Applying the PROSA Reference Architecture to Enable the

Interaction between the Worker and the Industrial Robot

Case Study: One Worker Interaction with a Dual-Arm Industrial Robot

Ahmed R. Sadik

1,2

and Bodo Urban

1,2

University of Rostock, Universitätsplatz 1, 18055, Rostock, Germany

Fraunhofer Institute for Computer Graphic Research IGD, Joachim-Jungius-Str. 11, 18059, Rostock, Germany

Keywords: Flexible Manufacturing Paradigm, Holonic Manufacturing System, Multi-Agent System, Worker-Industrial

Robot Cooperation, Semi-Autonomous System, Human-Robot Interaction, Gesture Recognition.

Abstract: Involving an industrial robot in a close physical interaction with the worker became quite possible, as a result

of the availability of different collaborative industrial robots in the market. The physical cooperation between

the industrial robot and the worker usually done under the umbrella of the flexible manufacturing paradigm,

where both the industrial robot and the worker need to change their tasks fast and efficiently, to cope with the

changes in the manufacturing process. This means that a reliable manufacturing control system must stand

behind this physical interaction to achieve the proper communication interaction. A holonic control

architecture is an ideal solution for this problem. Therefore, during this research we study the most commonly

applied model of the holonic control architecture, then we apply this architecture on our case study, where

one worker cooperates with a dual-arm industrial robot to build and produce any new product. Also the

research uses the worker’s hand gesture recognition as a method to interact with the industrial robot during

the execution of a cooperative production scenario.

1 INTRODUCTION

Human-Robot Interaction (HRI) is the field of

research focuses on understanding, designing and

evaluating robotic systems which involve the human

as an essential element of these systems (Goodrich

and Schultz, 2007). The HRI fundamental rules are

originally based on a fictional speculation as stated by

author Isaac Asimov’s novel “I, Robot” (Pinker,

1999). The first two fundamental rules of the HRI are

as following:

• A robot may not injure a human being or,

through inaction, allow a human being to

come to harm.

• A robot must obey orders given to it by

human beings.

As a scientific interpretation of these rules. The

first fundamental is mainly addressing the problem of

a safe physical HRI, while the second rule addresses

the problem of HRI information communication.

Accordingly the term interaction in the context of

HRI can either mean physical or information

interaction. A very good example where the HRI

physical close proximity interaction is more dominant

than the information interaction is an exoskeleton

robotic suit. In an exoskeleton the human literally

wears a robotic suit to magnify his strength and

endurance (Fontana, 2014). On the other hand, a tele-

operated robot arm operates in a nuclear reactor

(Parker and Draper, 1998)

is an example of a remote

HRI where the information interaction is more

dominant than the physical interaction. However in

most of the HRI scenarios, both forms of interaction

coexist together, regardless which one is more

obvious or dominant.

The HRI has gained a great attention in industrial

applications in particular (Lasota et al, 2014). As a

result of this attention, a new generation of safe

cooperative industrial robots are now available in the

commercial market. Example of these robots are

KUKA lightweight, Rethink Baxter, YuMi ABB dual

arm, and Universal Robots. These robots apply

different technologies and methodologies to grantee

the human safety during a close proximity interaction.

Therefore, safe physical HRI is achieved. However,

there is no clear control architecture which specifies

the information communication interaction between

190

R. Sadik A. and Urban B.

Applying the PROSA Reference Architecture to Enable the Interaction between the Worker and the Industrial Robot - Case Study: One Worker Interaction with a Dual-Arm Industrial Robot.

DOI: 10.5220/0006191801900199

In Proceedings of the 9th International Conference on Agents and Artiﬁcial Intelligence (ICAART 2017), pages 190-199

ISBN: 978-989-758-219-6

the safe industrial robot and the worker in a

cooperative production scenario.

In section 2 of this paper, we introduce some well-

known methodologies and terminologies which are

basic keys in Intelligent Manufacturing System (IMS)

design. It is important to know these methodologies

and terminologies in order to formulate the problem

statement in section 3. Section 4 explains the

technologies which are used to apply the solution

concept. Therefore an overview of the solution model

can be explained in section 5. Section 6 shows a case

study where we implement the solution concept.

