Sensors 20 03709 v2
Sensors 20 03709 v2
Article
Reliable Identification Schemes for Asset and
Production Tracking in Industry 4.0
Attila Frankó 1, * , Gergely Vida 1 and Pal Varga 2
1 Industrial IoT Division, AITIA International Inc., 48–50, Czetz János u., 1039 Budapest, Hungary;
[email protected]
2 Department of Telecommunications and Media Informatics, Budapest University of Technology and
Economics, 2, Magyar Tudósok krt., 1117 Budapest, Hungary; [email protected]
* Correspondence: [email protected]
Received: 25 May 2020; Accepted: 28 June 2020; Published: 2 July 2020
Abstract: Revolutionizing logistics and supply chain management in smart manufacturing is one
of the main goals of the Industry 4.0 movement. Emerging technologies such as autonomous
vehicles, Cyber-Physical Systems and digital twins enable highly automated and optimized solutions
in these fields to achieve full traceability of individual products. Tracking various assets within
shop-floors and the warehouse is a focal point of asset management; its aim is to enhance the efficiency
of logistical tasks. Global players implement their own solutions based on the state of the art
technologies. Small and medium companies, however, are still skeptic toward identification based
tracking methods, because of the lack of low-cost and reliable solutions. This paper presents
a novel, working, reliable, low-cost, scalable solution for asset tracking, supporting global asset
management for Industry4.0. The solution uses high accuracy indoor positioning—based on
Ultra-Wideband (UWB) radio technology—combined with RFID-based tracking features. Identifying
assets is one of the most challenging parts of this work, so this paper focuses on how different
identification approaches can be combined to facilitate an efficient and reliable identification scheme.
1. Introduction
The Industry4.0 movement [1] has a huge impact on all manufacturing related fields including
logistics and Supply Chain Management (SCM), which are often referred as Logistics 4.0 and SCM
4.0 [2]. While smart manufacturing considered as the main target of CPS (Cyber-Physical System)
based digitization trends; logistics—especially resource planning and warehouse management—and
SCM can also derive benefits from inter-twining physical and digital planes. A vision of CPS driven
warehouses includes autonomous vehicles moving assets from one place to another, based on the data
coming from different parts of the supply chain to achieve just-in-time and just-in-sequence delivery.
Meanwhile each asset is fully traceable via its Digital Twin counterpart and therefore the inventory
updates itself in an automated manner [3,4].
Many different levels of automation exist concurrently in the industry that depend mostly
on the size of a company, but the key element has always been the traceability of a product or
an asset. This includes not only tracking assets physically within the factory site (e.g., shop-floors and
warehouses), but also following their life-cycles and updating, managing their statuses.
The main objective of Productive 4.0, the biggest European research project on Digital Industry
to date [5] is to achieve improvement of digitalizing the European industry by electronics and ICT.
The three main pillars for the holistic system approach of Productive 4.0 are digital production,
supply chain networks and product lifecycle management, all of which interact and influence each
other. This current paper reports on an actual use-case of this approach. In this, the results of
the digital production (assets) [6] are distributed through the supply chain network (including tasks
of warehousing and logistics) [7], and the lifecycle steps [8] of the product can also be monitored
through its digital footprint (status and ownership changes get noted at its digital twin). The current
paper focuses on the asset tracking aspect of this use-case. The aim of this paper is to cover the gap
of the Industry 4.0 movement regarding fully digitalized asset tracking to be deployed by mid-scale
manufacturing and logistics sites.
Medium or small-sized companies that usually operate in less, or non-automated environments
lack nearly any digital capability to trace products in most of the cases; except administrating
the properties of a product on a single sheet of paper attached to the asset. This way of handling
products leads to warehouses where finding assets can be a burdensome task because of using Last
In, First Out (LIFO) stocking method. In such a situation, when it is onerous to locate products or to
access them, the preferred solution can be leaving the required asset as it is and searching for—or
sometimes producing—a new one which is easily accessible. This approach strongly reduces overall
efficiency and productivity, because of increased time of searching for goods and hence increased
manufacturing time.
Automated management of asset tracking can make significant improvement in this field by
eliminating the aforementioned issues. The key concept is following assets in real-time, which enables
us to verify the current location of a piece of product, therefore assets can be fully traceable
within the warehouse. This automation can highly enhance the efficiency of logistical tasks,
even in environments that do not go for full automation. Most SMEs keep themselves aloof from such
a development due to its high cost and the infrastructural changes that may be required.
The current paper presents a cost effective, automated asset tracking and management system
which can aid logistical tasks and optimize supply chain management related processes. The solution
consist of:
• a real-time indoor positioning subsystem (IPS) based on Ultra-wideband (UWB) radio technology
which provides accurate and precise location information;
• an Ultra High Radio Frequency Identification (UHF-RFID) based tracking subsystem using
a special identification scheme that is proposed and detailed in this paper;
• the so-called Core system, which handles the information exchange between these two underlying
subsystems as well as maintains the asset tracking logic;
• the communication system that allows uninterrupted information flow between the tracked
asset’s RFID reader, the IPS system, and the Core system;
• various interfaces to visualization front-ends and external data processing systems, such as ERP
(Enterprise Resource Planning) solutions.
Creating a system that is financially worthwhile to adopt for even small-sized companies is
one of the main motivations of this work. Therefore, our solution implements an indirect tracking
scheme—the IPS subsystem captures the movement of various vehicles and tools (later called as mule
in short) that can move actual products, meanwhile the tracking subsystem identifies which asset is
being moved at the moment. This method is comparably more economical than following each asset
individually, but it could also be error-prone due to the difficulties of determining when an asset have
been picked up or put down.
