Natural language instruction disambiguation method and system oriented to mechanical arm grabbing

A natural language and robotic arm technology, applied in natural language data processing, manipulators, neural learning methods, etc., can solve problems such as semantic disambiguation is difficult to achieve, and the grasping task of the robotic arm cannot be correctly guided.

Active Publication Date: 2021-12-03
SUZHOU UNIV
View PDF5 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

For complex natural language instructions with different temporal logic (same semantics), it is difficult to achieve s

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
  • Natural language instruction disambiguation method and system oriented to mechanical arm grabbing
  • Natural language instruction disambiguation method and system oriented to mechanical arm grabbing
  • Natural language instruction disambiguation method and system oriented to mechanical arm grabbing

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0052] refer tofigure 1 , Embodiments of the present invention provides a natural language instruction fetch oriented manipulator disambiguation method, comprising the steps of:

[0053] S1: the coordinates of the calibration, the calibration object gripping position coordinates of each region and the releasing region;

[0054] S2: word vector model, after Corpus word processing, first take Skip-gram model and one_hot coding method, and then get the word vector by the neural network training;

[0055] S3: vector model sentence, the sentence is processed to generate vector sentence word model vector;

[0056] S4: Sentence semantic similarity calculation, the similarity between two vectors is calculated by the loss function periods (cosin), setting the threshold value of 0.9, a preliminary determination semantic consistency between the two sentences; If the similarity value is greater than 0.9, proceeds to step S5; If the similarity value is less than 0.9, then step S2;

[0057] S5:...

Embodiment 2

[0088] In order to fully demonstrate the effectiveness of the present invention, the present embodiment has collected 200 of natural language instructions with temporal logic is verified, i.e., 100 pairs of different sequential logic (same semantics) of the natural language instructions. The natural language instructions into 4 experimental groups, the four groups of language instruction to operate according to the method of the present invention, wherein the test indicator comprises three parts: the sentence semantic similarity calculation accuracy, sentence acquired sequential logic accuracy, grasping manipulator take the task success rate. The experimental data shown in Table 2:

[0089]

[0090] Table 2

[0091] From the above table shows that the sentence semantic similarity calculation accuracy, sentence sequential logic acquisition of accuracy, the robotic arm grab mission success rate of over 90%, effectively guide the robotic arm to complete the crawl task.

Embodiment 3

[0093] Based on the same inventive concept, the embodiment of the present invention provides a natural language command disambiguation system that is grasped by a robotic arm, which solves the problem of a natural language command disambiguation method of the mechanical arm. The depends is not described again.

[0094] A natural language command disagressive system that is crawling toward a robotic arm, including:

[0095] Calibration unit, calibration object grabs the area and the point position coordinates of the release area;

[0096] Word vector model, after tentieratively processing of corpus, through the word prediction model training to generate word vector;

[0097] Instrument model, process the input sentence, generate the corresponding sentence based on the word vector model;

[0098] The similarity calculation unit, calculates the similarity between the two sentences by the loss function, and initially discriminates the semantic consistency between the two sentences. If t...

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 natural language instruction disambiguation method and system oriented to mechanical arm grabbing. The method comprises the following steps: calibrating position coordinates of all points of an object grabbing area and an object releasing area; generating a word vector model through word prediction model training; processing an input sentence, and generating a corresponding sentence vector based on the word vector model; calculating the similarity between the two sentence vectors through a loss function, and preliminarily judging the semantic consistency between the two sentences; extracting useful words from the sentences, and reordering according to a time adversity priority to obtain a specified sentence sequential logic form; according to the obtained sequential logic form of the sentences, judging semantic consistency between the sentences; determining a specified object in the sentence sequential logic through the stack; and enabling the mechanical arm to grab and place the specified object in combination with the position word and the position coordinate. According to the method, semantic consistency judgment between natural language instructions is achieved, and the mechanical arm is further guided to complete a grabbing task.

Description

Technical field [0001] The present invention relates to the technical field of natural language processing, and more particularly to a natural language command disambiguation method and system that faces a robotic arm. Background technique [0002] With the continuous development and breakthrough of natural language processing and artificial intelligence technology, the performance of service robots in human-machine interaction is more natural. At present, due to the strong support of artificial intelligence, the market of intelligent service robots is expanding, and the relevant mature products have been put into a variety of environments such as hotels, families, factories. The semantic disambiguation technology of natural language processing refers to the two-way formation, the same sentence, and the semantic consistency of the two can be realized. Currently, for the complicated, timing logic (semantic logic), the robot is difficult to correctly understand its semantic and sem...

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
IPC IPC(8): G06F16/33G06F40/194G06F40/279G06F40/30G06N3/04G06N3/08B25J9/16
CPCG06F16/3344G06F40/194G06F40/279G06F40/30G06N3/04G06N3/08B25J9/16B25J9/1658Y02P90/02
Inventor 迟文政叶荣广徐晴川刘杰洪阳孙立宁
Owner SUZHOU 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