CN-122008272-A - Object attribute sensing method, device and storage medium based on smart hand touch

CN122008272ACN 122008272 ACN122008272 ACN 122008272ACN-122008272-A

Abstract

The application provides an object attribute sensing method, device and storage medium based on smart hand touch, and belongs to the technical field of touch sensing. The method comprises the steps of controlling a smart hand provided with a three-dimensional force sensor array to conduct initial tact scanning on a target object to obtain initial contour information of the target object, determining a key area on the surface of the target object based on the initial contour information, wherein the key area comprises a high curvature area and/or a flat area, determining a corresponding perception strategy according to the type of the key area, and controlling the smart hand to execute the corresponding perception strategy in the key area to obtain a physical attribute perception result. The application can adapt to the characteristics of the object and perform cooperative sensing on multiple attributes of the object.

Inventors

CHEN LIYANG
YU YAYUN

Assignees

悟通感控(北京)科技有限公司
悟通感控(山东)科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260414

Claims (10)

1. A smart hand haptic based object property perception method, the method comprising: controlling a dexterous hand provided with a three-dimensional force sensor array to perform initial tact scanning on a target object, and acquiring initial contour information of the target object; determining a key area of the surface of the target object based on the preliminary profile information, wherein the key area comprises a high curvature area and/or a flat area; determining a corresponding perception strategy according to the type of the key region; And controlling the smart hand to execute a corresponding sensing strategy in the key area to acquire a physical attribute sensing result.
2. The smart hand haptic based object property sensing method of claim 1, wherein said determining a corresponding sensing strategy based on the type of the key region comprises: when the type of the key area comprises a high curvature area, determining that a corresponding sensing strategy is multi-angle touch, and performing shape-following touch on the target object along an angle with the largest curvature so as to refine contour sensing; and when the type of the key area comprises a flat area, determining a corresponding perception strategy to apply a fixed normal force to the target object to press so as to carry out hardness perception and/or smoothness perception.
3. The smart hand-tactile object attribute sensing method of claim 2 wherein the critical area comprises a high curvature area and the physical attribute sensing result comprises a contour model, the controlling the smart hand to execute a corresponding sensing strategy in the critical area to obtain a physical attribute sensing result comprises: controlling the smart hand to perform multi-angle touch in the high-curvature area, and acquiring real-time sensing data acquired by the three-dimensional force sensor array; determining a real-time stressed area, a real-time stressed direction and a real-time stressed size based on the real-time sensing data, taking the real-time stressed area, the real-time stressed direction and the real-time stressed size as input of a preset neural network model, and acquiring real-time curvature output by the preset neural network model; Determining an angle with the largest real-time curvature as a target angle in the multi-angle touch process, and controlling the smart hand to carry out random touch on the target object based on the target angle; And when the global outline touch is detected to be completed, constructing an outline model based on the movement track of the smart hand.
4. The smart hand-tactile object attribute sensing method of claim 2 wherein the critical area comprises a flat area, the physical attribute sensing result comprises a hardness level, the controlling the smart hand to execute a corresponding sensing strategy in the critical area, obtaining a physical attribute sensing result comprises: Controlling the smart hand to apply a first normal force to press the smart hand against the flat area; Determining a first point number of a sensing unit for sensing a force signal in the three-dimensional force sensor array when the force signal sensed by the three-dimensional force sensor array is stable; calculating to obtain a hardness sensing stressed area based on the single stressed area of the sensing unit and the first point number; And determining the hardness grade based on the hardness sensing stressed area and a preset hardness calibration function.
5. The smart hand tactile object property sensing method of claim 4 wherein said critical area comprises a flat area and said physical property sensing result further comprises a smoothness coefficient, and further comprising, after said determining a hardness level based on said hardness sensing footprint and a preset hardness calibration function: Controlling the smart hand to slide in the flat area according to a preset direction and a preset speed; Determining a total tangential force sensed in the three-dimensional force sensor array when the force signal sensed by the three-dimensional force sensor array is stable; a smoothness coefficient is determined based on the first normal force, the hardness level, and the total tangential force.
6. The smart hand-tactile object attribute sensing method of claim 2 wherein the critical area comprises a flat area, the physical attribute sensing result comprises a smoothness coefficient, the controlling the smart hand to execute a corresponding sensing strategy in the critical area, obtaining a physical attribute sensing result comprises: Controlling the smart hand to apply a second normal force to press the smart hand on the flat area, and sliding the smart hand on the flat area according to a preset direction and a preset speed; Determining a second point number of sensing units sensing force signals in the three-dimensional force sensor array and a tangential resultant force sensed when the force signals sensed by the three-dimensional force sensor array are stable; Calculating to obtain a smoothness sensing stressed area based on the single stressed area of the sensing unit and the second point number; and determining a smoothness coefficient based on the second normal force, the tangential resultant force, and the smoothness-aware surface area.
7. The smart hand haptic based object property sensing method of claim 6, wherein said determining a smoothness coefficient based on said second normal force, said tangential resultant force, and said smoothness sensing surface area comprises: When the smoothness sensing stressed area is smaller than or equal to a preset area threshold value, adopting a rigid body smoothness calculation strategy, and determining a smoothness coefficient based on the second normal force and the tangential resultant force; and when the smoothness sensing stressed area is larger than a preset area threshold, determining a smoothness coefficient by adopting a non-rigid body smoothness calculation strategy based on the second normal force, the tangential resultant force and a preset calibration formula.
8. The smart hand tactile object property sensing method of any one of claims 1-7 wherein the physical property sensing result comprises an elastic modulus, the method further comprising: Controlling a first fingertip and a second fingertip of the dexterous hand to execute tightening action, wherein when any finger is detected to be in contact with the target object through a three-dimensional force sensor array of the first fingertip and the second fingertip, stopping executing tightening action of the contact finger until the first fingertip and the second fingertip are in contact with the target object, and recording a first interval between the first fingertip and the second fingertip; Controlling the first fingertip and the second fingertip to clamp the target object with a third normal force; When the third normal force sensed by the three-dimensional force sensor arrays of the first fingertip and the second fingertip is stable, determining a third point number of the sensing units sensing the third normal force in the three-dimensional force sensor arrays and a second interval between the first fingertip and the second fingertip; based on the single stress area of the sensing unit and the third point number, calculating to obtain an elastic sensing stress area; And calculating to obtain elastic modulus according to the third normal force, the elastic sensing stressed area, the first interval and the second interval.
9. An electronic device, comprising: one or more processors; A memory for storing one or more programs, Wherein the one or more programs, when executed by the one or more processors, cause the one or more processors to perform the smart-hand haptic based object property perception method of any of claims 1 to 8.
10. A computer storage medium storing executable instructions which when executed by a processor cause the processor to perform the smart hand haptic based object property perception method of any one of claims 1 to 8.