Finally section 7 wraps up the work summary with

the conclusion and the future research.

2 BASIC KEY CONCEPTS

2.1 Flexible Manufacturing System

A Flexible Manufacturing System (FMS) is an

integrated system of machine modules and material

handling units which together can offer a certain

amount of flexibility (Lozano et al, 1994). A typical

FMS is composed of a series of manufacturing

workcells which controlled by a stand-alone

controller. A workcell can contain an automatic

machine, a material handling and storage station, an

industrial robot, or a human worker.

The flexibility term can be applied to different

aspects of the manufacturing system, such as: the

machine, the material handling and routing, the

control system, and the product building plan and

volume (Elmaraghy, 2005). The objectives of the

FMS are to increase the manufacturing system

reliability, optimizing the cycle time, reduce the lead

times and costs, overcome the internal changes like

breakdowns, and improve the worker productivity,

the machine efficiency, and the overall quality (Koren

et al, 1999; Kruger, 2015).

A Worker and Industrial Robot (W&IR)

cooperative workcell is a novel case study of the

FMS, where the worker can use the industrial robot

as an intelligent tool to achieve the FMS objectives.

The FMS control system could follow a centralized

or a distributed control topology. In a centralized

control topology, all the workcells controllers are

supervised by one central controller which carry out

the final decision making. In a distributed control

topology, the decision making task is distributed

between all the controllers. The scope of this research

focus on distributed control approach as a modern

intelligent method to control the W&IR cooperative

workcell.

2.2 Holonic Manufacturing System

In the late of sixties, the term holon has been

introduced for the first time by philosopher Koestler

(Koestler, 1967). Koestler developed the term as a

basic unit in his explanation for the evolution of

biological and social structures. Based on his

observations that organisms (e.g., biological cells) are

autonomous self-reliance units, which have a certain

degree of independent control of their actions, yet

they still subject to a higher level of control

instructions. His conclusion was that any organism is

a whole “holos” and a part “on” in the same time,

which derived the term holon (Giret, 2008). The

concept of holon has been adopted in the early of

nineties by the IMS consortium, to define a new

paradigm for the factory of the future. IMS has

defined the holon as an autonomous cooperative

building block of the manufacturing system, that can

be used to transform, transport, store and/or validate

the information and the physical objects (Radu and

Frank, 2006).

The Holonic Manufacturing System (HMS) is

basically a distributed control and communication

topology which divides the manufacturing process

tasks and responsibilities over different holon

categories. Two well-known reference models are

following the holonic manufacturing architecture

which are: Product-Resource-Order-Staff

Architecture (PROSA) model (Van Brussel et al,

2003), ADAptive holonic COntrol aRchitecture

(ADACOR) model (Leitao and Restivo, 2006).

The PROSA reference model implements three

basic holons as shown in Figure 1-a. These holons are

resource, product, and order holons. The resource

holon is a physical entity within the manufacturing

system, it can represent a robot, machine, worker, etc.

The product holon stores the process and the product

tasks needed to insure the correct manufacturing of a

certain product. An order holon is responsible for

assigning the tasks and making sure they have been

accomplished. The ADACOR reference model

(Leitao and Restivo, 2008) implements the same three

PROSA basic holons plus a fourth holon called a staff

or a supervisor holon. The supervisor holon is

providing coordination services when it is needed to

cooperate outside the boundaries of the workcell.

The holon generic structure is shown in Figure 1-

b. The resource holon is usually composed of two

components which are responsible for the physical

and communication interaction respectively. While

the other holons can have only the communication

component.

Applying the PROSA Reference Architecture to Enable the Interaction between the Worker and the Industrial Robot - Case Study: One

Worker Interaction with a Dual-Arm Industrial Robot

191

Figure 1: (a) PROSA HMS Reference Model – (b) Holon

Generic Structure.

3 PROBLEM FORMULATION

Close physical interaction between the worker and the

industrial robot is a new trend in manufacturing. This

trend was proposed as a part of Industrie 4.0

recommendations as a new vision for the smart

factory of the future. Especially for the Small and

Medium-Sized Enterprises (SMEs), as they apply a

highly customized industrial processes which follow

the flexible manufacturing paradigm. Accordingly,

the worker existence alongside the industrial robot is

essential in an SME, as the worker can easily adapt to

the fast changes in the production requirements.