The current paper contributes to the state-of-the-art by the following:
• it raises awareness that the Industry 4.0 movement have to keep refocusing to target small-sized
and mid-sized manufacturing and logistics scenarios as well, which require the digitalization
solutions to have low investment cost;
• it provides an overview of production asset-related reliable identification schemes, and their
underlying methods;
Sensors 2020, 20, 3709 3 of 24
• it introduces and details the idea of the indirect tracking scheme, where the assets get associated
with the domain where it resides—be it either a warehouse slot or a moving vehicle; and
in the latter case the asset’s position indirectly tied with the vehicle position;
• it describes a novel, intrinsically low-cost, highly scalable indoor positioning system—that
supports geo-fencing as well as the indirect tracking scheme—which allows the number of
trackable assets to be increased by magnitudes without any adjustments of the overall system
infrastructure;
• it presents the evaluation results of the system through real-life, practical measurements,
especially showing that positioning accuracy can be significantly improved by properly executed
simple (low calculation-cost) methods, such as moving averages.
This paper is organized to meet the methodology of information system design science [9] as
follows—describing problem relevance and context, related work and gap analysis, system architecture
as novel artifact description, evaluation setup, presentation of evaluation results, and finally,
the discussion of findings and contribution. Therefore, Section 2 describes the related work including
supporting technologies as well as revealing gaps, then Section 3 provides an overview of the system
architecture, the concept and the layouts for the measurements, whereas Section 4 details our
measurement scenarios used during validation, after which Section 5 evaluates the results related to
the identification scheme. Section 6 concludes the paper.
2. Related Works
where multiple technologies are used to fulfill different sets of requirements, however these are rather
two different implementations that are coupled [14,15].
A1 A2
r21
r11
r22
T1
T2
r31
r32 r42
A3 A4
Figure 1. An example system implementing time difference of arrivals (TDOA) and trailateration based
localization, where Ai -s are anchor points, Ti -s are UWB tags and rij -s are distances.
While Bluetooth, WiFi, ZigBee and RFID based localization techniques are considered low-cost,
UWB transceivers are usually more expensive, but this technology can be more accurate than the other
ones [16–18]. However, the accuracy of a system highly depends on the method itself and other
supplementary calculations for example, filtering, predictions [19–21] and auxiliary units (e.g., IMU-s,
cameras) and supporting technologies [22,23]. There are plenty of methods to improve the accuracy
of the aforementioned techniques, but generally speaking these methods often require much more
computational capacity or more complex and dense infrastructure. Contrarily, a TDOA method based
on UWB can provide an accuracy of 30 cm without any further calculations, while the response time
of the system is still really fast—for example, even under 100 ms, considering a cheap (under $5)
controller unit. Moreover, the complexity and the density of the infrastructure remains low, compared
to other solutions. It is also well known that UWB has much larger power consumption than
Bluetooth or ZigBee, however its energy efficiency is better, therefore it is more sufficient for
high-data rate applications—for example, providing real-time, highly granulated location information
as an Industry 4.0 compliant solution [24].
be interrogated in a more reliable way, although they are supplied by a battery and have a significantly
higher price tag.
The range of a High-frequency (HF) RFID reader is usually between 10 cm and 1 m which is
acceptable for identifying for smart cards, but it is inappropriate in use-cases where the distance
between a reader and the assets can not be decreased due to physical boundaries. In contrast to this,
the range of UHF-RFID reader is between 1 m and 10 m—that can be expanded to 100 m by using
active tags. Moreover, UHF-RFID standard has an extra capability that is not supported in HF-RFID,
the simultaneous or bulk reading. During bulk reading reader is able to interrogate multitude of passive
tags simultaneously in one reading cycle which can be extremely useful in logistical use-cases where
the individual assets are often bundled together. The hit-rate of the method—the ratio of identified
(read) tags to all tags—depends on the performance of the reader and the positions of antennas
as Ustundag et al., presented in Reference [26], therefore it seems the higher the required hit-rate,
the higher the cost of the reader. However, hit-rate can be improved with different methods such as
adaptive time and power control too [27].
It is worth to note that RFID itself can used as core technology for an IPS system along with
different methods. The most trivial one is based on active RFID tags [28], but there are numerous
solutions based on passive ones. However, these systems do not provide the desired accuracy or
require dense infrastructure [29].
Radio Frequency Identification is often considered as one of the core technologies of Internet of
Things paradigm, especially regarding industrial usage. It is commonly used in logistical task such as
identifying products, tracking items, tools and pieces of equipment (not location, but their presence).
Most of these applications aimed to facilitate RFID based management systems that synchronize with
company’s ERP system to keep the inventory up-to-date [30–32]. These systems can highly improve
the efficiency of logistics by eliminating manual updates; however, their scope is limited and they can
be used only in environments where only few special locations (e.g., workstations, gates) are equipped
with identifying capabilities, but not in dynamically changing ones.
especially where humans are involved in the scenario. These systems are often aided with different
algorithms such as Kalman-Filter to provide a better accuracy [40], however in this case their response
time becomes higher, therefore real-time systems can not be implemented this way. Some low-powered
solutions apply custom approaches to achieve energy-efficiency and to keep the costs down low,
but their characteristics are very similar to BLE based systems—mostly regarding to accuracy and
precision as well as the required density of readers [41,42].
UWB based solutions are usually coupled with RFID or UHF-RFID technology, since the first one
provides high accuracy with a short response time and the second one is a relatively cheap technology
that can enable identification or large number of assets—while some applications use both of them
to localization [14,43]. Most of these asset tracking solutions based on that approach, where tags are
capable of communicateing via UWB as well as RFID [44–47], so one device can serve both purposes.
While these systems are considered the state-of-the-art for on-site deployments, their scalability and
generic applicability is limited when it comes to economical considerations.
The main disadvantage of the previously described only-UWB and hybrid UWB/RFID approaches
is that each asset has to be equipped a tag, so a large number of assets can highly increase the cost of
a solution, while other solutions that rely upon only RFID technologies are more scalable in the manner
of cost, since the price of an RFID tag is 10 cent or less. In this case static readers can identify assets and
determine their locations, although it requires plenty of readers which prices can be really expensive
depending on the used frequency (HF or UHF) [48].
Hence, a research gap for a scalable, economical asset and production tracking solution needs to
be covered for wider applicability of Industry 4.0 results in small-sized and mid-sized stakeholders of
the value chain.
