds, the number of sampling points is recorded as N. The attitude information of each of the n sampling points of the two IMUs can be obtained. and Calculate the average value of each. and The attitude transition matrix, where, These are the pitch angle, roll angle, and yaw angle of the fingertip inertial measurement unit, respectively. These are the pitch angle, roll angle, and yaw angle of the inertial measurement unit on the algorithm controller at the wrist, respectively, n = 1, 2, ... N (sampling points);

[0064] Calculate the attitude transition matrix corresponding to the average attitude respectively. and This allows us to calculate the initial pose matrix of the fingertips relative to the wrist coordinate system. in, Let be the attitude transition matrix of the b-frame relative to the t-frame. Let be the attitude transfer matrix of the p-frame relative to the t-frame, and let x, y, and z be the position deviations of the fingertip inertial measurement unit relative to the inertial measurement unit on the algorithm controller at the wrist.

[0065] Step (3) Data Processing: The wearer grasps the target object, during which inertial measurement unit data is collected and recorded to obtain high-precision fingertip pose matrix information, array-type tactile sensor data is collected and recorded, and the hardness attribute information of the target object is identified. If any abnormality occurs during this process, the pedal controller can be quickly pressed to stop the data processing flow.

[0066] The high-precision fingertip pose matrix information obtained from the inertial measurement unit data is as follows:

[0067] Based on the output data of the inertial measurement units at the fingertips and the inertial measurement units on the algorithm controller at the wrist, the attitude transfer matrices of the two inertial measurement units relative to the geodetic coordinate system t-frame can be obtained, respectively. and The location information is {x b ,y b ,z b} and {x p ,y p ,z p This allows for the calculation of the pose matrix of the fingertips relative to the wrist coordinate system.

[0068] The target object hardness identification algorithm is as follows:

[0069] Taking a thumb array tactile sensor as an example, data processing is performed:

[0070] As the gripping force gradually increases, observe the combined force of the tactile sensors on the device. As the gripping force gradually increases, F gradually increases, and at time t, F increases to Ft. t When -F≤β, enter the window sampling time, when FF t When ≤β, the sampling time of the outgoing window, where β is the custom viewing window size, and the feature force F t Defined based on the gripping force achievable by target objects made of different materials, the preferred characteristic force F is... t The minimum force required to grasp any material target object can be selected. Within the window sampling time, the contact factor T = 0 is defined at each time step. At the same sampling time, the output force of each of the 5 × 10 haptic units is examined sequentially. And calculate the resultant force amplitude of a single haptic unit at each moment.

[0071] F i ≥F min When T = T + 1, its

[0072] China F min The minimum force that the sensor's tactile unit can detect, and the hardness factor at that moment. The hardness factor is calculated sequentially at each time point within the sampling window, and the average value is the hardness factor output by the sensor.

[0073] The same data processing is then performed on the other four fingers, and the average value of the effective hardness factors is calculated to obtain the hardness of the target object.

[0074] Step (4) Transmission: The joint angles of each finger, tactile sensor data, and hardness calculation values ​​are transmitted in real time to the wireless communication module, and then to the host computer or dexterous hand controller.

[0075] The present invention has been described in detail above with reference to specific embodiments and exemplary examples; however, these descriptions should not be construed as limiting the present invention. Those skilled in the art will understand that various equivalent substitutions, modifications, or improvements can be made to the technical solutions and embodiments of the present invention without departing from the spirit and scope of the invention, and all such modifications and improvements fall within the scope of the present invention. The scope of protection of the present invention is defined by the appended claims.

[0076] The contents not described in detail in this specification are common knowledge to those skilled in the art.

Claims

1. A system for identifying the hardness of a grasping target and collecting grip posture samples, characterized in that, include: Array-type tactile sensors, inertial measurement unit, algorithm controller, and pedal control module; An array of tactile sensors is fixed at the fingertips to collect the interaction force at the fingertips during grasping and output the interaction force to the algorithm controller. An inertial measurement unit is fixed at the fingernail position at the fingertip to collect fingertip pose data and output the fingertip pose data to the algorithm controller; The algorithm controller is fixed at the wrist position and is used to record the interaction force of the fingertips transmitted by the array of tactile sensors and the fingertips pose data transmitted by the inertial measurement unit. The hardness information of the grasping target is obtained based on the interaction force of the fingertips. The pedal control module is fixed at the wrist position and is used to control the array-type tactile sensor and inertial measurement unit to perform calibration before data acquisition, and to control the algorithm controller to stop data processing when data acquisition is abnormal.

2. The system for identifying the hardness of a grasping target and collecting grip posture samples according to claim 1, characterized in that, The array-type tactile sensor and inertial measurement unit are fixed to the fingertip and nail position at the end of the finger through a fixed structure; The fixed structure includes an inertial measurement unit fixed position located above and an array-type tactile sensor fixed position located below. The array-type tactile sensor and the inertial measurement unit are fixed to the array-type tactile sensor fixed position and the inertial measurement unit fixed position respectively by screws or adhesive. The fixation structure is secured to the fingertip with a strap.