Simultaneously, the industrial robot is important

resource on the factory shop floor, as it is reliable in

terms of speed, load lifting, and accuracy, etc.

The first fundamental rule of the HRI can be

easily obtained in a W&IR cooperative workcell, due

to the existence of a variety of safe industrial robots

in the commercial market. The problem is to achieve

the second fundamental rule of the HIR, which is to

connect the worker and the industrial robot together

from the information communication point of view.

The PROSA reference model is a well-known

control and communication architecture which has

commonly used to solve different FMS problems.

However implementing the PROSA model over this

specific case of the FMS (i.e. a W&IR cooperative

workcell) is a new problem. The main purpose of

applying the PROSA model is to afford the worker

the privilege to use the industrial robot as an

intelligent tool within the production process. In a

semi-autonomous control manner, the control system

should allow the worker to interfere the production

process and teach the robot a new task. This new task

can be stored in the system and used as a building

block of more complicated products.

Furthermore, in order to complete the W&IR

interaction, the control system must be able to

recognize the worker activities. Therefore, the second

part of this research problem is to find an appropriate

recognition method which can be used to express the

different tasks done by the worker.

4 BASIC KEY TECHNOLOGIES

4.1 Autonomous Artificial Agent and

Multi-Agent System

A software agent is a computer system situated in a

specific environment that is capable of performing

autonomous actions in this environment in order to

meet its designer objective (Jennings & Wooldridge,

1998). An agent is responsive, proactive and social.

Responsive means that the agent can perceive its

environment and respond in a timely fashion to the

changes occurring in it. Proactive means that the

agent is able to exhibit opportunistic, goal directed

behaviour and take initiative. Social means that the

agent can interact with other artificial agents or

humans within its environment in order to solve a

problem.

Conceptually, an agent is a computing machine

which is given a specific task to execute. Therefore,

it chooses certain set of actions and formulate the

proper plans to accomplish the assigned task. The set

of actions which are available to be chosen by the

agent are called a behavior. The agent behaviors are

mainly created by the agent programmer. An agent

can formulate one or more plan to reach its target. The

selection of an execution plan among others would be

based on a certain criteria which has been defined by

the agent programmer. Building an execution plan is

highly depending on the information which inferred

by the agent from its environment. A Multi-Agent

System (MAS) is a collective system composed of a

group of artificial agents, teaming together in a

flexible distributed topology, to solve a problem

beyond the capabilities of a single agent (Shen et al,

2006).

ICAART 2017 - 9th International Conference on Agents and Artiﬁcial Intelligence

192

Figure 2: Java Agent Development (JADE) Framework.

JAVA Agent Development (JADE) is a

distributed MAS middleware framework (JADE,

n.d.). JADE applies reactive agent architecture which

complies with the Foundation for Intelligent Physical

Agent (FIPA) specifications, and provides a graphical

interface to deploy and debug a MAS (Bellifemine et

al, 2007). FIPA is an IEEE computer society

standards organization that promotes agent-based

technology and the interoperability of its standards

with other technologies (FIPA, n.d.). JADE agents

use FIPA-Agent Communication Language (FIPA-

ACL) (Poslad, 2007), to exchange messages either

inside its own platform or with another platform in a

distributed MAS as shown in Figure 2.

Each JADE instance is an independent thread

contains a set of containers. A container is a group of

JADE agents run under the same JADE runtime

instance. Every platform must contain a main

container. A main container contains two necessary

agents which are: an Agent Management System

(AMS) and a Directory Facilitator (DF). AMS

provides a unique ID for every agent under its

platform, to be used as an agent communication

address. While the DF announces the services every

agent can offer under its platform, in order to

facilitate agent service exchange, so that each agent

can obtain its specific goal (Caire, 2009; Teahan,

2010).

4.2 Hand Gesture Recognition

One of the most successful gesture recognition

techniques is the vision-based sensing that applies

active technique approach. In active sensing a signal

of a burst of (light, microwaves or sound) waves is

emitted, then the reflected signal by the surrounding

is received back by the sensor. Often the sensor is non

active as long as no motion occurs in its sensing

range, till some object moves within this sensing

range, therefore the change in the reflected signal

activates the sensor. The most famous sensors which

belong to this category are the Kinect and Leap

Motion (Berci and Szolgay, 2007).

Figure 3: (a) Leap Controller Dimensions – (b) Leap

Controller Coverage Range.

The Leap Motion controller (Leap Motion, n.d.) is

a sensor device that aims to translate the hand

movements into computer commands. The Leap

dimensions are 8 cm in length and 3cm in width, and

it can be connected to the computer using a USB

connection as shown in Figure 3-a. The controller

range of sensing is a hemispherical volume which

extends to of 60 cm in radius as shown in Figure 3-b.

Using two monochromatic IR cameras and three

infrared LEDs the device observes its sensing

volume. The infrared LEDs emit a 3D pattern of IR

light dots, simultaneously the cameras reconstruct the

reflected data in a rate of a 300 frames per second.

The constructed data transfers to the host computer

via the USB connection, where it can be parsed by the

Leap Motion controller software (Potter et al, 2013).

Applying the PROSA Reference Architecture to Enable the Interaction between the Worker and the Industrial Robot - Case Study: One

Worker Interaction with a Dual-Arm Industrial Robot

193

4.3 Cooperative Industrial Robots:

Baxter Rethink Robot

The robot Baxter is a collaborative dual-arm robot.

Every arm is a 7 DOF articulated industrial robot.

Collaborative robotics is to share the process between

two robots or a robot and a human being, unlike the

conventional industrial robot that repeatedly

performs the very same task on the production line

without a direct interaction with the human coworkers

in the process. Baxter is slower than the traditional

industrial robot and is "compliant", i.e. its movements

are elastic, non-hazardous to the humans and closer to

that found in the nature. The objective of the

collaborative robotics, rather than performing a single

task constantly, which is the case of a conventional

industrial robot, is to be able to carry several small

tasks, alongside humans in its work environment.

Baxter is particularly suitable for SMEs that do not

have sufficient volume to automate sophisticated

tasks. But they can achieve higher productivity if they

automate many simple tasks (Rethink Robotics, n.d.).

The maximum payload of one of Baxter’s arm is

2.3 Kg. The two arms are programmable within the

same code. Baxter has mechanisms of "anti-self-

collision" for compensating its movements to avoid

collisions of the two arms. These mechanisms may be

deactivated in specific cases (passage of an object

from one hand to the other, handling two hands close

together). It is therefore possible to synchronize the

two arms movements to coordinate them together or

controlling any of them separately.

5 SOLUTION MODEL

While the PROSA is a conceptual model focuses on

HCS description, it does not specify a certain

technology to apply this concept. On the other hand,

artificial agent technology is a general purpose

solution which can apply the PROSA concept. Thus

during this research, JADE agent framework has been

used to implement the concept of the PROSA. Figure

4 illustrates the implementation of the PROSA basic

holons over a W&IR cooperative workcell. With the

assumption that a single worker cooperates with the

Baxter dual-arms in flexible manufacturing scenario.

On the worker platform, there are four different

holons, two of them are resource holons which locate

on the shop floor layer. The first resource holon is the

worker User Interface (UI). This worker UI holon has

a physical component which is the worker laptop. The

communication component of this holon is

implemented in the automation layer in the form of

Figure 4: Worker & Industrial Robot Flexible Cooperative

Scenario based on Holonic Manufacturing Concept.

the worker task display agent. The worker UI is

basically responsible for creating the new tasks either

for the worker or the robot, and constructing new

products. Also the worker UI displays the assignment

task for the worker during the product execution. The

second resource holon is the worker task execution.

The worker task execution uses the Leap

controller as a physical competent to capture the

worker’s hand gestures which are associated with

different work events. The communication

component of this holon is the worker task execute

agent. The last two holons on the worker side are the

product and the order holons, which only implement

communication components, therefore they are only

located in the automation layer in a form of

autonomous artificial agents.

On the robot platform, there are two resource

holons. The first resource holon is the robot display.

Similar to the worker UI holon, the robot display

holon has two components, its physical component is

the Baxter screen, which is connected to the robot

task display agent as its communication component.

The robot display holon displays the baxter task while

execution. The second resource holon is the robot task

execution. The physical component of this holon

could be Baxter right or left arm. In either cases the

Baxter arms are connected with the robot task

execution agent as their communication component.

The robot task execute holon can assign a task either

for the right arm, the left arm, or both.

The last holon in this solution model is the robot

task building resource holon. This holon is in fact

distributed over both the worker and the robot

platforms. This is because it is used by the worker to

teach one of the Baxter arms a new task. The worker

platform and the robot platform are communicating

with each other’s using a Wireless Local Area

ICAART 2017 - 9th International Conference on Agents and Artiﬁcial Intelligence

194

Network (WLAN) technology which deploys FIPA-

ACL Communication Protocol.

6 CASE STUDY

During this section, a simple product will be built and

executed using the previously mentioned solution

concept. The aim is to show as simple as possible how

the holons interact. Therefore, a product composed of

three consequent tasks will be constructed. The first

task will be assigned to the Baxter right arm, the

second task will be assigned to the worker, and the

final task will be assigned to the Baxter left arm.

Figure 5: A Case Study for a Three Tasks Product.

6.1 Product Building

Figure 5 shows the worker UI and the different steps

to build a new product. Figure 5-a shows the Main

App UI, which is mainly used to navigate to create a

new task either for the robot or the worker, and

create/execute a new product. Figure 5-b shows the

creation of new robot task via the robot task UI. In

our case study, two tasks are assigned to the Baxter.

Before the worker starts teaching the Baxter a new

task, he gives the task a name and selects which arm

will be used. Then presses start recording button to

physically teach the robot arm the new task,

ultimately the worker presses stop recording button

when finish teaching.

Figure 6: JADE Interaction to Teach a New Task for the

Baxter.

Figure 6 shows the communication interaction

between Baxter-Teach@WorkerPlatform and

Baxter_Teach@BaxterPlatform in order to teach

Baxter a new task. An INFORM message is sent from

the worker side in case of start/stop recording, and the

message is replied by a CONFIRM message from the

Baxter side to complete the handshaking process as it

can be seen in Figure 7. The start teaching message

holds the name of the task and which arm should

perform this task. Accordingly the task will be

Applying the PROSA Reference Architecture to Enable the Interaction between the Worker and the Industrial Robot - Case Study: One

Worker Interaction with a Dual-Arm Industrial Robot

195

recorded on Baxter Platform using this exact task

name as a reference.

Figure 7: ACL-Messages exchange to start task recording.

Figure 5-c shows the creation of a new worker

task via the worker task UI. In our case study four

different gestures have been programmed to be

recognized by the Leap Motion Controller. Those

four gestures are Pick and Place an object, Horizontal

Swipe the hand from left to right, Lean the right hand

back, and finally a Tool (Screw Driver) recognition.

Swipe hand and Tool are pre-defined gestures by the

Leap SDK. However Pick and Place and Lean right

hand back are customized gestures by our code which

will be shown in the next section. Pick and Place has

been selected as an indication of the worker task.

After defining all the needed tasks from both the robot

and the worker to build a product. A product building

plan can be defined using the product UI as shown in

Figure 5-d.

6.2 Product Execution

After pressing execute product button in the Main

App UI, an ACL-Message with a communicative act

“AGREE” is sent from the Product Order agent to the

Order Execute agent as shown in Figure 8.

Figure 8: JADE Interaction to execute a Three Tasks

Product.

ICAART 2017 - 9th International Conference on Agents and Artiﬁcial Intelligence

196

The AGREE message is an indication to start the

execution of a new product. The message content

holds the task target and the task name as shown in

Figure 9-a. the Order Execute agent answers the

Product Order agent with a CONFIRM message the

same way described previously in section 6.1 during

teaching Baxter a new task. Based on the task target

within the AGREE message, the Order Execute agent

assigns a new task execution for that target. Therefore

it informs the task to the

Baxter_Task@BaxterPlatform and simultaneously to

the Task_Dispaly@BaxterPlatform.