There are numerous RTLS and tracking system on the market right now with different capabilities
and costs to serve different purposes. In this paper we introduce a different tracking approach which
unifies the advantages of the aforementioned approaches. In this case the assets are tracked indirectly,
therefore the required infrastructure do not have to scale with the number assets. This means that
the cost of the system could be as low as possible to make it an Industry 4.0 compliant asset tracking
solution that can be applied at minor companies too as a tool of digitization across the value chain.
Hence local clouds can be defined for the System-of-Systems that communicate on the manufacturing
floor, others on the warehouse, yet another ones on the enterprise planning level, or in the logistics
network, and so forth. Service producers and consumers are orchestrated together within the local
clouds with the help of mandatory core systems defined by the Arrowhead framework—such as Service
Registry, Orchestrator, or the Authorizer system, as shown by Figure 2. Furthermore, the inter-cloud
servicing is introduced as a novelty, and also supported [52] practically as well. Inter-cloud servicing
is a necessary concept for flexible but secure information exchange between the systems of different
stakeholders—so service consumers and producers can connect even if they reside in different local
clouds, as Figure 3 shows.
Figure 2. The core elements of the Arrowhead Framework are connected in a service-oriented way.
Figure 3. Local clouds with serving different industrial domains have their internal requirements (such
as latency or security) guaranteed—and they can produce and consume services to/from other Local
clouds as well. [51]
The high-level interoperability concept for the entire supply chain—that is supported by
Arrowhead—is visualized by Figure 4 [7].
Sensors 2020, 20, 3709 8 of 24
Figure 4. System-of-Systems established throughout the supply chain network can exchange
information in a service oriented architecture.[7]
Figure 5. An example scenario for asset tracking with integrated UWB-based indoor-positioning and
Ultra High Radio Frequency Identification (UHF-RFID)-based asset identification.
Since anchors are the reference points, their location is fixed: they are usually deployed on walls
or surfaces of static objects, meanwhile UWB tags represent the vehicles or tools that can replace
assets to be followed which we can call mules in general. In this case the only restriction regarding
Sensors 2020, 20, 3709 9 of 24
to the placement of the devices within the tracked object is: it must not be placed in shielded area
(e.g., under metal surfaces).
The workflow of the system relies on the tight coupling of localization and identification.
The localization method follows the general scheme where the relative position of an UWB tag
can be calculated from the measured distances between the UWB tag and the nearby anchor points.
This relative position can be converted into an absolute position if the locations of anchors is known
in the tracked area. The calculation and conversion are processed by the so-called Core System, which
also handles the trigger events and command messages.
The identification scheme is basically trigger-based and event-driven; in other words: the system
identifies an asset only if some specific conditions are met. Events are changes in the physical world
such as when a mule starts or stops of its motion. These events can be detected by various sensors that
generates trigger signals to command the RFID reader on the mule to read nearby RFID tags. The set of
events that can be used as triggers is limited and highly depends on the use-case, but generally picking
assets up and putting them down are the core events that can be possibly found in any circumstances.
Full traceability of assets is provided by the unified localization and identification workflow.
The IPS subsystem tracks each mule and object that can relocate products or assets. When an asset has
been picked up, it triggers the identification subsystem which identifies the product or products and
commands the Core system to attach the read RFID tags virtually to the tracked mules. Similarly, when
an asset has been removed from a mule, the identification subsystem commands the Core system to
detach the related RFID tags from the tracked object. This scheme is an indirect form of asset tracking,
since we do not follow the asset itself, but the mules that can move them.
Within this scheme, movement of assets can be traced even if they are not tracked directly;
however, the corresponding vehicles and tools are. We call this scheme indirect asset tracking, since
we do not follow the asset itself, only the mules that can move them. This approach enables our
solution to be inexpensive and easy to deploy by avoiding the usage of more complex methods
and infrastructure regarding asset identification, because the number of assets does not really have
an impact on the size of the infrastructure. Moreover, due to this indirect scheme the system is scalable
in a wide range, and expandable as well, therefore it fits well into medium-, and small-sized companies’
workflow. An example for this workflow applied in a warehouse scenario is shown by Figure 5.
charging periods, which is also not desirable from logistical perspective. Note: Peak voltage reduction
of battery from 14 V to 10 V is included.
Trigger-based identification can be implemented in many different ways—this paper focuses
on geo-fencing and motion based triggers. The term geo-fencing covers defining virtual perimeters
for real-word geographic areas to enable the usage of location-based services. Using geo-fencing is
beneficial in those environments where boundaries of staging, storing, loading or other special areas
are well-defined [55]. In this case, these areas can be fenced virtually and if an AT device enters or
leaves a pre-defined, geo-fenced area, the RFID reader will be triggered to scan which assets entered
or left the area [56]. This method can be useful when geo-fenced areas cover the full map, although
it provides less accurate asset location data.
Motion-based methods refer to schemes when events related to the movement of AT devices
(e.g., start and stop) trigger identification. The main difference between these schemes is the source of
trigger that is, how these events are detected. One approach is the position based detection scheme,
where the source of trigger is our existing infrastructure—the movement of AT devices can be estimated
based on the real-time location data-streams. The other approach includes the usage of external trigger
sources such as separate modules or attached sensors—for example, Inertial Measurement Unit
(IMU) [57]. These units provide information about an object’s special force (acceleration) and angular
rate, so it can be applied as a supplementary unit for tracking [58], but this technique is based on sensor
fusion which is computation heavy process. Therefore, in this paper we use a different approach,
since most of these units can also detect and sign specific event like start and stop of movement.
These sensors are really sensitive to even minor movements hence they detect motions that should
not be tracked (e.g., small movements during loading), so we have to eliminate these false-detections
in our identification scheme.
• Geo-fencing based trigger generation: It can be detected if an AT device enters or leaves an area.
– Advantages: Out-of-the-box method (no further costs), provides really accurate asset
location;
– Disadvantages: false detection due to start/stop events while mule is loaded, false detection
due to ranging error.
• External source based trigger generation: Based on the external sensor such as IMU-s, it can be
determined if an AT devices is moving or not.
– Disadvantages: false detection due to high sensitivity and start/stop events while mule
is loaded.
Yes
Has Core-system No
detected motion?
Yes
Yes
It is worth noting that non-event driven approaches—for example, periodic polling—can be used
as supplementary methods. Periodic polling is also a great way to eliminate false detection when
geo-fencing is not an option, because assets can be unloaded anywhere in the warehouse. In this case
the Received Signal Strength Indication (RSSI) values are used to determine the distance of certain
Sensors 2020, 20, 3709 12 of 24
RFID tags from the reader. This can be applied as a filter during movement to avoid false detection:
if distances of RFID tags are the same before stop and after start, then RFID tags should not be detached
from the AT device.
4. Measurements
4.1. Overview
To achieve the full potential of our unified identification scheme, we have to refine the position
based method, in particular, what type and amount of motion is considered as moving. If an AT device
moves 50 cm back and forth during loading, for instance, it does not count as moving.
A major issue here is false detection due to ranging error—this occurs if the absolute distance
error—the distance between the physical position and the estimated one—is greater than the value
that we defined as the act of moving previously.
This approach requires to measure the limits and characteristics of the presented solution to see
if the application is capable of using the unified method without any modification. Each measurement
was performed indoors with a static layout that is shown in Figure 7.
5m
Figure 7. A schematic map of the test environment for measurements, where purple thumbtacks
are anchors.
80
70
μ = 8.2194, σ = 80.0421
60
50
Frequency
40
30
20
10
0
00
50
00
50
00
50
00
0
0
50
0
0
0
0
0
0
−5
10
15
20
25
30
35
−4
−3
−3
−2
−2
−1
−1
Error [mm]
Figure 8. Statistical characteristics of UWB ranging.
UWB samples
12250
12000
11750
X [mm]
11500
11250
11000
10750
11500 11750 12000 12250 12500 12750 13000 13250 13500 13750
Y [mm]
12500
Visualisation of the samples 12500
Moving average #1
12000 12000
X [mm]
X [mm]
11500 11500
11000 11000
10500 10500
11500 12000 12500 13000 13500 11500 12000 12500 13000 13500
Y [mm] Y [mm]
12500
Moving average #2 12500
Moving average #3
12000 12000
X [mm]
X [mm]
10000 10000
8000 8000
6000 8000 10000 12000 14000 6000 8000 10000 12000 14000
Y [mm] Y [mm]
Moving average #2 Moving average #3
14000 14000
12000 12000
X [mm]
X [mm]
10000 10000
8000 8000
6000 8000 10000 12000 14000 6000 8000 10000 12000 14000
Y [mm] Y [mm]
Figure 11. Using moving average with different window sizes on a path tracked by a moving device.
Sensors 2020, 20, 3709 16 of 24
• UWB based positioning subsystem: the forklift is continuously being positioned by its
UWB tracker,
• RFID reader: the forklift has an RFID reader on its front to identify the carried goods,
• IMU: the accelerometer detects the current state of motion.
The test scenario consists of three phases which covers all the possible cases that are handled by
the identification scheme, while being executed within a short time interval and physical path. In the
first phase the forklift arrives at workstation A. At this station there are goods that should get loaded
onto the mule and transferred to workstation B. After stopping at the station A, in the second phase
the forklift follows its schedule and takes the goods to workstation B. At the halfway point of the path,
the forklift stops as it faces temporary obstacles. When the path is free again, the forklift continues
its motion to workstation B. After arrival, in the third phase the transferred goods get unloaded and
the forklift leaves the station. To simulate a realistic scenario there are additional goods both on
workstation A and B. These boxes can be also tracked by the RFID reader of the forklift.
Time Event
14:39:50 Start
14:40:20–14:40:30 Arrival at workstation A
14:41:10–14:41:20 Stopping at the obstacle
14:41:50–14:42:00 Arrival at workstation B
14:42:30 Leaving workstation B
Sensors 2020, 20, 3709 17 of 24
Geofencing
14000
path
13500
X [mm]
13000
12500
12000
4000 4500 5000 5500 6000 6500 7000 7500 8000
Y [mm]
The three phases are also visible in Figure 13—the graph shows the data collected by
the accelerometer. The four significant peaks mark each listed event from the table. In the demo
setup, IMU subsystem is providing a simple one-bit trigger (interrupt) for the system, to reduce
the computational cost to the minimum possible level. On the contrary, UWB ranging is a more
complex operation that uses more resources to provide a more sophisticated output.
IMU
1
Motion detection
0
14:39:50
14:40:00
14:40:10
14:40:20
14:40:30
14:40:40
14:40:50
14:41:00
14:41:10
14:41:20
14:41:30
14:41:40
14:41:50
14:42:00
14:42:10
14:42:20
14:42:30
14:42:40
14:42:50
14:43:00
Time
Figure 13. Inertial Measurement Unit (IMU) triggers the system by setting "1" its output.
E2000017230C01990890C123 E2000017230C01990890C123
30 −40
RSSI [dB]
20
RC
10 −50
0 −60
E2000017230C02170980B737 E2000017230C02170980B737
30 −40
RSSI [dB]
20
RC
10 −50
0 −60
E2000017230C02410980B767 E2000017230C02410980B767
30 −40
RSSI [dB]
20
RC
10 −50
0 −60
E2000017230C02520980B780 E2000017230C02520980B780
30 −40
RSSI [dB]
20
−50
RC
10
0 −60
E2000017230C01640980B6D0 E2000017230C01640980B6D0
30 −40
RSSI [dB]
20
−50
RC
10
0 −60
E2000017230C02400890C16C E2000017230C02400890C16C
30 −40
RSSI [dB]
20
−50
RC
10
0 −60
14:39:50
14:40:00
14:40:10
14:40:20
14:40:30
14:40:40
14:40:50
14:41:00
14:41:10
14:41:20
14:41:30
14:41:40
14:41:50
14:42:00
14:42:10
14:42:20
14:42:30
14:42:40
14:42:50
14:43:00
14:39:50
14:40:00
14:40:10
14:40:20
14:40:30
14:40:40
14:40:50
14:41:00
14:41:10
14:41:20
14:41:30
14:41:40
14:41:50
14:42:00
14:42:10
14:42:20
14:42:30
14:42:40
14:42:50
14:43:00
Time Time
(a) (b)
Figure 14. RFID measurements. (a) The Read Count of the tracked RFID tags. (b) The RSSI value of
the detected RFID tags.
As it was shown in the position data, the RSSI and Read Count also follow the three phases of
the test scenario. The RFID reader starts tracking the matrices when the forklift arrives at workstation
A and two of them are continuously being identified in the next two minutes. When the forklift is
reversing from workstation A, it loses the signal of the remaining two. As the forklift is heading
towards workstation B, the reader identifies 2 more RFID matrices, and keeps on tracking until it leaves
the station.
the five phases: the last column of Table 2 indicates that valid events happened at the two workstations,
while the stop at the obstacle is irrelevant.
Table 2. Summary of the Use-Case Scenario: IMU Triggers UWB Ranging, the RFID Reader Identifies
the Assets in Sight (Numbers in Brackets Mark Measurements, That Ore Not Part of the Identification
Scheme). Current State Stands for the Mode the System Operates in at the Moment, Asset Tracking
Events Describe The Outcome of the Operation.
Time IMU Trigger UWB Ranging Number of RFID Tags Current State Asset Tracking Events
14:39:50 (0) STAND–BY
14:40:00 (0) STAND–BY
14:40:10 (0) STAND–BY
14:40:20 X (3) START
14:40:30 X 2 MOTION new products detected
14:40:40 4 STOP new products detected at A
14:40:50 (4) STAND–BY
14:41:00 (3) STAND–BY
14:41:10 X (3) START
14:41:20 X 2 MOTION products are taken
14:41:30 X 2 MOTION products are taken
14:41:40 2 STOP products arrived at obstacle
14:41:50 X (2) START
14:42:00 X 4 MOTION new products detected
14:42:10 X 4 MOTION new products detected
14:42:20 4 STOP products arrived at B
14:42:30 X (4) START
14:42:40 X 0 MOTION empty forklift
14:42:50 0 STOP empty forklift
14:43:00 (0) STAND–BY
As it is shown, the tracking scheme is reliable since the assets that are relocated are fully tracked
between the two workstations despite of the several intermediate stop and start events. According to
this, the system can serve its purposes as a low-cost tracking solution for certain logistic tasks across
the whole supply chain.
arrangements where merely a set of vehicles or tools can actually move the objects to be tracked,
since each of the mules has to be equipped with tag devices. Other limitations are derived from
the underlying technology stack—the current UHF-RFID readers can handle only approx. 200 RFID
tags simultaneously [54], which maximizes the number of assets that can be relocated together
as a batch. This is a theoretical maximum, however, it highly depends on the actual reader—which may
provide higher rates then the aforementioned one [60]. Additionally, while we do not have to deal with
density issues related to UWB, but the usual limitations of such a wireless technology—shielding metal
surfaces, certain amount of interference—still affect the system. This, however is a common challenge
for all indoor-localization systems. Together with the description of the method and the utilized
technologies and integrated systems, the paper presented a concrete proof of concept scenario as well,
to demonstrate how the indirect asset tracking scheme works and how reliable this concept really is
in a specific use case.
To summarize the evaluation results, the system is an efficient implementation of asset tracking
due to the cooperation of the different subsystems and technologies and the presented identification
scheme. The main advantages of the system are its high scalability, and the possibility of continuous
adaptation. When applying our indirect identification scheme, the number of assets to be tracked is
not bound to the overall infrastructure of the system, thus the number of assets can be increased by
magnitudes without extending the infrastructure itself. Moreover, the architecture itself is modular,
so if there is a need for extending the number of tags (mules) or anchors, it does not require significant
financial effort, since each new device can be added to the system one by one. Similarly to other
state-of-the-art systems, this one supports energy-efficiency as well, by involving inactive periods
during stand-by mode that lower power consumption of the RFID reader, while there is no information
loss, because all of the necessary asset identifications happens in motion (unnecessary measurements
are in brackets in Table 2). Furthermore, forklifts operate as searchlights in the warehouse: they not
only track the goods during moving them between stations, but also discover the remaining supplies
and update their position, as well.
Nevertheless, future work is still ongoing in relation to similar live use-cases, which further
highlight the advantages and eliminate the limitations of the presented solution.
As for wider perspectives, digital twins appear in all areas of production and logistics,
hence the proceedings of digital-twin related research, development and innovation activities are
expected to expand [7]. Associated with digital twins, mass individualization and lot size one paradigms
are reshaping the production logics deep inside the manufacturing process level. Since smart assets
are getting traceable on-the-fly, and their digital twin can contain not only status logs but actual
production recipes for the asset, they can potentially drive autonomous production re-organization
to meet the lot size one requirement on the spot. Such self-organization of manufacturing process for
mass individualization [61] requires high flexibility and interoperability. This can be achieved by
Service Oriented IoT Architectures, based on which the Arrowhead Framework also enables such
workflow choreography [62].
For an even wider view, the "ABCDE5G" technologies—artificial intelligence, blockchains,
cloud computing, big data analytics, edge computing, private 5G campus networks—are fostering
the industrial IoT domain [63]. When it comes to asset traceability, blockchains and related
technologies (such as distributed ledgers and smart contracts) also provide added value to data security
traceability of smart assets. Some blockchain-based industrial models, such as ManuChain [64] and
Makerchain [65] have already been proposed to address these very issues.
Author Contributions: Conceptualization, F.A., G.V. and P.V.; methodology, F.A. and P.V.; software, F.A. and G.V.;
investigation, F.A. and G.V.; resources, P.V.; writing–original draft preparation, F.A. and G.V.; writing–review and
editing, P.V.; visualization, A.F., G.V. and P.V.; supervision, P.V.; funding acquisition, P.V. All authors have read
and agree to the published version of the manuscript.
Sensors 2020, 20, 3709 21 of 24
Funding: This research is supported by the EU ARTEMIS JU funding, under grant agreement #737459 (project
Productive4.0) and from National Research, Development and Innovation Office, Hungary, under the agreement
2018-1.1.1-MKI.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. DIN Standards. Reference Architecture Model Industrie 4.0 (RAMI4.0); DIN SPEC 91345; DIN: Berlin,
Germany, 2016.
2. Kozma, D.; Soos, G.; Varga, P. Supporting Digital Production, Product Lifecycle and Supply Chain
Management in Industry 4.0 by the Arrowhead Framework–A Survey. In Proceedings of the IEEE
International Conference on Industrial Informatics (INDIN), Helsinki, Finaland, 22–25 July 2019.
3. Barreto, L.; Amaral, A.; Pereira, T. Industry 4.0 implications in logistics: An overview. Procedia Manuf.
2017, 13, 1245–1252. [CrossRef]
4. Perez-Guirao.; Kull, B.; Luediger, H. An ultra-wideband approach towards autonomous radio control
and positioning systems in manufacturing & logistics processes. In Proceedings of the 4th Workshop on
Positioning, Navigation and Communication, Hannover, Germany, 22–22 March 2007.
5. Productive 4.0. Available online: https://productive40.eu/about/ (accessed on 30 June 2020).
6. Halme, J.; Jantunen, E.; Hästbacka, D.; Hegedűs, C.; Varga, P.; Björkbom, M.; Mesiä, H.; More, R.; Jaatinen,
A.; Barna, L.; et al. Monitoring of Production Processes and the Condition of the Production Equipment
through the Internet. In Proceedings of the 2019 6th International Conference on Control, Decision and
Information Technologies (CoDIT), Paris, France, 23–26 April 2019; pp.1295–1300.
7. Kozma, D.; Varga, P.; Hegedűs, C. Supply Chain Management and Logistics 4.0 - A Study on Arrowhead
Framework Integration. In Proceedings of the 2019 8th International Conference on Industrial Technology
and Management (ICITM), Cambridge, United Kingdom, 2–4 March 2019; pp. 12–16.
8. Soós, G.; Kozma, D.; Janky, F.N.; Varga, P. IoT Device Lifecycle–A Generic Model and a Use Case for
Cellular Mobile Networks. In Proceedings of the 2018 IEEE 6th International Conference on Future Internet
of Things and Cloud (FiCloud), Barcelona, Spain, 6–8 August 2018; pp.176–183.
9. Wieringa, R.J. Design Science Methodology for Information Systems and Software Engineering; Springer-Verlag:
Berlin/Heidelberg, Germany, 2014.
10. Niemelä, V.; Haapola, J.; Hämäläinen, M.; Iinatti, J. An Ultra Wideband Survey: Global Regulations and
Impulse Radio Research Based on Standards. IEEE Commun. Surv. Tutor. 2017, 19, 874–890. [CrossRef]
11. Dardari, D.; Closas, P.; Djurić, P.M. Indoor Tracking: Theory, Methods, and Technologies. IEEE Trans. Veh.
Technol. 2015, 64, 1263–1278, doi:10.1109/TVT.2015.2403868. [CrossRef]
12. Zafari, F.; Gkelias, A.; Leung, K.K. A Survey of Indoor Localization Systems and Technologies.
IEEE Commun. Surv. Tutor. 2019, 21, 2568–2599. [CrossRef]
13. Lee, J.; Su, Y.; Shen, C. A Comparative Study of Wireless Protocols: Bluetooth, UWB, ZigBee, and Wi-Fi.
In Proceedings of the IECON 2007–33rd Annual Conference of the IEEE Industrial Electronics Society,
Taipei, Taiwan, 5–8 November 2007; pp.46–51, doi:10.1109/IECON.2007.4460126. [CrossRef]
14. Huang, S.; Guo, y.; Zha, S.; Wang, F.; Fang, W. A Real-time Location System Based on RFID and UWB
for Digital Manufacturing Workshop. Procedia CIRP 2017, 63, 132–137, doi:10.1016/j.procir.2017.03.085.
[CrossRef]
15. Kolakowski, J.; Djaja-Josko, V.; Kolakowski, M.; Broczek, K. UWB/BLE Tracking System for Elderly People
Monitoring. Sensors 2020, 20, 1574. [CrossRef]
16. Sadowski, S.; Spachos, P. RSSI-Based Indoor Localization With the Internet of Things. IEEE Access 2018,
6, 30149–30161. [CrossRef]
17. Xu, B.; Sun, G.; Yu, R.; Yang, Z. High-Accuracy TDOA-Based Localization without Time Synchronization.
IEEE Trans. Parallel Distrib. Syst. 2013, 24, 1567–1576. [CrossRef]
18. Son, S.; Kim, B.; Park, H.; Baek, Y. Hierarchical asset tracking system using IEEE 802.15.4a radio in container
terminals. In Proceedings of the 2016 URSI Asia-Pacific Radio Science Conference (URSI AP-RASC), Seoul,
Korea, 21–25 August 2016; pp.276–278.
Sensors 2020, 20, 3709 22 of 24
19. Li, Z.; Braun, T.; Zhao, X.; Zhao, Z.; Hu, F.; Liang, H. A Narrow-Band Indoor Positioning System by
Fusing Time and Received Signal Strength via Ensemble Learning. IEEE Access 2018, 6, 9936–9950,
doi:10.1109/ACCESS.2018.2794337. [CrossRef]
