=Paper=
{{Paper
|id=Vol-1940/paper03
|storemode=property
|title=Development of an Algorithm for Determining the Movement of Products Between Racks Based on Data from Their Radio Frequency Tags
|pdfUrl=https://ceur-ws.org/Vol-1940/paper03.pdf
|volume=Vol-1940
|authors=Alexandr Astafiev,Alexey Orlov,Dmitry Popov,Maxim Pshenichkin
}}
==Development of an Algorithm for Determining the Movement of Products Between Racks Based on Data from Their Radio Frequency Tags==
https://ceur-ws.org/Vol-1940/paper03.pdf
Development of an Algorithm for Determining the
Movement of Products Between Racks Based on Data
from Their Radio Frequency Tags
Alexandr Astafiev, Alexey Orlov, Dmitry Popov, and Maxim Pshenichkin
Murom Institute of Vladimir State University, Murom, Russia
AlexeyAlexOrlov@gmail.com
Abstract. At present, in connection with the need to move to new intellectual
digital production technologies and implement international quality standards, it
is necessary to introduce new science-based approaches to controlling the
movement of products and small-scale mechanization of warehouses. This is
due to the fact that the warehouses of large industrial organizations, at the
current level of hardware and software, can not fully comply with domestic and
foreign quality standards in the field of product tracking, regulated by GOST
and ISO. The article is devoted to the development of an algorithm for
determining the movement of products between racks based on data from their
radio frequency tags to automate the control of the movement of products in
industrial enterprises. The main methods of development of such systems based
on technical vision and radio frequency identification are considered. The main
normative documents and interstate standards regulating the described
processes are given. The review of modern publications and automation
systems for the control of the movement of products of domestic and foreign
manufacturers is conducted. The developed algorithm operates radio-frequency
labels of racks and products. Data processing is performed using statistical
methods of analysis. The results of experimental research at Vyksa Steel Works
are presented.
Keywords: positioning, radio frequency identification, racks.
1 Introduction
One of the most important elements of the quality management system of industrial
production, which largely determines the efficiency of its operation, is the
identification mechanism, which makes it possible to ensure traceability of products
throughout the technological production cycle. Traceability in production helps to
ensure compliance with the requirements of government and international quality
standards, to perform a fast and purposeful tracking of the entire technological cycle
of manufacturing products, which, in turn, minimizes financial consequences.
Especially relevant is the question of tracking products at the enterprise, if the
production cycle consists of many stages, implemented in large production areas. The
organization of the mechanism for tracking products is possible by automating the
control of the movement of industrial products [1,2].
18
Automation of traffic control is currently predominantly made using two
technologies: technical vision and radio frequency identification (RFID). The use of
technical vision approaches is complicated by the need for graphic marking of proper
quality, which is difficult to realize in real production conditions and requires
significant financial and human resources. The use of radio frequency identification is
less demanding in the process of marking products. Traffic control, based on radio
frequency identification methods, is an advanced information technology for the
construction of warehouse accounting systems [3,4].
2 Subject Overview
When solving problems of controlling the movement of products during its
stochastic movement by various transport devices, either methods of digital image
processing coming from video sensors or methods of radio frequency identification
are often used. The use of methods of digital image processing provides great
opportunities for implementing various ways of identifying symbolic markings
printed on the surface of manufactured products. In most cases, developers of such
solutions prefer modern controlled cameras, which are equipped with rotary devices,
optical zoom and automatic focus adjustment, to obtain a variety of images and
analysis of disparate features. To develop such solutions it is necessary to have a high
level of expertise in the field of digital image processing, and the algorithms
themselves are very narrow-based. During the implementation of the project, it is
proposed to develop methods and algorithms for the detection, localization and
recognition of symbolic markings, which make it possible to reliably identify
industrial products on a variety of images using a wide variety of disparate features.
