Industrial device and method of reading a tag located on an object using a model

Patent metadata

Field	Value
Publication number	US-11922256-B2
Application number	US-202117927718-A
Country	US
Kind code	B2
Filing date	Mar 22, 2021
Priority date	May 27, 2020
Publication date	Mar 5, 2024
Grant date	Mar 5, 2024

Abstract

A method of reading a tag located on an object, wherein the tag includes tag values associated with the object, where the method includes obtaining a first sensor output, by reading the tag using a tag reader, the first sensor output being at least one of an erroneous and incomplete representation of the tag values of the tag and includes deriving at least one tag value from the first sensor output using a model, where the model is based on a first set of sensor outputs generated by the tag reader by reading a plurality of tags and tag values associated with the plurality of tags, and where the first set of sensor outputs includes a plurality of sensor outputs, where each sensor output includes at least one of an erroneous and incomplete representation of at least one tag value of the corresponding tag.

First claim

What is claimed is:

1. A method of reading a tag located on an object, the tag comprising at least one tag value associated with the object, the method comprising:
a) obtaining a first sensor output, by reading the tag using a tag reader, the first sensor output being at least one of an erroneous and incomplete representation of the at least one tag value of the tag; and
b) deriving the at least one tag value from the first sensor output utilizing a model;
wherein the model is based on a first set of sensor outputs generated by the tag reader by reading a plurality of tags and tag values associated with the plurality of tags; and
wherein the first set of sensor outputs comprises a plurality of sensor outputs, each sensor output of the plurality of sensor outputs from the first set of sensor outputs comprising at least one of an erroneous and incomplete representation of the at least one tag value of a corresponding tag.

2. The method as claimed in claim 1 , wherein the object is moving with a first velocity from a first position to a second position; and wherein the tag reader is affixed at a third position between the first and second positions.

3. The method as claimed in claim 1 wherein the model is trained utilizing a training data set comprising a first set of sensor outputs generated by the tag reader by reading a plurality of tags and a second set of sensor outputs from a second tag reader by reading the plurality of tags; and wherein each sensor output from the second set of sensor outputs, is generated by the second tag reader by reading a corresponding tag from the plurality of tags and forms a complete representation of at least one tag value of the corresponding tag.

4. The method as claimed in claim 3 , wherein the object is moving with a second velocity from a fourth position to a fifth position, the second tag reader being affixed at a sixth position between the fourth and fifth positions, the second velocity being less than the first velocity.

5. A method of training a model, the method comprising:
a) receiving a first set of sensor data from a first tag reader, the first set of sensor data comprising a plurality of sensor outputs generated by the first tag reader by reading a plurality of tags, each tag of the plurality of tags being located on a corresponding object moving at a first velocity, and each sensor output from the first set of sensor data being at least one of an erroneous and incomplete representation of at least one tag value of a corresponding tag;
b) receiving a second set of sensor data from a second tag reader, the second set of sensor data comprising a plurality of sensor output generated by the second tag reader by reading the plurality of tags, each tag of the plurality of tags being located on a corresponding object moving at a second velocity, and each sensor output from the second set of sensor data forming a complete representation of at least one tag value of the corresponding tag; and
c) training the model by mapping the first set of sensor data with the second set of sensor data, each sensor output from the first set of sensor data of a corresponding tag being mapped to a sensor output from the second set of sensor data of the corresponding tag based on a tag identifier of the corresponding tag.

6. An industrial device for determining at least one tag value of a tag located on an object, the industrial device comprising:
a) a network interface connected to a tag reader;
b) at least one processor connected to a memory module, the at least one processor being configured to:
i) receive a first sensor output from the tag reader via the network interface, the first sensor output being generated by the tag reader by reading the tag located on the object and the first sensor output being at least one of an erroneous and incomplete representation of the at least one tag value of the tag; and
ii) derive the at least one tag value from the first sensor output utilizing a model;
wherein the model is based on a first set of sensor outputs generated by the tag reader by reading a plurality of tags, and tag values associated with the plurality of tags; and
wherein the first set of sensor outputs comprises of a plurality of sensor outputs, each sensor output from the first set of sensor outputs comprising at least one of an erroneous and incomplete representation of the at least one tag value of the corresponding tag.

7. A training system for training a model for determining at least one tag value of a tag, the training system comprising:
a) a first tag reader installed between a first and a second position, within a proximity of a transportation system, the first tag reader being configured to generate a first set of sensor outputs by reading a plurality of tags, each tag of the plurality of tags being located on a corresponding object moving at a first velocity on the transportation system, and each sensor output from the first set of sensor outputs from the first set of sensor outputs being at least one of an erroneous and incomplete representation of at least one tag value of a corresponding tag;
b) a second tag reader installed between a third and a fourth position, within a proximity of the transportation system, the second tag reader being configured to generate a second set of sensor outputs by reading the plurality of tags, each tag of the plurality of tags being located on a corresponding object moving at a second velocity on the transportation system, and each sensor output from the second set of sensor outputs forming a complete representation of at least one tag value of the corresponding tag; and
c) an industrial device connected to the first and second tag readers, the industrial device being configured to:
i) receive the first set of sensor outputs from the first tag reader and the second set of sensor outputs from the second tag reader; and
ii) train the model by mapping the first set of sensor data with the second set of sensor data, each sensor output from the first set of sensor data of a corresponding tag being mapped to a sensor output from the second set of sensor data of the corresponding tag based on a tag identifier of the corresponding tag.

8. A non-transitory computer-readable storage medium encoded with a plurality of program instructions, which when executed by at least one processors, cause the at least one processor to:
a) obtain a first sensor output, by reading a tag located on an object using a tag reader, the tag comprising at least one tag value associated with the object, and the first sensor output being at least one of an erroneous and incomplete representation of the at least one tag value of the tag; and
b) derive the at least one tag value from the first sensor output utilizing a model;
wherein the model is based on a first set of sensor outputs generated by the tag reader by reading a plurality of tags and tag values associated with the plurality of tags, the first set of sensor outputs comprising a plurality of sensor outputs, and each sensor output from the first set of sensor outputs comprising at least one of an erroneous and incomplete representation of at least one tag value of a corresponding tag.

9. A non-transitory computer-readable storage medium for training a model, the non-transitory storage medium being encoded with a plurality of program instructions, which when executed by at least one processor, cause the at least one processor to:
a) receive a first set of sensor data from a first tag reader, the first set of sensor data comprising a plurality of sensor outputs generated by the first tag reader by reading a plurality of tags, each tag from on the plurality of tags being located on a corresponding object moving at a first velocity, a

Assignees

• Siemens Ag

Inventors

• Bär Felix
• Gross Ralf
• Lavrik Viadimir

Classifications

• G06K7/10366 Primary

  the interrogation device being adapted for miscellaneous applications · CPC title

• G06K7/1482 Primary

  using fuzzy logic or natural solvers, such as neural networks, genetic algorithms and simulated annealing · CPC title

• G06K7/10

  by electromagnetic radiation, e.g. optical sensing; by corpuscular radiation · CPC title