20. Paul, A.S.; Wan, E.A. RSSI-Based Indoor Localization and Tracking Using Sigma-Point Kalman Smoothers.
IEEE J. Sel. Top. Signal Process. 2009, 3, 860–873, doi:10.1109/JSTSP.2009.2032309. [CrossRef]
21. You, W.; Li, F.; Liao, L.; Huang, M. Data Fusion of UWB and IMU Based on Unscented Kalman Filter for
Indoor Localization of Quadrotor UAV. IEEE Access 2020, 8, 64971–64981. [CrossRef]
22. Lu, C.; Uchiyama, H.; Thomas, D.; Shimada, A.; Taniguchi, R.I. Indoor Positioning System Based on
Chest-Mounted IMU. Sensors 2019, 19, 420. [CrossRef] [PubMed]
23. Hu, X.; Luo, Z.J.W. AGV Localization System Based on Ultra-Wideband and Vision Guidance. Electronics
2020, 9, 448. [CrossRef]
24. Delamare, M.; Boutteau, R.; Savatier, X.; Iriart, N. Static and Dynamic Evaluation of an UWB Localization
System for Industrial Applications. Sci 2020, 1, 62. [CrossRef]
25. Want, R. An introduction to RFID technology. IEEE Pervasive Comput. 2006, 5, 25–33. [CrossRef]
26. Ustundag, A.; Kilinc, S.; Kabadurmus, O. Evaluation of Operational Parameters Affecting Bulk Reading
Performance of UHF RFID System. In Proceedings of the 2007 1st Annual RFID Eurasia, Istanbul, Turkey,
5–6 September 2007; pp.1–4.
27. Gan, O.P.; Aw, L.L.; Sheng, H. Reliable RFID bulk reading using adaptive time and power control.
In Proceedings of the 2016 IEEE Region 10 Conference (TENCON), Singapore, 22–25 November 2016;
pp.130–134. doi:10.1109/TENCON.2016.7847974. [CrossRef]
28. Zhang, D.; Yang, L.T.; Chen, M.; Zhao, S.; Guo, M.; Zhang, Y. Real-Time Locating Systems Using Active
RFID for Internet of Things. IEEE Syst. J. 2016, 10, 1226–1235, doi:10.1109/JSYST.2014.2346625. [CrossRef]
29. Scherhäufl, M.; Pichler, M.; Müller, D.; Ziroff, A.; Stelzer, A. Phase-of-arrival-based localization of passive
UHF RFID tags. In Proceedings of the 2013 IEEE MTT-S International Microwave Symposium Digest
(MTT), Seattle, WA, USA, 2–7 June 2013; pp. 1–3, doi:10.1109/MWSYM.2013.6697558. [CrossRef]
30. Lian, X.; Zhang, X.; Weng, Y.; Duan, Z. Warehouse Logistics Control and Management System Based on
RFID. 2007 IEEE International Conference on Automation and Logistics, Jihan, China, 18–21 August 2007.
31. Jing, X.; Tang, P. Research and Design of the Intelligent Inventory Management System Based on RFID.
In Proceedings of the 2013 Sixth International Symposium on Computational Intelligence and Design,
Hangzhou, China, 28–29 October 2013.
32. Liu, Z.; Yi, J.; Chen, C.; Gu, C.; Fan, T. Development of a management system of laboratory equipment
based on UHF RFID. In Proceedings of the 2nd International Conference on Information Science and
Engineering, Hangzhou, China, 28–29 October 2013.
33. Decawave Limited DWM1000. Available online: https://www.digikey.com/product-detail/en/decawave-
limited/DWM1000/1479-1002-1-ND/4805335 (accessed on 30 June 2020).
34. Website of Wirlesstag Store. Available online: https://store.wirelesstag.net/collections/all (accessed on
30 June 2020).
35. Website of Estimote Products. Available online: https://estimote.com/products/ (accessed on 30 June 2020).
36. Website of Gimbal Store. Available online: https://store.gimbal.com/ (accessed on 30 June 2020).
37. Kotankt IO’s Website. Available online: https://store.kontakt.io/ (accessed on 30 June 2020).
38. Rosa, R.L.; Dehollain, C.; Livreri, P. Advanced Monitoring Systems Based on Battery-Less Asset Tracking
Modules Energized through RF Wireless Power Transfer. Sensors 2020, 20, 3020. [CrossRef]
39. Chen, X.; Zhu, Y.; Li, J.; Wen, Y.; Gong, Z. Efficiency and Privacy Enhancement for a Track and Trace
System of RFID-Based Supply Chains. Information 2015, 6, 258–274. [CrossRef]
40. Lee, C.K.M.; Ip, C.M.; Park, T.; Chung, S.Y. A Bluetooth Location-based Indoor Positioning System for
Asset Tracking in Warehouse. In Proceedings of the 2019 IEEE International Conference on Industrial
Engineering and Engineering Management (IEEM), Macao, Macao, 15–18 December 2019; pp.1408–1412.
41. Robert, J.; Lindner, T.; Milosiu, H. Sub 10µW wake-up-receiver based indoor/outdoor asset tracking
system. In Proceedings of the 2015 IEEE 20th Conference on Emerging Technologies Factory Automation
(ETFA), Luxembourg, 8–11 September 2015; pp.1–3.
42. Tegou, T.; Kalamaras, I.; Tsipouras, M.; Giannakeas, N.; Votis, K.; Tzovaras, D. A Low-Cost Indoor Activity
Monitoring System for Detecting Frailty in Older Adults. Sensors 2019, 19, 452. [CrossRef]
Sensors 2020, 20, 3709 23 of 24
43. Ramos, A.; Guillén, A.; Villarino, R.; Girbau, D. Time-domain UWB RFID tags for smart floor applications.
In Proceedings of the 2014 IEEE RFID Technology and Applications Conference, Tampere, Finland,
8–9 September 2014; doi:10.1109/RFID-TA.2014.6934221. [CrossRef]
44. Cruz, C.C.; Costa, J.R.; Fernandes, C.A. Hybrid UHF/UWB Antenna for Passive Indoor Identification and
Localization Systems. IEEE Trans. Antennas Propag. 2013, 61, 354–361. [CrossRef]
45. Vizziello, A.; Savazzi, P. Efficient RFID Tag Identification Exploiting Hybrid UHF-UWB Tags and
Compressive Sensing. IEEE Sens. J. 2016, 16, 4932–4939. [CrossRef]
46. Decarli, N.; Guidi, F.; Dardari, D. Passive UWB RFID for Tag Localization: Architectures and Design.
IEEE Sens. J. 2016, 16, 1385–1397. [CrossRef]
47. D’Errico, R.; Bottazzi, M.; Natali, F.; Savioli, E.; Bartoletti, S.; Conti, A.; Dardari, D.; Decarli, N.;
Guidi, F.; Dehmas, F.; et al. An UWB-UHF semi-passive RFID System for localization and tracking
applications. In Proceedings of the 2012 IEEE International Conference on RFID-Technologies and
Applications (RFID-TA), Nice, France, 5–7 November 2012; pp.18–23.