3. The system for identifying the hardness of a grasping target and collecting grip posture samples according to claim 1, characterized in that, An array-type tactile sensor includes no fewer than 5 × 10 tactile units; The interactive force output by the array-type tactile sensor includes the contact force of each tactile unit; the contact force of each tactile unit is synthesized by the algorithm controller into the resultant force of the entire array-type tactile sensor.

4. The system for identifying the hardness of a grasping target and collecting grip posture samples according to claim 1, characterized in that, Fingers and wrists include the fingers and wrists of a human hand or dexterous hand.

5. A method for identifying the hardness of a grasping target and collecting grip posture samples, characterized in that, The system for identifying the hardness of a grasping target and collecting grip posture samples, as described in any one of claims 1-4, includes: S1 fixes the array-type tactile sensor and the inertial measurement unit to the fingertip and nail position at the fingertips, respectively; and fixes the algorithm controller and pedal control module to the wrist position. S2 uses a pedal control module to control an array of tactile sensors and an inertial measurement unit for calibration; The S3 array-type tactile sensor collects the interaction force of the fingertips during the grasping process and outputs the interaction force to the algorithm controller; the inertial measurement unit collects the fingertips pose data and outputs the fingertips pose data to the algorithm controller; the algorithm controller records the interaction force of the fingertips transmitted by the array-type tactile sensor and the fingertips pose data transmitted by the inertial measurement unit, and obtains the hardness information of the grasping target based on the interaction force of the fingertips. When data acquisition is abnormal, the pedal control module controls the algorithm controller to stop data processing.

6. The method for identifying the hardness of a grasping target and collecting grip posture samples according to claim 5, characterized in that, Step S2, the method for calibrating the array-type tactile sensor and inertial measurement unit using the pedal control module, includes: When the fingertip is not in contact with the grasping target, the pedal control module is triggered to enter the array tactile sensor calibration stage, so that the contact force of each tactile unit in the array tactile sensor and the resultant force of the entire array tactile sensor are both 0. With the fingers extended, the pedal control module is triggered to enter the inertial measurement unit calibration phase, obtaining the initial pose matrix of the inertial measurement unit relative to the algorithm controller.

7. The method for identifying the hardness of a grasping target and collecting grip posture samples according to claim 6, characterized in that, In the inertial measurement unit (IMU) calibration process, methods for obtaining the initial pose matrix of the IMU relative to the algorithm controller include: The coordinate system of the inertial measurement unit at the fingertip is defined as the b system, the coordinate system of the inertial measurement unit on the algorithm controller at the wrist is defined as the p system, and the geodetic coordinate system is defined as the t system. During the inertial measurement unit (IMU) calibration process, after a period of stillness of at least 3 seconds, the number of sampling points is recorded as N. This yields the attitude information of each of the n sampling points from two IMUs: the fingertip IMU and the IMU on the algorithm controller at the wrist. and in, These are the pitch angle, roll angle, and yaw angle of the fingertip inertial measurement unit, respectively. These are the pitch angle, roll angle, and yaw angle of the inertial measurement unit on the algorithm controller at the wrist, respectively, n = 1, 2, ... N; Calculate the average value of each. and The attitude transition matrix is ​​given, and the attitude transition matrix corresponding to the average attitude is calculated respectively. and Thus, the initial pose matrix of the b-frame relative to the p-frame can be calculated. in, Let be the attitude transition matrix of the b-frame relative to the t-frame. Let be the attitude transfer matrix of the p-frame relative to the t-frame, and let x, y, and z be the position deviations of the fingertip inertial measurement unit relative to the inertial measurement unit on the algorithm controller at the wrist.

8. The method for identifying the hardness of a grasping target and collecting grip posture samples according to claim 6, characterized in that, In step S3, the algorithm controller calculates the pose matrix of the fingertip relative to the wrist coordinate system, specifically including: Based on the output data of the inertial measurement units at the fingertips and the inertial measurement units on the algorithm controller at the wrist, the attitude transfer matrices of the two inertial measurement units relative to the geodetic coordinate system t are obtained, respectively. and The position information of the inertial measurement unit at the fingertip and the inertial measurement unit on the algorithm controller at the wrist are respectively {x b ,y b ,z b } and {x p ,y p ,z p }, thus obtaining the pose matrix of the fingertips relative to the wrist coordinate system.

9. The method for identifying the hardness of a grasping target and collecting grip posture samples according to claim 6, characterized in that, In step S3, the method by which the algorithm controller obtains the hardness information of the grasping target based on the interaction force at the fingertips includes: Calculate the resultant force of the entire array of tactile sensors; When |FF t When |≤β, the contact factor T = 0 is defined at each moment; where F t β is the characteristic force, determined based on the required gripping force of the target object made of different materials; β is the custom viewing window size. The output force of each tactile unit in the array tactile sensor is examined sequentially at each moment, and the resultant force of each tactile unit at each moment is calculated. When the resultant force of each tactile unit at any given moment is greater than or equal to F min When T = T + 1; where F min This is the minimum force value that the tactile unit can be sensitive to; The hardness factor at that moment

Images