Description

Object attribute sensing method, device and storage medium based on smart hand touch Technical Field The application relates to the technical field of touch perception, in particular to an object attribute perception method, device and storage medium based on smart hand touch. Background With the deep application of the intelligent robot technology in the fields of industrial precise assembly and dangerous environment operation, especially minimally invasive surgery robots, the robots often need to operate target objects in unfamiliar environments without vision or limited vision. In such a scenario, haptic perception is a core alternative to robots understanding the key properties of objects, enabling precise smart operations. Currently, existing haptic sensations typically rely on a pre-set fixed program to sequentially measure a single or multiple properties of an object. For example, one possible implementation is to first control the tactile sensor to make a traversing touch to the object surface according to a preset trajectory to reconstruct its outline, then perform a fixed pressing or sliding action on a specific point to measure hardness or roughness, etc. This sequential measurement approach, while capable of obtaining some of the object's properties to some extent, results in poor perception efficiency and lack of overall cooperativity due to the dispersion of its perception process. Therefore, a new method capable of adapting to the characteristics of the object and cooperatively sensing multiple attributes of the object is needed. Disclosure of Invention The object of the present application is to provide a smart hand touch-based object attribute sensing method, apparatus and storage medium to solve the above problems. To achieve the above object, in a first aspect, the present application provides an object attribute sensing method based on smart hand touch, the method comprising: controlling a dexterous hand provided with a three-dimensional force sensor array to perform initial tact scanning on a target object, and acquiring initial contour information of the target object; determining a key area of the surface of the target object based on the preliminary profile information, wherein the key area comprises a high curvature area and/or a flat area; determining a corresponding perception strategy according to the type of the key region; And controlling the smart hand to execute a corresponding sensing strategy in the key area to acquire a physical attribute sensing result. In some embodiments, the determining a corresponding perception policy according to the type of the key region includes: when the type of the key area comprises a high curvature area, determining that a corresponding sensing strategy is multi-angle touch, and performing shape-following touch on the target object along an angle with the largest curvature so as to refine contour sensing; and when the type of the key area comprises a flat area, determining a corresponding perception strategy to apply a fixed normal force to the target object to press so as to carry out hardness perception and/or smoothness perception. In some embodiments, the key region includes a high curvature region, the physical attribute sensing result includes an outline model, and the controlling the smart hand to execute a corresponding sensing policy in the key region, to obtain the physical attribute sensing result includes: controlling the smart hand to perform multi-angle touch in the high-curvature area, and acquiring real-time sensing data acquired by the three-dimensional force sensor array; determining a real-time stressed area, a real-time stressed direction and a real-time stressed size based on the real-time sensing data, taking the real-time stressed area, the real-time stressed direction and the real-time stressed size as input of a preset neural network model, and acquiring real-time curvature output by the preset neural network model; Determining an angle with the largest real-time curvature as a target angle in the multi-angle touch process, and controlling the smart hand to carry out random touch on the target object based on the target angle; And when the global outline touch is detected to be completed, constructing an outline model based on the movement track of the smart hand. In some embodiments, the key region includes a flat region, the physical attribute sensing result includes a hardness level, and the controlling the smart hand to execute a corresponding sensing policy in the key region, to obtain the physical attribute sensing result includes: Controlling the smart hand to apply a first normal force to press the smart hand against the flat area; Determining a first point number of a sensing unit for sensing a force signal in the three-dimensional force sensor array when the force signal sensed by the three-dimensional force sensor array is stable; calculating to obtain a hardness sensing stressed area based on the single st