La boîte à outils sur les compétences psychosociales (CPS

Loading...

Taking too long?

Reload document

| Open in new tab

Article
Blockchain-Oriented Coalition Formation by CPS
Resources: Ontological Approach and Case Study

Alexey Kashevnik 1,* ID and Nikolay Teslya 2 ID

1

2

International research center “Computer Technologies”, ITMO University, 49 Kronverksky Pr.,
197101 St. Petersburg, Russia
Laboratory of Computer Aided Integrated Systems, St. Petersburg Institute for Informatics and Automation
of the Russian Academy of Sciences (SPIIRAS), 39, 14th Line, 199178 St. Petersburg, Russia;
teslya@iias.spb.su

* Correspondence: alexey@iias.spb.su; Tel.: +7-911-211-9880

Received: 15 February 2018; Accepted: 7 May 2018; Published: 8 May 2018

Abstract: Cyber-physical systems (CPS), robotics, Internet of Things, information and communication
technologies have become more and more popular over the last several years. These topics open new
perspectives and scenarios that can automate processes in human life. CPS are aimed at interaction
support in information space for physical entities communicated in physical space in real time.
At the same time the blockchain technology that becomes popular last years allows to organize
immutable distributed database that store all significant information and provide access for CPS
participants. The paper proposes an approach that is based on ontology-based context management,
publish/subscribe semantic interoperability support, and blockchain techniques. Utilization of these
techniques provide possibilities to develop CPS that supports dynamic, distributed, and stable
coalition formation of the resources. The case study presented has been implemented for the scenario
of heterogeneous mobile robots’ collaboration for the overcoming of obstacles. There are two types of
robots and an information service participating in the scenario. Evaluation shows that the proposed
approach is applicable for the presented class of scenarios.

Keywords: blockchain; ontology; context; cyber-physical systems; robotics; interaction; coalition

1. Introduction

The second generation of cyber-physical systems (CPS) consider different aspects of artificial
intelligence aimed at self-awareness and/or self-adaptation between CPS resources. There is a lot of
research in the areas of mobile robotics [1,2], unmanned vehicles [3,4], wearable electronics [5,6] and
others that are required by second generation CPS for effective operation. This paper considers the
problem of dynamical decentralized coalition formation by heterogeneous CPS resources for joint task
performing. For this purpose, resources should not only exchange information with each other but
also have to understand the semantics in CPS, take into account the current situation for self-adaptive
behavior, and recognize compromised situations.

Tasks distribution and assignment among the coalition participants requires a special method.
An example of such a method is the sensor assignment to missions approach, based on knowledge
representation using ontologies and reasoning techniques to distribute tasks among coalition
participants [7]. However, the participants of the coalition do not have a full picture of how the
tasks were distributed. They can get an approximate idea of this, only by making full reasoning among
the knowledge of all coalition members. This behavior is more typical for swarm interaction models
and reduces overall efficiency of the coalition. Usually the existing method is characterized by the
absence of a central or distributed processing log, in which results or coalition participants’ work are

Electronics 2018, 7, 66; doi:10.3390/electronics7050066

www.mdpi.com/journal/electronics

electronics(cid:1)(cid:2)(cid:3)(cid:1)(cid:4)(cid:5)(cid:6)(cid:7)(cid:8)(cid:1)(cid:1)(cid:2)(cid:3)(cid:4)(cid:5)(cid:6)(cid:7) Electronics 2018, 7, 66

2 of 16

displayed. Such a log allows us to reduce the number of calculations on the coalition formation stage
and to distribute tasks taking into account the capabilities and characteristics of all the participants
in the coalition. At the same time, with such a log each participant at any time will have a view of
task distribution as well as when it is expected to receive the result from other participants and what
resources are involved in solving the overall problem. The overall view makes it possible to reconfigure
the coalition on the fly by adding or replacing participants who will know what they should do from
the very beginning.

The paper presents a blockchain-oriented approach to coalition formation of CPS resources that is
based on context management methodology and publish/subscribe-based semantic interoperability
mechanism as well as utilization of a distributed, decentralized, immutable transaction log technology,
also known as a blockchain for creation of the coalition working log. These technologies allow us
to implement self-adaptive behavior of CPS resources for coalition formation. Its distinctive feature
is the combination of private transactions in a block of a given length (defined by the developer of
a particular block) using hash functions, as well as the connection of each new block with the last
block in the chain that is also provided by using the hash function. It is proposed to use two types of
transactions. The first one contains task assignment for coalition participants. It can be used to control
the overall progress of the coalition participants and adjust CPS resource’s ratings, as well as adjusting
coalition structure by replacing the resource that cannot complete the assigned task. The second type
of transaction contains the facts of the system’s consumables spending. Each CPS resource can use
consumables to process the assigned task. The fact of spending is written to the distributed ledger and
can be checked by the CPS manager. Each transaction in the block is signed by the private key of the
transaction owner. Such a mechanism makes it possible to unequivocally determine who entered the
information in the block, track its sequence of changes, and ensure the immutability of information
contained in the log.

The main contribution of the paper is an extension of the classical multi-agent approach to
resource interaction by the ontology-based smart space and blockchain technology. The main features
of the approach are usage of context management techniques for determination of the resources
knowledge as well as its competencies that are related to the current task. The smart space technique
used supports ontology-based publish/subscribe mechanism that allows to create the actual for the
task model of physical space in the information space. This model provides interoperability support for
CPS resources. Based on resource interaction in information space, the coalitions are formed to perform
the task. Every resource is described by the rating, as well as its competencies. Ratings are taken into
account for decision about the resource participation in a coalition. To support the immutability of
the resource ratings the blockchain technology is used for keeping the history of resources interaction,
as well as consumables spending, and preventing compromised situations.

For the approach evaluation, a case study has been presented that considers the scenario of
heterogeneous mobile robots collaboration work for obstacles overcoming. The scenario is based on
the proposed approach and Smart-M3 semantic interoperability platform that implement concept
of smart spaces and provides possibilities of ontology-based mobile robots interaction organization
for their semantic interoperability. The platform makes possible to significantly simplify further
development of CPS, include new information sources and services, and to make the system
highly scalable. Smart-M3 assumes that devices and software entities can publish their embedded
information for other devices and software entities through simple, shared information brokers.
The platform is open source and accessible for download at Sourceforge (Smart-M3 at Sourceforge,
URL: http://sourceforge.net/projects/smart-m3). There are two types of mobile robots that have
been developed for participation in the case study: manipulating robots and measuring robots.
The task is aimed at cargo delivery by the manipulating robot with increased passability. During the
delivery, several obstacles should be overcome. For obstacle overcoming, the manipulating robot
should form a coalition with the measuring robot and knowledge base service to get the appropriate
overcoming algorithm.

Electronics 2018, 7, 66

3 of 16

The prototype of the manipulating robot has been assembled for this scenario based on Lego
Mindstorms EV3 kit. EV3 control block has been re-flashed for LeJOS operating system that allows
to connect it to Wi-Fi network and Smart-M3 information sharing platform. The robot consists of
four motors, four pair of wheels, and an ultrasonic sensor. The measuring robot has been assembled
based on the Makeblock robotics kit. The Raspberry Pi 3 Model B single-board microcomputer was
connected to the control board to expand its computation power. The main task of the Raspberry Pi is
to connect the robot to the Smart-M3 platform, as well as to process data from sensors and calculate
the further robot actions.

The rest of the paper is organized as follows. The next section describes related work in the
area of mobile robot interaction and blockchain technology utilization for next generation CPS. Then
the authors present a blockchain-oriented approach to coalition formation in CPS. After that the
CPS ontology is described. Next, the authors propose a case study that is aimed at heterogeneous
mobile robots collaboration work for the overcoming of obstacles. Finally, the conclusion summarizes
the paper.

2. Related Work

The section presents the related work analysis in the area of interaction between CPS resources
on the example of mobile robots and blockchain technology utilization for next generation CPS.
The following main technologies are actively used for CPS resources coalition formation in the
considered research papers: smart space, ontology management, context management. The usage of
blockchain technologies for CPS resources coalition formation is a rather new research field despite the
great potential of the technology in distributed trustless systems. However, there are several research
works aiming to utilize blockchain technology in the field of coalition operation and Internet of Things
(IoT).

2.1. Related Work Analysis in the Area of Interaction between CPS Resources on the Example of Mobile Robots

The aim of the RoboEarth project is building the Internet of robots. Authors determine the Internet
of robots as a web community that autonomously shares descriptions of tasks they have learned,
object models they have created, and environments they have explored [8]. Authors developed
a formal language for encoding this information and propose methods for solving the inference
problems related to finding information, to determining if information is usable by a robot, and to
grounding it on the robot platform; the infrastructure for using this representation to reason about
the applicability of information in a given context and to check if all required robot capabilities are
available, and mechanisms for creating and uploading shared knowledge.

The authors of the paper [9] deal with an intelligent system of robot and human interaction. This
system uses adaptive learning mechanisms to account for the behavior and preferences of a human.
It includes algorithms for speech and face recognition, natural language understanding, and other
ones that provide the robot possibility of recognition the human actions and commands. A robot used
in this paper (Pioneer 3-AT) is equipped with multiple sensors, scanners and a camera. The paper
focuses on the process of teaching the robot and its mechanisms of interaction with a human.

In the paper [10] authors describe a coordinated process for the agents (in particular, mobile robots)
jointly perform a task, using the knowledge about the environment, their abilities and possibilities
for communication with other agents. The effective use of knowledge about the agents’ capabilities
is implemented by using suitability rates which allow agent to choose the most appropriate action.
This approach was tested during a football match between mobile robots. The interaction in this article
is considered in terms of the choice of an agent to perform an action.

The authors of the paper [11] propose a methodology for coordination of autonomous mobile
robots in jams in congested systems. The methodology is based on reducing a robot’s speed in jams
and congested areas, as well as in front of them. Thus, the robots slow down if detecting other robots

Electronics 2018, 7, 66

4 of 16

in front of themselves. The effectiveness of the methodology was tested in a simulation of a congested
system with a jam.

The paper [12] discusses the combining of wireless sensor networks and mobile robots. Authors
suggest layered swarm framework for the interaction of decentralized self-organizing complex
adaptive systems with mobile robots as their members. This framework consists of two layers (robots
and wireless LAN) and communication channels within and between layers. During the experimental
verification of framework mobile robots independently reached the destination.

The paper [13] is devoted to the problem of robots group control in non-deterministic, dynamically
changing situations. The authors propose a method for solving formation task in a group of quadrotors.
The method makes it possible to ensure accurate compliance with distances between quadrotors in the
formation, as well as featuring low computational complexity.

The authors of the paper [14] shows a system of robots interaction based on BML—Battle
Management Language. This language is intended for expressing concise, unambiguous orders
read by both humans and robots. The orders are transmitted from the system of multiple robots and
processed for further distribution specific commands between them.

In the paper [15], design of the human-robot interaction, improving human situational awareness
on the basis of an agent-oriented approach is described. The approach is applied when mobile agents
move (autonomous robots and astronauts in protective gear) on the Moon’s surface in a limited
geographically area.

The paper [16] presents the concept of the learning factory for cyber-physical production systems
(LVP) and presents robotics and human-robot collaboration. Authors use ontologies in the context of
flexible manufacturing scenarios to equip autonomous robots with the necessary knowledge about
their environment, opportunities and constraints. They discuss the use of semantic technologies
together with cyberphysical systems for integrating decision making into smart production machinery.
The paper [17] presents extensions to a core ontology for the robotics and automation field.
Authors of the paper proposed ontological approach aims at specifying the main notions across
robotics subdomains.

The paper [18] defines the current issues and solutions possible with ontology development for
human-robot interaction. It describes the role of ontologies in robotics at large, provide a thorough
review of service robot ontologies, describe the existing standards for robots, along with the future
trends in the domain, and defines the current issues and solutions possible with ontology development
for human-robot interaction.

The paper [19] evaluates the proposed ontology through a use case scenario involving both
heterogeneous robots and human-robot interactions, showing how to define new spatial notions
using an ontology. It discusses experimental results, presenting the ontology strengths. The paper is
concerned with the development of model-based systems engineering procedures for the behavior
modeling and design of CPS. Three independent but integrated modules compose the system: CPS,
ontology and time-reasoning modules. This approach is shown to be mostly appropriate for CPS or
which safety and performance are dependent on the correct time based prediction of the future state
of the system. Consequently, the ontological approach applicability for modeling the problem area
is justified.

The approaches described by [9,11] require a well-defined functionality of robots, so the set of
tasks undertaken is restricted. The paper [10] describes the knowledge of the capabilities of robots that
is generated dynamically during the execution of tasks what allows to manage the robots effectively.
Thus, no limitation to robots’ functionality and dynamic determination of the options available can
provide the widest range of tasks to execute. Authors of the paper [12] show the effectiveness of the
indirect interaction of robots, that is, the interaction through an intermediary that is Smart-M3 platform
in the paper. The paper [13] discusses a formal model that can be used for interaction. In the paper [14]
the effectiveness of using specialized language for robots control is shown. Authors of the paper [15]
demonstrate the applicability of the multi-agent systems approach for robot and human interactions

Electronics 2018, 7, 66

5 of 16

in various fields. Many researches inherit the idea of ontologies usage for modelling context in
social-cyber physical systems. An ontology should include description of physical components, agents,
model of problem area. It consists of classes, subclasses, properties, and relationships between classes.

2.2. Related Work Analysis in the Area of Blockchain Technology Utilization for Next Generation CPS

The paper [7] considers the use of blockchain technology to create a logically centralized system
for managing the distribution of tasks among coalition members. The blockchain technology itself is
considered as a mechanism for recording and tracking transactions between participants performing
control over the distribution of tasks. At the same time, it is possible to create a virtual log that contains
all distributions of the assets available among the coalition participants.

The paper [20] provides an example of the E-Business model for decentralized autonomous
corporations based on IoT and blockchain. The blockchain network is used here to create smart
contracts and perform payment transactions using Bitcoin and the system’s own currency—IoTcoin.
The approach presented provides the following advantages: internal resources of the system are
presented in the form of payment system tokens; distributed ledger for transactions, built-in mechanism
of digital signature; and protection of transactions by linking blocks with a hash function of a given
complexity. Interaction between elements is carried out by transactions passing through four phases:
preparing a transaction, negotiating, signing a contract, and fulfilling the terms of the contract.

A similar method is considered in relation to a Smart Home scenario in [21]. Also, there
is an unchanged transaction log as an advantage of integrating IoT and Blockchain technologies.
The proposed framework provides the following transaction processing functions: transaction storage,
access to the transactions history and real-time monitoring. Also, due to the adaptation of the
hierarchical structure of the proposed framework (which includes three levels: the smart home,
overlay network, and cloud storage), optimization of resource consumption is achieved, and scalability
of the network is increased. The functions of the blockchain are focused on the level of interaction of
elements in the overlay network.

It is also noted that the combination of the peer-to-peer network and the cryptographic algorithms
that are underlying the blockchain technology allow for a negotiation process and consensus building
without the presence of any controlling authorities. The distributed nature of the blockchain can be
used in swarm robotics to store global knowledge about swarm actions [22]. At the same time, due to
blockchain, the security of the transmitted data is ensured (garbage data can affect the achievement of
a common goal), distributed decision making (creating a distributed voting system for the solution
and use of the multi-signature), separation of robots behavior (switching between behavior patterns
depending on the role in the swarm), the emergence of new business models using the swarm.
In addition, the availability of a distributed transaction ledger allows new robots to join the swarm
and gain all the knowledge they have gained prior to the moment of inclusion by downloading and
analyzing the transaction history.

The results of the research presented can be summarized in the following way. The blockchain
is mostly used as immutable storage for information exchange between robots. Information stored
in the blockchain could contain records about task and assets distribution [7,21], smart contracts and
payment transactions [20], or global knowledge about coalition actions [22]. In combination with
IoT, concept blockchain technology can provide more trust for transaction through IoT, due to the
storing information about transactions in immutable log that can be verified by every IoT participant.
In contrary to existing approaches, blockchain does not require a central authority that provides trust
for all nodes. All nodes negotiate with each other and come to consensus using one of the possible
mechanisms: Proof of Work, Proof of Stake, or practical Byzantine fault tolerance [23]. Assuming that
robots are forming a coalition without any control center, the blockchain can be used for safe and
trustiness logging of robots’ task distribution and reward for task solving.

Electronics 2018, 7, 66

6 of 16

3. A Blockchain-Oriented Ontological Approach to Coalition Formation

For coalition formation in CPS it is required to support the semantic interoperability between
resources, as well as providing infrastructure for information & knowledge sharing between them
and keeping agreements between resources made in scope of interaction in the system. The CPS
concept is aimed at tight integration of the physical and cyber spaces based on interactions between
these spaces in real time. Resources are exchanging the information with each other in information
space while their physical interaction occurs in physical space. For interoperability support, it is
needed to create the model of CPS and support the interaction of resources based on this model.
One of the possible approaches to problem domain modelling is ontology management. The ontology
formally represents knowledge as a set of concepts within a domain, using a shared vocabulary to
denote the types, properties, and interrelationships of those concepts. For current situation modelling
and reducing the search space of potential coalition members, the utilization of context management
technology is proposed. The aim of this technology is a context model creation that is based on the
ontology of problem domain and information about current situation in physical space. Context is
defined as any information that can be used to characterize the situation of an entity. An entity is a
person, place or object that is considered relevant to the interaction between a user and an application,
including the user and application themselves [24]. Context is suggested to be modeled at two levels:
abstract and operational. These levels are represented by abstract and operational contexts, respectively.
The process of coalition formation based on abstract and operational contexts is shown in Figure 1
(enhanced based on [25]).

Figure 1. Abstract and operational contexts for coalition formation.

Abstract context is an ontology-based model of a potential coalition participant related to the
current task. Abstract context is built based on integrating information and knowledge relevant to
the current problem situation. Operational context is an instantiation of the domain constituent of
the abstract context with data provided by the contextual resources. Thereby, coalition formation is
implemented with the following phases. In the first phase the resources are waiting to participate
in the process of task performing. Every resource is described by an ontology. For the second phase
the abstract context is created based on the resource ontology, a task appeared in CPS, and current
situation. Abstract context includes all accessible knowledge relevant to the task. For the third phase,
the process of concretization of these knowledge by information accessible in information space for
operational context formation is implemented. The operational context is published in information

Electronics 2018, 7, 66

7 of 16

space and becomes accessible for other potential coalition participants. The coalitions of resources are
created based on their operational context intersections (see

For information space organization, a semantic interoperability platform is used that is based on
blackboard architecture that provides possibilities for potential coalition participants to implement
indirect interaction with each other. Thereby, a virtual coalition is created in information space and then
a physical coalition appears in physical space (CPS resources implement the joint task). Interaction
of potential coalition participants in information space is implemented using the ontology-based
publish/subscribe mechanism that provides possibilities for resources publishing their information
and knowledge, and subscribing to interesting information using ontologies. When a resource registers
to be a potential coalition participant it uploads its own ontology to the semantic interoperability
platform. This resource ontology formalizes the main resource capabilities and constraints that have
to be satisfied to use the capabilities. Thereby, the interoperability is supported based on open
information space and ontology-based publish/subscribe mechanism. A potential coalition participant
can participate in a joint task if the operational context in information space is matched with its abstract
context. More details about the ontology matching can be found in the paper [26].

Simultaneously with the publishing of the operational context to the semantic interoperability
platform the signed transaction is writing to the blockchain network. The transaction consists of
the information published to the platform by a particular resource on each stage of the joint task
performance, signed by a private key of the resource. The transaction is created on the coalition
participant hardware and includes the new block that writes to the local copy of the blockchain. Then
the blockchain spreads the new block among the other coalition participants that can verify block’s
source as well as transactions inside it using the public key. Figure 2).

Figure 2. Blockchain-Oriented Conceptual model for resource interaction for coalition formation.

Figure 3 shows the resource profile model that is aimed at automating the process of coalition
creation and takes into account the ratings of the resources considered for coalition formation. Every
time the resource participated in the coalition, other resources and users have the possibility to estimate
their competencies’ ratings and the overall rating of the resource. The model consists of the following
components: basic resource information, its competence & constraints, and resource interaction history.
Basic resource information includes the following information about the resource: identifier, name,
description, and rating. Competence and constraints of resource is aimed at description of the tasks
the resource can perform and constraints that have to be satisfied to perform these tasks. Competence

Electronics 2018, 7, 66

8 of 16

and constraints of resource includes: basic competencies, extended competencies, and constraints.
Basic competencies and extended competencies can be estimated with ratings by another resource,
and users. The blockchain network is used for the following purposes. First is a coalition formation.
During the coalition formation process the potential resources share their competencies and get
subtasks that are parts of the task. Each subtask assignment is stored in the blockchain network.
In this case for other resources it is easy to check who is responsible for a certain subtask. If some
resource fails in the subtask, its competence rating is decreased, and the resource can be replaced
by another one with a higher competence rating. In the case of a subtask performed successfully,
the ratings of competencies related to this subtask are increasing. All rating changes are stored in
the blockchain network and can be tracked for preventing the unauthorized changes. The second
purpose is aimed at information exchange and consumables tracking. During the subtask performance,
each resource can spend some of the system’s consumables (e.g., power, time or task-specific resource,
like seeds for planting). The details of consumables spending is stored in the blockchain network.
Every resource can check at any time the value of consumables that was spent by another resource for
subtask performance. It is helpful to track the task performance as well as for in-time consumables
supplying and increasing the efficiency of joint work by adjusting the coalition. In these cases the
blockchain network is used to prevent the submitting and use of malicious information. Only signed
information from coalition members can be used in joint task performance. The immutability property
of the blockchain allows us to track actions performed by each resource and detect the guilty resource
that causes the failure and decrease its competence rate. Unlike existing approaches, the blockchain
can provide an immutable log with distributed trustiness between robots without any central point.
In cases when robots are acting only with each other, without any central node, the blockchain can
simplify trustiness providing.

Figure 3. Resource Profile Model.

4. CPS Ontology

Developed CPS ontology is based on definitions and abbreviations from Suggested Upper Merged
Ontology (SUMO), proposed in [27]. The model is built for CPS, where the mobile robots are interacting
with each other for joint task performance. The ontology includes description of the main processes
that occur when coalitions of robots are created to solve problems in cooperation. The protégé
editor has been used to create the ontology. The ontology classification consists of two main classes:
“Physical Space” and “Information Space”.

Class “Physical Space” described all physical entities which participated in a coalition (devices,
sensors, motors, end etc.). “Information Space” refers to computational resources and processes for
coalition formation. Considering the classification of “Physical Space” in more detail (see Figure 4),