48. Website of AtlasRFIDStore. Available online: https://www.atlasrfidstore.com/rfid-readers/ (accessed on
30 June 2020).
49. Delsing, J.; Varga, P.; Ferreira, L.L.; Albano, M.; Pereira, P.P.; Eliasson, J.; Carlsson, O.; Derhamy, H.
The Arrowhead Framework architecture: Arrowhead Framework. In IoT Automation: Arrowhead Framework;
Delsing, J., Ed.; CRC Press: Cleveland, OH, USA, 2017; Chapter 3, pp. 44–89.
50. Varga, P.; Blomstedt, F.; Ferreira, L.L.; Eliasson, J.; Johansson, M.; Delsing, J.; De Soria, I.M. Making system
of systems interoperable–The core components of the arrowhead framework. J. Netw. Comput. Appl. 2017,
81, 85–95, doi:10.1016/j.jnca.2016.08.028. [CrossRef]
51. Delsing, J.; Eliasson, J.; van Deventer, J.; Derhamy, H.; Varga, P. Enabling IoT automation using local
clouds. In Proceedings of the 2016 IEEE 3rd World Forum on Internet of Things (WF-IoT), Reston, VA,
USA, 12–14 December 2016; pp.502–507.
52. Hegedűs, C.; Kozma, D.; Soós, G.; Varga, P. Enhancements of the Arrowhead Framework to Refine
Intercloud Service Interactions. In Proceedings of the 42nd Annual Conference of IEEE Industrial
Electronics Society (IECON), Florence, Italy, 23–26 October 2016.
53. Hegedűs, C.; Frankó, A.; Varga, P. Asset and Production Tracking through Value Chains for Industry 4.0
using the Arrowhead Framework. In Proceedings of the 2019 IEEE Industrial Cyber-Physical Systems
(ICPS), Taipei, Taiwan, 6–9 May 2019.
54. ThingMagic Nano (UHF) RAIN RFID Desgin Guide. Available online: https://www.jadaktech.com/
products/rfid/embedded-uhf-rfid-readers/nano/ (accessed on 30 June 2020).
55. Modica, P.W.; Loria, M.P.; Toja, M.; Carchiolo, V.; Malgeri, M. A Geofencing Algorithm Fit for Supply Chain
Management. In Proceedings of the 2018 Federated Conference on Computer Science and Information
Systems (FedCSIS), Poznan, Poland, 9–12 September 2018; pp.737–746.
56. Oliveira, R.R.; Cardoso, I.M.; Barbosa, J.L.; A., C.; Prado, M.P. An intelligent model for logistics management
based on geofencing algorithms and RFID technology. Expert Syst. Appl. 2015, 42, 6082–6097. [CrossRef]
57. Yao, L.; Wu, Y.A.; Yao, L.; Liao, Z.Z. An integrated IMU and UWB sensor based indoor positioning system.
In Proceedings of the 2017 International Conference on Indoor Positioning and Indoor Navigation (IPIN),
Sapporo, Japan, 18–21 September 2017; pp.1–8.
58. Corrales, J.A.; Candelas, F.A.; Torres, F. Hybrid tracking of human operators using IMU/UWB data fusion
by a Kalman filter. In Proceedings of the 2008 3rd ACM/IEEE International Conference on Human-Robot
Interaction (HRI), Amsterdam, The Netherlands, 12–15 March 2008; pp.193–200.
59. Djaja-Josko, V.; Kolakowski, J. Application of Kalman Filter for positioning precision improvement in UWB
localization system. In Proceedings of the 24th Telecommunications Forum (TELFOR), Belgrade, Serbia,
22–23 November 2016.
60. ThingMagic M6E Micro (UHF) UHF RAIN RFID Description. Available online: https://www.jadaktech.
com/products/thingmagic-rfid/thingmagic-m6e-micro-uhf-rain-rfid-series/ (accessed on 30 June 2020)
61. Ding, K.; Chan, F.T.; Zhang, X.; Zhou, G.; Zhang, F. Defining a Digital Twin-based Cyber-Physical
Production System for autonomous manufacturing in smart shop floors. Int. J. Prod. Res. 2019,
57, 6315–6334, doi:10.1080/00207543.2019.1566661. [CrossRef]
Sensors 2020, 20, 3709 24 of 24
62. Varga, P.; Kozma, D.; Hegedus, C. Data-Driven Workflow Execution in Service Oriented IoT Architectures.
In Proceedings of the 2018 IEEE 23rd International Conference on Emerging Technologies and Factory
Automation (ETFA), Turin, Italy, 4–7 September 2018; pp.203–210.
63. Varga, P.; Peto, J.; Franko, A.; Balla, D.; Haja, D.; Janky, F.; Soos, G.; Ficzere, D.; Maliosz, M.; Toka, L. 5G
support for Industrial IoT Applications–Challenges, Solutions, and Research gaps. Sensors 2020, 20, 828,
doi:10.3390/s20030828. [CrossRef] [PubMed]
64. Leng, J.; Yan, D.; Liu, Q.; Xu, K.; Zhao, J.L.; Shi, R.; Wei, L.; Zhang, D.; Chen, X. ManuChain:
Combining Permissioned Blockchain With a Holistic Optimization Model as Bi-Level Intelligence for
Smart Manufacturing. IEEE Trans. Syst. Man Cybern. Syst. 2020, 50, 182–192. [CrossRef]
65. Leng, J.; Jiang, P.; Xu, K.; Liu, Q.; Zhao, J.L.; Bian, Y.; Shi, R. Makerchain: A blockchain with
chemical signature for self-organizing process in social manufacturing. Cleaner Prod. 2019, 234, 767–778,
doi:10.1016/j.jclepro.2019.06.265. [CrossRef]
c 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).