Pattern recognition method and system based on airbag array tactile sensing

A tactile perception and pattern recognition technology, applied in the field of pattern recognition methods and systems based on airbag array tactile perception, can solve the problems of over-fitting and low recognition accuracy, and achieve good robustness, high accuracy, and improved The effect of the model prediction effect

Pending Publication Date: 2020-08-07
CHONGQING UNIV
View PDF4 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The tactile information collection of the traditional robot skin pattern recognition method usually relies on the contact of objects in a limited pose and position. The tactile perception pattern recognition method only refers to a single machine learning algorithm to train the network, the recognition accuracy is low, and it is prone to overfitting And other issues

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Pattern recognition method and system based on airbag array tactile sensing
  • Pattern recognition method and system based on airbag array tactile sensing

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0041] Such as figure 1 As shown, the airbag array type tactile sensing pattern recognition method provided by this embodiment is aimed at the tactile sensing process of the airbag array robot skin sensor. The method trains the collected tactile data to obtain a pattern recognition model and uses the training method Realize the identification of new data.

[0042] The method provided in this embodiment uses a flexible robot skin sensor based on an airbag array, and the airbag array collects air pressure information when it contacts an object, and trains a machine learning fusion algorithm model to recognize the shape type and object type of the contacting robot skin; from the object contact angle, the airbag The haptic data extracted under the dynamic environment where the array contact position changes are trained, so that the trained model has better robustness. The use of geometric characteristics of different shapes of objects and different physical objects (including human-c...

Embodiment 2

[0051] The new machine learning algorithm fusion pattern recognition method based on the airbag array touch sensing skin sensor provided in this embodiment requires data collection and response dynamic dynamic characteristic measurement of the airbag array touch sensing mechanism to ensure the reliability and accuracy of the experimental data , Including the following steps:

[0052] Step 1. For the airbag array skin sensor attached to the smart robot, there are n airbag units in the horizontal and vertical directions to form an n*n linear array, which is calibrated in a specific test environment to test the reliability of the skin sensor's tactile perception;

[0053] The airbag array skin sensor provided in this embodiment is fixed on a horizontal test platform, and the static characteristics of the skin sensor are calibrated to obtain the stiffness, force displacement and other static characteristics of the airbag array skin. In this embodiment, the MTS pressure testing machine ...

Embodiment 3

[0063] Such as figure 2 As shown, figure 2 It is a schematic diagram of the stacking model integration strategy process. The fusion algorithm model provided in this embodiment is established by stacking layered model integration method and machine learning algorithm. The stacking layered model integration method is an effective method to solve the fusion between heterogeneous models , Can effectively improve the generalization ability of the model, reduce the phenomenon of over-fitting, and improve the overall effect of the model. The specific steps are as follows:

[0064] S1: Initialize the data set, m=1, representing the m-th single model, and the single model includes one of decision trees, random forests, XG-Boost or LightGBM;

[0065] S2: Set i=1, i represents the number of cross-validation;

[0066] S3: Cross training for a single model, and then set i=i+1;

[0067] S4: Obtain the prediction results of the training set and the test set under a single model;

[0068] S5: judg...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a pattern recognition method and system based on air bag array tactile sensing. The method mainly comprises the following steps that: firstly, the tactile data of different objects are collected by an air bag array robot skin sensing system; training is performed by utilizing the tactile sensing data to obtain an algorithm model for pattern recognition; and finally, for thecontact type of unknown tactile sensing data, recognition is performed by utilizing the machine learning algorithm model. According to the method, multiple machine learning algorithms such as an XG-Boost, a LightGBM mainstream distributed machine learning framework and a random forest are adopted to construct a fusion mode recognition algorithm based on a Stacking integration strategy; and therefore, a model prediction effect can be further improved, and good robustness is achieved for a large data set.

Description

Technical field [0001] The present invention relates to the technical field of robot skin perception, in particular to a pattern recognition method and system based on airbag array tactile perception. Background technique [0002] With the widespread application of robotics in the fields of industry and biomedical engineering, intelligent robots need to perform human-machine collaboration in various complex and changeable multi-information environments. When the robot's vision is limited, the sense of touch is an important source of receiving external information, and as the medium of interaction with the environment, the sense of touch of the robot skin can provide effective feedback of environmental information for the robot to perform multiple tasks. The tactile information collection of traditional robot skin pattern recognition methods usually relies on the contact of objects in limited poses and positions. Tactile sensing pattern recognition methods only refer to a single m...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/62G01L25/00G06N3/00G06N20/00
CPCG01L25/00G06N3/006G06N20/00G06F18/25G06F18/24G06F18/214
Inventor 皮阳军颜泽荣刘凡郭琦
Owner CHONGQING UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap