<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Archiving and Interchange DTD v1.0 20120330//EN" "JATS-archivearticle1.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink">
  <front>
    <journal-meta />
    <article-meta>
      <title-group>
        <article-title>Blockchain and IoTA Systems for Automotive sector: A Comparative Analysis</article-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author">
          <string-name>Lorenza Cotugno</string-name>
          <email>lorenza.cotugno@uniroma2.it</email>
          <xref ref-type="aff" rid="aff0">0</xref>
          <xref ref-type="aff" rid="aff2">2</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Simona Franci</string-name>
          <email>simona.franci@uniroma2.it</email>
          <xref ref-type="aff" rid="aff0">0</xref>
          <xref ref-type="aff" rid="aff2">2</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Giordano Galli</string-name>
          <email>giordano.galli@uniroma2.it</email>
          <xref ref-type="aff" rid="aff0">0</xref>
          <xref ref-type="aff" rid="aff2">2</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Alessandro Vizzarri</string-name>
          <email>alessandro.vizzarri@uniroma2.it</email>
          <xref ref-type="aff" rid="aff0">0</xref>
          <xref ref-type="aff" rid="aff2">2</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Giuseppe Perrone</string-name>
          <email>giuseppe.perrone@it.ey.com</email>
          <xref ref-type="aff" rid="aff1">1</xref>
          <xref ref-type="aff" rid="aff2">2</xref>
        </contrib>
        <aff id="aff0">
          <label>0</label>
          <institution>Department of Enterprise Engineering "M. Lucertini", University of Rome Tor Vergata</institution>
          ,
          <addr-line>Rome</addr-line>
          ,
          <country country="IT">Italy</country>
        </aff>
        <aff id="aff1">
          <label>1</label>
          <institution>EY Italy S.p.A</institution>
        </aff>
        <aff id="aff2">
          <label>2</label>
          <institution>ICYRIME 2020: International Conference for Young Researchers in Informatics, Mathematics, and Engineering</institution>
          ,
          <addr-line>Online</addr-line>
        </aff>
      </contrib-group>
      <fpage>9</fpage>
      <lpage>15</lpage>
      <abstract>
        <p>In the last years the automotive sectors are afected by the efect of the innovation digital technologies in several vertical applications. In this ecosystem, the automotive sector is also impacted by the adaption of innovative technologies as blockchain and IoTA as decentralized network systems. In the first part of the paper the authors provide a first analysis of the two systems. In the second, a comparison analysis is presented and commented. tions, e.g. micropayments and more. The authors present the overview of the blockchain In the last years the digital technologies have been and IoTA systems in the context of the “Blockchain the center of the innovation in diferent sectors, as and IoT Lab”, a joint research laboratory between The health [1, 2], automotive [3], space [4] or energy [5, 6, University of Rome Tor Vergata and EY Italy Spa. 7]. To provide connectivity to persons and to smart ob- The Section 2 the application of the blockchain sysjects worldwide, some telecommunication architectu- tem in the automotive sector. In the section 3 the smart res have been proposed in the literature, including fixed contract are described. In the Section 4 the IoTA sysaccess, 5G and ultra-dense wireless networks and satel- tem and its application to the automotive sector are delite [8], including the IoT communication technologies scribed. The Section 5 provides a comparison analysis as wireless sensor networks and Bluetooth [9, 10, 11, of the blockchain and IoTA systems. In the Section 6 12]. In this ecosystem, the blockchain systems have the main conclusions are summarized. gained the interest of the scientific communities since they are able to managed the data in a secure way and anti-tampered way [13]. This happens thanks to the 2. Blockchain System for adoption of a decentralized network topology, asym- Automotive metric cryptography schemes and digital signature techniques. In the case of automotive sector, the possibil- The new mobility models that are being studied and ity to connect the on-board sensors and the Electronic implemented involve an increasing use of data and inControl Unit (ECU) is very important in order to col- formation, leading these systems, which can be delect data with high levels of integrity. ifned as smart, to become increasingly complex and The Tangle is the technology heart of the IOTA, a critical. The automotive sector, such as construction, permissionless, feeless and multi-dimensional distribu- insurance and entertainment, has realized that, by now, ted ledger [14]. the concept of ownership has become completely “liqThe IoTA system is based on the Tangle. It is a tech- uid” and there is no longer interest in the product “car” nology that enables Internet of Things (IoT) innova- but in the mobility service. However, this new paradigm tion through innovative machine-to-machine applica- implicitly provides some problems:</p>
      </abstract>
      <kwd-group>
        <kwd>eol&gt;5G</kwd>
        <kwd>Wireless networks</kwd>
        <kwd>Blockchain</kwd>
        <kwd>IoTA</kwd>
        <kwd>Automotive</kwd>
        <kwd>Smart contract</kwd>
      </kwd-group>
    </article-meta>
  </front>
  <body>
    <sec id="sec-1">
      <title>1. Introduction</title>
      <p>© 2020 Copyright for this paper by its authors. Use permitted under Creative
CPWrEooUrckReshdoinpgs IhStpN:/c1e6u1r3-w-0s.o7r3g CCoEmUmoRns WLiceonrsekAsthtriobuptioPnr4o.0cIneteerdnaitniognasl ((CCC EBYU4R.0)-.WS.org)
• Electronics already for several years, take
control in the car system. Each car, even in the
cheapest models, provides an on-board computer that
constantly monitors the "health status" of the
vehicle, in the most advanced models this
becomes a collector for all the latest safety
sensors and must be sure that everything is working</p>
      <sec id="sec-1-1">
        <title>A team of researchers from the University of Not</title>
        <p>tingham, Ningbo in collaboration with China’s
iCarForce is developing a smart contract-based (automated
software contracts installed on blockchain) insurance
system that involves the use of sensors on board the
car and a driver’s remuneration system based on its
driving mode, a dynamic pricing that does not remain
constant but which can decrease and increase
depending on the use that is actually made of the car [1].</p>
        <p>Instead, there are many proposals for a
blockchainifne; based supply chain management system, a unique
system that certifies, mathematically, all transactions in
• The software, therefore, that manages the elec- the supply chain, from receiving demand, to taking
tronics described above, must be updated con- charge from the warehouse, following all logistical
transstantly paying attention to bugs that could un- actions up to receiving and installing, ensuring the
cusdermine the operation of the car itself, especially tomer that the work is done well. The same approach
in critical functions such as safety. One of the is given to raw material, this ensure that all the supply
issues experiencing now is managing software chain from mine to factory is guaranteed and
traceupdates, knowing which release is installed on able, BMW is creating a blockchain-based platform to
a specific car can be a problem; do so.
• If the new paradigm does not involve the pur- A team from Cornell University research into the
chase of the car, the manufacturers will have all topic of software update and release monitoring,
proposthe interest to keep the vehicle in the best possi- ing a blockchain-based system that would ensure the
ble way to ensure that it still has market finished traceability of software releases on individual vehicles,
the rental ofer. Thanks to connection services their updating and the security of an immutable
systhe car company constantly monitors the status tem that therefore would not allow attackers to
operof the car remotely and optimizes maintenance. ate vehicle compromise activities.</p>
        <p>However, the customer is still not in any way Blockchain can also resolve the problem of payment
guaranteed that maintenance is done properly in e-charging stations using traceable transaction in a
and with quality parts, unless personally check- unique platform that give the customer the freedom of
ing the installed parts; payment old fuel stations give [3].</p>
        <p>However, the typical features of blockchain do not
• Data collected remotely on cars many times is guarantee useful transaction recording times for IoT
redundant and falsified; applications, such as the connected cars we’ve talked
about so far. For this reason, another type of DLT has
• The e-station for charging vehicles are operated developed over time that still guarantees the
characby diferent providers means diferent account teristics of the blockchain but adds in its favor rapid
to recharge the car. exchanges of transactions even on high volumes, the
TANGLE.</p>
        <p>DLT and specifically blockchain fit comfortably into
this system becoming an essential tool for this
constantly and rapidly evolving industry. In fact, the block- 3. Design of Blockchain Smart
chain ensures that the data entered in the ledger, the
main transaction’s database is, as well as safe and trace- Contract
able, also immutable, ensuring that critical
transactions such as those of software installation on the car
or the tracking of certified spare parts or the
management of litigation by accident are managed, tracked
and archived in a simple and searchable way. Each
transaction to be modified requires confirmation of
more than half of the participants in the blockchain itself,
making it perfectly transparent to change attacks.</p>
      </sec>
      <sec id="sec-1-2">
        <title>As per the perfect definition by Andreas M. Antonopou</title>
        <p>los and Gavin Wood, “smart contracts are immutable
computer programs that run deterministically in the
context of an Ethereum Virtual Machine as part of the
Ethereum network protocol. The word contract has
no legal meaning because the contracts are so defined
only because they can control valuable things like ETH
or other digital assets. Are immutable because once
4. IOTA for Automotive
written and implemented, the smart contract cannot 4.1. Description of IOTA System
be modified and requires another smart contract to
modify its system logic. Runs deterministically be- IOTA was founded in 2015 by David Sønstebø, Sergey
cause implements a series of if-than-else that allow Ivancheglo, Dominik Schiener, and Dr. Serguei Popov.
it to function in a deterministic way upon the occur- IOTA refers to the ecosystem which has two main
comrence of a specific event. Smart contracts are used ponents:
whenever there is a need to automate transactions that
require unnecessary human intervention for example • The Tangle represents the architecture of the
disare heavily used in logistics. We can summarize the tributed ledger,
strengths and drastic changes that the use of this tech- • IOTA cryptocurrency.
nology will bring:</p>
        <p>Smart contracts can be used in diferent ways and The project was born specifically to create a DLT
diferent fields: trade finance, property ownership, in- that could play the role of backbone for data from the
surance, peer to peer transaction and a lot others, chan- Internet of Things (IoT) world.
ging deeply the concept of transaction. Experts estimate that by 2020 30 billion IoT devices
will be active, distributed in every imaginable area.</p>
        <p>The problems arising from the need to manage the
large amount of data generated by IoT devices could
be solved by DLT technologies. In this scenario IOTA
is one of the projects that aims to made this come true.</p>
        <p>One of the principal goals of IOTA is the
Machine-toMachine (M2M) Payments, which consists of an
economy managed entirely autonomously by machines. To
achieve success some limits ofered by the classic block- Figure 5: Iota transactions.
chain architectures have been overcome: several
innovations were introduced, and several changes were
made to the algorithms normally used in previous DLTs. faster the consent mechanism and the transaction
con</p>
        <p>The Tangle is a data structure created by IOTA as a firmation is put in place [15].
basis for cryptocurrency. A tangle is a special type of
distributed ledger architecture based on a direct acyclic 4.2. Transactions in IOTA
graph (DAG). A DAG is a graph in which each node is
connected to other nodes via direct connections and The transaction is the basic data structure of any DLT,
there are no cycles. In a tangle there is no concept of it describes the transfer of tokens from one address to
blocking, of chains or of mining. In order to be able another. Transactions are logged in the DLT database
to execute a transaction, each user must first validate and are visible in plaintext within the log.
two other transactions, efectively becoming a miner. To create a new transaction on the network, a node</p>
        <p>The imposed rule serves to ensure the functioning does the following steps:
of the network, since there is no incentive mechanism
that remunerates the miners to approve the transac- • prepare the transaction
tions. • select two more transactions</p>
        <p>Unlike blockchain technologies, IOTA does not
create a sequence of synchronized static blocks (each con- • sign the transaction
taining a number of transactions) but a single
transaction can be associated with itself and in parallel with • calculate Proof of Work
other transactions. • transmit across the network.</p>
        <p>The IOTA system makes each node part of the
validation process, efectively eliminating any scalability When a user tries to execute a transaction, the
translimit: as more transactions are created, as more trans- action will be first prepared. This contains the filling
actions are validated. The result is a self-sustainable in most of the fields of the transaction subject, such
and infinitely scalable system (there is no theoretical as the amount of transactions, the recipient’s address.
tps limit). The next step is to select two more transactions. The</p>
        <p>In the Bitcoin blockchain (but also in its derivatives) new transaction will validate these two existing
transthere is a balance of the dificulty of the PoW in order actions. The transaction can now be signed using the
to limit the creation of a block in an average time of user’s private key. Before the transaction can be
transten minutes. In IOTA, the absence of blocks implies mitted over the network, the PoW must be calculated.
that the waiting time for the validation of transactions After the transaction has been transmitted, the user
is very low. Being the consensus parallelized one has will have to wait for other transactions to select this
the ability to scale infinitely with the increase in the transaction for validation. Once enough other
transnumber of nodes. The more nodes they belong to, the actions directly or indirectly approve the transaction,
it will get the “confirmed” status. Usually, when 95% z and y. From the calculation of the   and  
probof new transactions approve a given transaction, this abilities (considering  = 1 for simplicity) we obtain:
will be enough to have the status “confirmed”.</p>
      </sec>
      <sec id="sec-1-3">
        <title>As we can see the probability of going towards y is negligible. This is because z is already approved by many transactions and its cumulative weight is high, while the branch of y has a much lower weight.</title>
      </sec>
      <sec id="sec-1-4">
        <title>The term "consensus" refers to the process that leads</title>
        <p>all nodes to agree on the same state as the DLT. In
IOTA the entire network is responsible for the
consensus, as opposed to what happens in Bitcoin, where
only miners have the right to decide. 4.4. Coordinator
suaCnocnespernoscuessisnaInOdTiAtisisoinnetroifntshiceimn athinefteraatnusraecstioofnthisis- jIenctthteo enaurlmyesrtoaugsesatotfacthkes. TAanugslee,rtwhehonectownotrrkolsismsuobs-t
ireaaaodpDsetcccfuoLectttdIrraiiirTtttieooollni.imylfinnnnosgewnsoesbidaaartsteyhhisliaidsunegartsaeortdthiahetfdiathettrsasrtherettetaeocoehocwntnfehnblacoyeaeemwitionlbgtlceyopgryhoitaeueghtinantrntholaofirastldtmstfaptiieoociorapepfftenrtditscthPoo.rht(wvanarlweoTytnehaotdhhsaporefa)eeparpn-.cronprTitecinforisuth-otoavdWhmeincveisertesowupievtoralcliieaaknutnptiltlgsisoidyapdvhta.tnieaherhtrArdteeeefpewcroidttastsrtrrelrurmyaapdiaagmnnrimen-ohossdsr----t- IftttteoBCmOeihhhrfoioaaseeatTtnoc,snthAoi.CrIyibeaOdinoBlncftiheTonotryIiaauAwaOvtsncnpotehhTrosedfleieayAreoranarcxiut(getoiteCiensaiodcospsanodniuunocoammteiaaswt)esi.tlias.itoltelnryuTraenraeeTceshnicrtksosioioasssseacpnasnupronuefceirudmrtneocsoianood.dgotptdem,enedeEzcaassotphetmvne.ureaaesaoaeTbtsrgetn-wlsyhtevaasptae-iiaatetcncsrowllhckpehusntifotereealbaoanyclslertmnklduercnwspia,oocinaononhtmwicdgthuisthne.heaeatiotesUcretnctsntnfmnaami,ou,ocllndaontilnikheknneorcssdeees--dt,,
ifpnoladrTyitrhsheeecatMlskyea.elreykcotrivoolneChoinafitnthheMetoswencotuertCiiptaysrltooof abtlhegeorresiyftehsrtmeenmicsesdui.nsecIdet ta1hl0le0t%Crao“ncosoainscfirtamiocanetsnioatnrpatplrrueonsvtte”i.tdyTbhiynisittihsaetrhteaetnmrgealaeit.nedreaassohnavwinhgy a
protection from double spending attacks comes from
a right choice of tips. The idea is to put some ran- 4.5. Real use cases for IoTA in
dom walkers in the Tangle area that will be confirmed Automotive
transactions, and then let them move towards the tips
in a pseudo-random way. The tips so reached are those
that are taken to be approved. Along the way, the
choice of the next transaction to move depends on the
cumulative weight of each one. Starting from
anywhere, the set of upcoming transactions is defined by
all transactions that directly refer to the current
transaction. Assume that the walker is on transaction x.</p>
        <p>The next set is indicated with w. A y transaction of
the w set has a   probability to be selected to
continue the path. This probability is given by:</p>
      </sec>
      <sec id="sec-1-5">
        <title>One of the areas in which IOTA is applying a lot is</title>
        <p>precisely that of mobility so much to participate in the
conference of MOBI (mobility open blockchain
initiative) a consortium of companies that deal with
applying blockchain to the new mobility system.</p>
        <p>One of IoTA’s most important partners is Jaguar Land
Rover, for which they have developed a car wallet
system for their I-Pace car that can, in this way, send and
receive payments for physical and digital services and
can communicate with parking meters and tolls. The
wallet allows the user to make micro-transactions in
 − (  −  ) almost instantaneous time and then collect this
infor  = ∑  − (  −  ) (1) wmiatthiotnhefocirtyfuotfuTrerounsdeh.eIinmatdodciotimonm, utnhiecyatceotlolatbhoeruatseedr
The equation indicates that the walker tends to cho- the source of the energy they are using [6].
ose a transaction with the minimum cumulative weight "Smart Wallet" eliminates the need for small coins
gradient between the current one and the next poten- or to sign up for multiple accounts to pay for a variety
tial transaction. of services. Unlike other similar systems and thanks to</p>
        <p>In the figure the walker is located on x and it must its structure, it does not require a commission to
operchoose the next transaction on which to move between ate; over time, transactions will become faster across
the entire network, which is estimated to reach 75
billion connected devices by 2025. Users will be able to
integrate the "Smart Wallet" with conventional
payment methods.</p>
        <p>Another interesting application of the Tangle is the
one described in the previews chapter. IoTA in
collaboration with ElaadNL (knowledge and innovation
centre in the field of smart charging infrastructure in the
Netherlands) created an IOTA-enabled Smart Charger,
so the problem of diferent accounts for diferent
energy providers is no longer an issue.</p>
        <p>When connecting cars to the charging station, these
devices will be able to independently resolve
transactions and related payments. The technique is safe,
reliable, fast and it has a derisory cost [16].</p>
      </sec>
    </sec>
    <sec id="sec-2">
      <title>5. Blockchain and IOTA: a comparison analysis</title>
      <sec id="sec-2-1">
        <title>The features that a blockchain network must have to ensure its use in IoT, can be summarized here: • Scalability: the network must ensure its performance even with millions of connected devices,</title>
        <p>• Applicability, must ensure fast, economic, pri- The cons of this system are undoubtedly the great
vacy and automated transactions, complexity of development by new developers as
opposed to Ethereum that has now become the standard
• Consensus techniques that enable scalability and for industrial applications, plus it does not guarantee
applicability, a certain time (average) of reaching consensus, it can
happen that your transaction remains stalled for a long
• Developability to ensure ease of development of time before being validated. This is why the IoTA
foundevices operating on the network itself, dation is working, together with industrial partners, to
• Interoperability to work with other blockchain improve the structure and the system in general.</p>
        <p>networks without problems.</p>
        <p>In the beginning, blockchain technology was not
born for IoT purposes but, from Nakamoto’s
whitepaper, to build a tax-free, high-speed global money
exchange system. Over time, it was realized that the
characteristics of the blockchain could be applied to
the most diverse applications, especially in the IoT field,
and many realities were born in a short time with
different protocols, consensus algorithms and network
topology, one on all IoTA. The diferences that IoTA
brings to the standard blockchain are many and
specific:
• We are no longer talking about blockchain but
about tangles, an acyclic graph composed of
nodes in which the individual nodes represent the
individual transactions,</p>
      </sec>
    </sec>
    <sec id="sec-3">
      <title>6. Conclusion</title>
      <p>The blockhain and the IoTA Systems are two importat
decentralized network systems very important. Their
implementations in the automotive sector is very
interesting due to the involvement of several on-board
sensors node and the possibility to enhance their
security. In this way, the two systems provide several
features in terms of security, eficiency and trusted
transactions among the nodes on the basis of common and
diferent features analyzed in the paper. The authors
also provide a comparison analysis of the two systems.
The blockchain is mainly characterized by the
decentralized topology, a chain of blocks, the presence of
the miner, their corresponding costs and a transaction
speed of about 6 minutes. The IoTA system is based</p>
    </sec>
    <sec id="sec-4">
      <title>Acknowledgment</title>
      <sec id="sec-4-1">
        <title>The authors would like to thank EY company for supporting the Blockchain and IoT joint lab activities.</title>
        <p>on the tangle, the absence of miners and relative fees
and a transaction speed of about 3 minutes.</p>
        <p>In the next steps, the authors will perform
simulation activities in order to provide numerical results to
be analyzed.</p>
        <p>Clean Electrical Power: Renewable Energy
Resources Impact, ICCEP 2015, 2015, pp. 602–609.
[8] F. Mazzenga, R. Giuliano, F. Vatalaro, Fttc-based
fronthaul for 5g dense/ultra-dense access
network: Performance and costs in realistic
scenarios, Future Internet 9 (2017) 71.
[9] S. L. Ullo, G. Sinha, Advances in smart
environment monitoring systems using iot and sensors,</p>
        <p>Sensors 20 (2020) 3113.
[10] R. Giuliano, G. C. Cardarilli, C. Cesarini,</p>
        <p>L. Di Nunzio, F. Fallucchi, R. Fazzolari,
F. Mazzenga, M. Re, A. Vizzarri, Indoor
localization system based on bluetooth low energy
for museum applications, Electronics 9 (2020)
[1] M. Asif-Ur-Rahman, F. Afsana, M. Mahmud, M. S. 1055.</p>
        <p>Kaiser, M. R. Ahmed, O. Kaiwartya, A. James- [11] C. Napoli, G. Pappalardo, E. Tramontana,
ImTaylor, Toward a heterogeneous mist, fog, proving files availability for bittorrent using a
and cloud-based framework for the internet of difusion model, in: 2014 IEEE 23rd International
healthcare things, IEEE Internet of Things Jour- WETICE Conference, IEEE, 2014, pp. 191–196.
nal 6 (2018) 4049–4062. [12] G. Capizzi, S. Coco, G. Lo Sciuto, C. Napoli, A
[2] G. Capizzi, G. Lo Sciuto, C. Napoli, D. Polap, new iterative fir filter design approach using a
M. Woźniak, Small lung nodules detection based gaussian approximation, IEEE Signal Processing
on fuzzy-logic and probabilistic neural network Letters 25 (2018) 1615–1619.
with bio-inspired reinforcement learning, IEEE [13] A. Pieroni, N. Scarpato, L. Felli, Blockchain and
Transactions on Fuzzy Systems 6 (2020). iot convergence—a systematic survey on
tech[3] J. P. Trovao, Trends in automotive electronics nologies, protocols and security, Applied
Sci[automotive electronics], IEEE Vehicular Tech- ences 10 (2020) 6749.</p>
        <p>nology Magazine 14 (2019) 100–109. [14] IoTA Foundation, Iota developer documentation,
[4] G. C. Cardarilli, L. Di Nunzio, R. Fazzolari, D. Gia- 2020. URL: https://docs.iota.org/.
rdino, M. Matta, M. Re, L. Iess, F. Cialfi, G. De An- [15] Y. Chen, Blockchain in enterprise: An
innogelis, D. Gelfusa, et al., Hardware prototyping vative management scheme utilizing smart
conand validation of a w- dor digital signal proces- tract, in: 2020 9th International Conference
sor, Applied Sciences 9 (2019) 2909. on Industrial Technology and Management
(IC[5] F. Bonanno, G. Capizzi, G. L. Sciuto, C. Napoli, ITM), IEEE, 2020, pp. 21–24.</p>
        <p>G. Pappalardo, E. Tramontana, A novel cloud- [16] Elaadnl, Iota charging station, 2020. URL: https:
distributed toolbox for optimal energy dispatch //www.elaad.nl/projects/iota-charging-station/.
management from renewables in igss by using
wrnn predictors and gpu parallel solutions, in:
2014 International Symposium on Power
Electronics, Electrical Drives, Automation and
Motion, IEEE, 2014, pp. 1077–1084.
[6] S. Chen, B. Guo, H. Yan, Q. Qin, B. Li, B. Qi,</p>
        <p>Application and prospect of integrated energy
interoperability management system based on
blockchain, in: 2019 IEEE International
Conference on Energy Internet (ICEI), IEEE, 2019, pp.</p>
        <p>421–425.
[7] F. Bonanno, G. Capizzi, G. Sciuto, C. Napoli,</p>
        <p>Wavelet recurrent neural network with
semiparametric input data preprocessing for
microwind power forecasting in integrated generation
systems, in: 5th International Conference on</p>
      </sec>
    </sec>
  </body>
  <back>
    <ref-list />
  </back>
</article>