Analysis and selection of signs of symbolic markings should be carried out in
accordance with existing international and interstate standards [5-11]:
1. GOST R ISO / IEC 15459-3-2007 "Automatic identification. Identifiers are
unique international. Part 3. General rules for unique identifiers »
2. GOST R ISO / IEC 15459-4-2007 "Automatic identification. Identifiers are
unique international. Part 4. Unique Identifiers of Single Items for Supply Chain
Management »
3. GOST 27465-87 "Information processing systems. Symbols. Classification,
name and designation »
4. GOST R 51294.2-99 "Automatic identification. Encoding is a dashed.
Description of the format of the requirements for symbolism »
5. GOST 30832-2002 "Automatic identification. Encoding is a dashed. Linear
symbols of the bar code. Testing requirements for print quality »
6. GOST ISO 15394-2013 "Packing. Linear bar code symbols and two-
dimensional symbols on labels for shipping, transportation and acceptance. General
requirements"
7. GOST ISO / IEC 15459-1-2008 "Automatic identification. Identifiers are unique
international. Part 1. Unique identifiers of transported units»
19
The development of new methods and algorithms for the detection, localization
and recognition of symbolic markings is aimed at improving the two main technical
and operational indicators: - reduction in the speed of the processing of incoming
multi-image images on a set of dissimilar features; - increasing the reliability of the
results obtained during the analysis of different digital images. The requirements for
performance and reliability indicators are determined experimentally. However, the
use of symbolic markings is complicated by the fact that the execution of the
hardware, its location and the quality of the installation do not always allow us to
cover the entire area of potential identification of the product identifier, which in turn
has a negative impact on the reliability of the results obtained. Also, the symbolic
markings applied to the surface of the monitored products may be damaged during
transport, which leads to the impossibility of identifying the current product, which
also has a negative effect on the reliability of the results obtained. The use of radio
frequency identification, at present, is a very popular and effective approach to
identification, because Allows to bypass a number of restrictions imposed by digital
image processing. For implementation of radio frequency solutions, both manual and
stationary readers can be used with the option of installing remote antennas. To
develop such solutions, it is necessary to have a high level of expertise in the field of
radio frequency identification and to take into account limitations on the use of this
equipment. During the implementation of the project, it is proposed to develop
methods and algorithms for the statistical analysis of data of radio frequency
identifiers to determine the position relative to storage locations and identifiers of
captured items. Identification is proposed to be made by standard means of radio
frequency readers, when positioning - using statistical data. The determination of the
movement of products is planned to be controlled by analyzing the space-time
information on the state of the transporting device. The analysis and choice of the
signs of radio frequency identifiers and space-time characteristics should be carried
out in accordance with existing international and interstate standards [12-18]:
1. GOST R 54621-2011 "Information technology. Radio frequency identification
for the management of objects. Recommendations for use. Part 1. Labels and
packaging with RFID tags according to ISO / IEC 18000-6 (type C) »
2. GOST R ISO / IEC 19762-3-2011 "Information technology. Automatic
identification and data collection technologies (AISD). Harmonized dictionary. Part 3.
Radio Frequency Identification (RFID) »
3. GOST R ISO / IEC 18000-6-2013 "Information technology. Identification of
radio frequency for the management of objects. Part 6. Radio interface parameters for
the frequency range 860 - 960 MHz. General requirements"
4. GOST R ISO / IEC 18000-7-2012 "Information technology. Identification of
radio frequency for the management of objects. Part 7. Active radio interface
parameters for communication at 433 MHz »
5. GOST R ISO / IEC 15459-3-2007 "Automatic identification. Identifiers are
unique international. Part 3. General rules for unique identifiers »
6. GOST R ISO / IEC 15459-4-2007 "Automatic identification. Identifiers are
unique international. Part 4. Unique Identifiers of Single Items for Supply Chain
Management»
20
7. GOST R ISO / IEC 15963-2011 "Information technology. Radio frequency
identification for the management of objects. Unique identification of radio frequency
tags»
The development of new methods and algorithms to determine the position and
identifier of radio frequency marking is aimed at improving the two main technical
and operational indicators: - reducing the speed of processing incoming RF data for
positioning the product relative to storage locations; - increasing the reliability of the
results obtained during the statistical analysis of radio frequency identifiers. With all
the advantages of radio frequency identification, there are a number of limitations that
can adversely affect the result of the automatic identification system. One such
limitation is the directivity of the signal of radio frequency identifiers fixed on the
products. The presence of this parameter rigidly sets the requirements for the
composition, location and installation of equipment for identification. Another
limitation is that radio frequency identifiers poorly pass radio waves through
obstacles, and especially through metal. Failure to comply with the rules for the
operation of radio frequency equipment can have a negative impact on the reliability
of the results obtained.
A great contribution to the development of radio frequency identification
technology and product movement control systems in various spheres of life was
made by Bondarevsky A.S., Zolotov R.V., Do Zuy Nyat, Kamozin D.U., Manish B.,
Shahram M., Ke-Sheng Wang , Worapot Jakkhupan, Somjit Arch-int, Yuefeng Li,
Mahir Oner, Alp Ustundag, Aysenur Budak and many others [5-10].
Application of these knowledge-intensive technologies makes it possible to
automate the processes of controlling the product movement at industrial plants,
ultimately, to increase the efficiency and reliability of transportation control and
warehouse inventory control of manufactured products.
However, they are not without flaws. The use of existing software and hardware
solutions is more aimed at organizing automated warehouse inventory control and less
suitable for automating product movement control, in the absence of universal
methods and algorithms. In support of this, at a number of industrial enterprises,
developers of RFID systems attempted to organize traceability of products by
automatic movement control based on radio frequency identification. As a result, it
became clear that automatic control of the product movement is possible only in
certain areas of the production process. Such areas are conveyor lines and transport
tunnels, where the transportation of products is carried out along the permanently
installed radio frequency identification equipment (RFID tunnels). In other production
and warehouse areas, automatic control over the movement of products is impossible.
This is due to the lack of universal methods and algorithms for product identification
in the process of its transportation along unmapped routes.
Positioning of objects and people using information technology is quite a
substantial task. These technologies can be used to solve social, industrial and other
types of tasks. Currently, there are a large number of approaches to positioning using
a large number of technologies, among them:
- Satellite navigation technologies (GPS, GLONASS);
- Local positioning technology (infrared and ultrasonic);
21
- Technology of technical vision;
- Radio-frequency technologies.
The use of satellite navigation technology and positioning are tightly integrated
into our daily lives. They are used for navigation and transport tracking, monitoring
and coordination of various kinds of events. The accuracy of positioning is 10-15
meters outdoors. Unfortunately, the application of this technology inside production
facilities is almost impossible. An exception is the installation of expensive
equipment for organizing GPS-positioning indoors, the unit of which can cover no
more than 10 square meters, which is unacceptable for most industrial plants, whose
sizes can be tens of kilometers.
Local positioning technologies are highly accurate - about 2 centimeters, but with a
short range of 5-10 meters. With these attributes, they are used to achieve local
accurate results and, in general, are used for flaw detection (analysis of welds,
detection of chips, dents, etc.). The use of local positioning technology for small-scale
mechanization is not economically effective and will lead to huge financial costs.
The use of vision technology for solving positioning problems is a relatively young
concept. Currently, there are a huge number of methods and algorithms for solving
localization and positioning problems, but their effectiveness depends a lot on
meeting a large number of requirements, which include the quality of materials used
for production of visual labels, cleanliness and lighting of premises, staff
attentiveness, etc. Failure to comply with even one of the requirements can lead to a
significant reduction in positioning accuracy or make it completely inoperative.
Radio-frequency technologies have found wide application in sales (organization
of security in stores). Positioning based on radio frequency technologies can be
divided into two categories: positioning on passive RFID tags (distance up to 5
meters) and active RFID beacons (distance up to 80 meters), but all of them are based
on the principle that the moved object is marked with an RFID tag and the reading
equipment is stationary. This approach allows to effectively automate production
processes, where the product movement routes are strictly limited, for example,
conveyor lines. However, for the positioning of chaotically moving small
mechanization means, this approach will lead to a significant increase in the cost of
the positioning system. Instead of a few readers they would need ten times as many.
Considering all the information stated above it is possible to draw a conclusion,
that development of technology and software for the construction of positioning and
control systems for small mechanization in industrial plants based on radio frequency
identification methods is a substantial scientific and technical task.
The development of software and hardware for movement control systems is
carried out by: PCT-Invent (Russia, Saint-Petersburg), AiTiProject (Russia, Moscow),
Impinj (USA, Seattle), Motorola (USA, Morrisville), Nordic ID (Finland, Salo), FEIG
(Germany, Weilburg).
Development of positioning systems based on radio frequency identification is
carried out by the following scientific organizations:
- Human positioning systems, in particular patients in medical institutions: Shonan
Institute of Technology (Japan, Fujisawa), Institute of Medicine (Kathmandu, Nepal),
National Patient Safety Foundation (USA, Boston) and others.
22
- Systems for positioning moving non-metallic objects: East China Jiaotong
University (China, Nanchang), Universiti Sains Malaysia (Malaysia, Nibong Tebal),
University of Adelaide (Australia, Adelaide), Wellness Convergence Research Center
(Korea, Daegu) and many others.
However, the tasks of developing and implementing automatic systems for
tracking products in production are still unresolved. Currently, industrial enterprises
still have a number of problems, the solution of which is not realized with the help of
modern product movement control systems.
3 Development of an Algorithm for Determining the Movement
of Products Between Racks Based on Data from Their Radio
Frequency Tags
Let J be the log of product movements (the log is the result of the algorithm), g (t1,
t2) is the function that returns the set of identifiers of the moved products in the time
interval (t1, t2) (to extract the identifier of the moved product, the time-averaged value
of the signal power level is compared with the threshold value):
〈 〉
, , ∈ ,
,
Where is the set of product label identifiers, is the threshold
value (the minimum value of the signal power level from the tag of the product being
〈 〉
moved), is the vector of signal strength levels from the product labels and racks at
time t.
The power level of the signal from the label is calculated as the average value in a
time τ:
〈 〉 1
,
where N is the number of read requests during the time interval (t, t + τ), n is the
number of reads of the label i, μ is the signal power level from the label i, μave is the
average value of the signal power level from the label i.
We also introduce the intermediate variables: Ibegin is the identifier of the initial
rack (the rack from which the load is moved), I is the identifier of the current rack
(the rack over which the load is at the time t), Iprevious is the identifier of the previous
rack (the previous value of I), K is the number of shifts of the racks (the number of
changes in the value of I starting from the initial Ibegin), tfirst is the time of the first
change of the rack, tlast is the time of the last change of the rack, taverage_of_void is the left
boundary of the time interval during which there is no signal from the shelves.
23
Fig. 1 Scheme of moving between racks without a void
Fig. 2 Scheme of movement between the racks with a void
Algorithm steps:
1) Zeroing the values of intermediate variables (let zero mean the uncertainty of
the value):
_ _ 0.
2) All the following steps (iteration) of the algorithm are performed for each time
interval τ, during which a representative sampling of data from the marks is
performed (each interval will be identified by its left boundary t).
3) The identifier of the rack label is defined as the index of the maximum value of
the signal power level (the zero identifier will denote the absence of a signal from the
labels, i.e., I = 0):
24
〈 〉 〈 〉 〈 〉
match max , , max
,
0, ,
Where match (·, ·) is the function that returns the index of the given element (the
first parameter) in the vector (the second parameter), is the signal level
threshold below which it will be considered to be absent, 〈 〉 is the vector of signal
power level values from labels of the racks at time t.
Note. ∈ ( – is set of label identifiers for racks).
4) If the initial rack is not defined (Ibegin = 0) and the signal from the marks is
present (I ≠ 0), then the initial rack is defined as the current:
Ibegin = I.
5) If the previous rack is not defined (Iprevious = 0) and the signal from the marks is
present (I ≠ 0), then the previous rack is also determined as long as the current one
(the value of the identifier of the previous rack is necessary at the next iteration):
Iprevious = I.
6) If the values of the previous rack are determined (Iprevious ≠ 0 и I ≠ 0) and the
identifiers of the current and previous rack are different (Iprevious ≠ I ), This means that
at a given moment of time t, when moving the load with the tag reader, the signal
power level from the current rack marker has exceeded (at that moment the load is
somewhere between the two racks), and the steps are performed 7-13.
7) If there is no signal from the labels between the racks, the instant of time for
changing the racks is refined and taken as the middle of the absence interval:
_ _
_ _ .
2
8) If the change of the racking occurred for the first time (while k = 0), then the
obtained time is taken as the time of the first event of changing the rack:
_ _ .
9) As the time of the last event of changing the rack, time is always
taken _ _ :
_ _ .
10) If during the movement of the load the rack has changed several times (
), but the identifiers of the labels of the transferred products have also changed,
i.e. the composition of cargo changed
∆ , ∆ ∆ , ,
where ∆ is half the time of moving the captured cargo over the racks, it means that it
is necessary to create a record of the movement of the load from the initial to the
previous rack (steps 11, 12 are performed).
Note. In the above formula to not added ∆ , i.e. The composition of cargo in
the future has not yet been determined.
11) In the journal J information is written about that from the rack Ibegin in rack
Iprevious there was a movement of cargo with identifiers ∆ , ∆ :
∪〈 , , ∆ , ∆ 〉.
12) The values of the intermediate variables are overridden. Because The
movement of a new cargo has already begun, then:
25
- the initial rack becomes the previous one
I begin = Iprevious;
- the number of shifts of the rack will be equal to one
k = 1;
- the time of the first event of the change of the rack will be the time of the last
change of the rack
.
13) Otherwise (the condition of step 10 is not fulfilled, i.e., when the rack is
changed the composition of the cargo has not changed), the current rack is taken as
the previous rack, and the number of shifts of the racks is increased by one (the
resulting values of these variables are used at the next iteration):
Iprevious = I,
k = k + 1.
14) If the signal from the rack label is present (I ≠ 0) and assuming that the signal
from the rack label disappears at the next iteration, the left boundary of the time
interval during which the signal from the racks is absent will be the time instant at the
next iteration:
_ _ .
If the signal from the label of the rack is not lost for the next iteration, then the
value of this time is also redefined and will be used in the subsequent iteration.
15) If the moment of the last change of the rack is determined (tlast> 0), and from
that moment the time passed exceeds the threshold value T (waiting time including
the time Δt and the unloading time):
,
then steps 16 and 17 are performed.
16) In the journal J information is written about that from the rack Ibegin to rack I
there was a movement of cargo with identifiers ∆ , ∆ :
∪〈 , , ∆ , ∆ 〉.
Note. If the signal from the shelving labels is not present at the moment (I = 0),
then to determine the freelance situation in this log record, it is also necessary to add
information about the absence of this signal "the signal from the rack marks is
missing".
17) The values of the intermediate variables for the next iteration again become
undefined:
0.
4 Experimental Research
During the pilot studies, a large number of different typical close to production
situations were modeled (Figure 3). Among them:
– movement between two storage areas;
– movement between three or more storage areas;
26
– movement between storage areas with the presence of "noise" (other radio
frequency tags that are not tags of storage areas)
– movement between storage areas with partial overlapping of non-metallic and
metal barriers.
Fig. 3 The results of experimental studies
Experimental research was carried out at the industrial enterprise of JSC Vyksa
Steel Works. During the research, the labeled metal products were moved between the
shelves by means of small-scale mechanization, in particular a bridge crane with a
load-carrying beam. The technological map of the product movement is shown in
Figure 4.
Fig. 4. A technological map for the transport of products by overhead cranes with a
load-carrying cross-beam
Figure 5 shows the interpreted data on four experiments on beam movement
between racks.
27
Fig. 5 - The results of experimental research at JSC Vyksa Steel Works
Experimental studies have shown the correctness of the algorithm for determining
the current storage zone in laboratory and production conditions.
The work is executed at financial support of the grant of the President of the
Russian Federation № МК-991.2017.9.
References
1. Astafiev A., Development of Methods for Determining the Locations of Large Industrial
Goods During Transportation on the Basis of RFID // A. Provotorov, D. Privezentsev, A.
Astafiev / Procedia Engineering Volume 129, Pages 1005-1009 (2015) DOI:
10.1016/j.proeng.2015.12.163
2. Astafiev, A.V. The localization algorithm of symbolic and bar-code labels on industrial
products for the control of product movements // Astafiev, A.V. Orlov, A.A. ; Provotorov,
A.V. / "Stability and Control Processes" in Memory of V.I. Zubov (SCP), 2015
International Conference, St. Petersburg, 5-9 Oct. 2015, pp. 615 - 616, DOI:
10.1109/SCP.2015.7342230
3. Astafiev, A.V Method of controlling the movement of large metal products with the use of
algorithms for localization and recognition of bar code markings / Astafiev, A.V, Orlov,
A.A., Privezentsev, D.G. // 2016 Dynamics of Systems, Mechanisms and Machines,
Dynamics 2016, Omsk. DOI: 10.1109/Dynamics.2016.7818969
4. Zhiznyakov, A.L. Using fractal features of digital images for the detection of surface
defects / Zhiznyakov, A.L., Privezentsev, D.G., Zakharov, A.A. // Pattern Recognition and
Image AnalysisVolume 25, Issue 1, 2015, Pages 122-131
5. GOST R ISO / IEC 15459-3-2007 "Automatic identification. Identifiers are unique
international. Part 3. General rules for unique identifiers »
6. GOST R ISO / IEC 15459-4-2007 "Automatic identification. Identifiers are unique
international. Part 4. Unique Identifiers of Single Items for Supply Chain Management »
7. GOST 27465-87 "Information processing systems. Symbols. Classification, name and
designation »
8. GOST R 51294.2-99 "Automatic identification. Encoding is a dashed. Descrip-tion of the
format of the requirements for symbolism »
9. GOST 30832-2002 "Automatic identification. Encoding is a dashed. Linear symbols of the
bar code. Testing requirements for print quality »
10. GOST ISO 15394-2013 "Packing. Linear bar code symbols and two-dimensional symbols
on labels for shipping, transportation and acceptance. General requirements"
11. GOST ISO / IEC 15459-1-2008 "Automatic identification. Identifiers are unique
international. Part 1. Unique identifiers of transported units»
28
12. GOST R 54621-2011 "Information technology. Radio frequency identification for the
management of objects. Recommendations for use. Part 1. Labels and packaging with
RFID tags according to ISO / IEC 18000-6 (type C) »
13. GOST R ISO / IEC 19762-3-2011 "Information technology. Automatic identifi-cation and
data collection technologies (AISD). Harmonized dictionary. Part 3. Radio Frequency
Identification (RFID) »
14. GOST R ISO / IEC 18000-6-2013 "Information technology. Identification of ra-dio
frequency for the management of objects. Part 6. Radio interface parameters for the
frequency range 860 - 960 MHz. General requirements"
15. GOST R ISO / IEC 18000-7-2012 "Information technology. Identification of ra-dio
frequency for the management of objects. Part 7. Active radio interface parame-ters for
communication at 433 MHz »
16. GOST R ISO / IEC 15459-3-2007 "Automatic identification. Identifiers are unique
international. Part 3. General rules for unique identifiers »
17. GOST R ISO / IEC 15459-4-2007 "Automatic identification. Identifiers are unique
international. Part 4. Unique Identifiers of Single Items for Supply Chain Management»
18. GOST R ISO / IEC 15963-2011 "Information technology. Radio frequency identification
for the management of objects. Unique identification of radio frequency tags»