Figure 9: ACL-Messages to start tasks execution.

Both the last mentioned agents send back

confirmations that they received a new task to the

Order Execute agent. When Baxter Task agent

receives an INFORM message to start executing a

new task, it searches into the message contents for the

task name. When Baxter Task agent finds a match of

the task name with the tasks previously recorded. It

starts to play back this exact task, and finally sends an

ACL-Message with communicative act “PROPOSE”

to the Product Order agent. The PROPOSE message

indicates that a task done from the product. As the

product tasks still not finished, the Product Order

agent sends an ACL-Message with communicative

act “ACCEPT-PROPOSAL” to the Order Execute

agent.

Figure 10: The Python Server Code to define the Worker

Hand Gestures.

The ACCEPT-PROPOSAL message that has

been sent to the Order Execute Agent is an indication

of a new task assignment. The same previously

mentioned mechanism to assign a new task for the

robot is followed to assign a worker task. As this time

the ACCEPT-PROPOSAL message content indicates

a Worker Task with type Pick and Place as shown in

Figure 9-b. Thus the Order Execute agent informs the

Applying the PROSA Reference Architecture to Enable the Interaction between the Worker and the Industrial Robot - Case Study: One

Worker Interaction with a Dual-Arm Industrial Robot

197

task to the Worker-Task@WorkerPlatform and to the

Task-Dispaly@WorkerPlatform. Both the last

mentioned agents send back confirmations that they

received a new task to the Order Execute agent. When

the Worker Task agent receives an INFORM message

to start executing a new task, it extracts the task name,

initialized a unique socket, and waits the Leap server

to capture this specific gesture. The definitions of the

different gestures are written in the Leap Server using

Python language as shown in the Figure 10.

When the Leap server captures the specified

gesture by the Worker Task agent, it informs the

Worker Task agent that this gesture was detected.

Accordingly the Worker Task agent turns off the

connection socket and sends a PROPOSE message to

the Product Order agent to inform a task done event.

The Baxter Task-2 is achieved exactly the same way

Baxter Task-1 done as can be seen in Figure 9-c,

except one important difference. At the final step of

the product execution when the Baxter Task agent

sends a PROPOSE message to the Product Order

agent. The Product Order agent realizes that all the

product tasks have been fulfilled, therefore it sends

the Order Execute agent an ACL-Message with a

communicative act “REJECT-PROPOSAL”, to

indicate the end of this routine.

7 SUMMARY, CONCLUSION

AND FUTURE WORK

During this research, we introduced the basic

fundamental rules of the HRI. In the field of industrial

robots, the first fundamental rule has become easily

achievable, because of the availability of safe

cooperative industrial robots in the robotics market.

However to achieve the second HRI fundamental

rule, an intelligent control and communication

architecture must be found and implemented. The

purpose of this control architecture not only to

connect the industrial robot and the worker together

from the information point of view, but also to enable

the flexible cooperation between them.

The research has applied the PROSA holonic

control reference model to solve the addressed

problem. The PROSA reference model specifies the

three basic holons which are required in intelligent

manufacturing scenarios. That was the reason of

choosing this model specifically, as the scale of our

study still limited to one worker in cooperation with

one or two robots.

Based on the PROSA specifications, the physical

and information interaction model between a worker

and a dual-arm has been constructed. JADE

autonomous agent framework has been used to

implement and deploy the communication

components of the PROSA holons. A robot task

teaching resource holon has been used by the worker

to teach any of the Baxter arms a new task. While a

robot task execution resource holon was responsible

for executing this task. The Leap sensor has been

chosen to track the gestures of the worker’s hand,

accordingly a dedicated resource holon was assigned

to monitor the Leap Motion controller server and pass

the gestures to this resource holon when it is needed.

As has been shown in the case study, the

production flexibility has been accomplished using

the proposed solution. Thus, the worker could teach

the robot many different tasks at any time during the

manufacturing process. Simultaneously, he can

assign different hand gestures which are associated

with his activities during the production process.

Ultimately, the worker can use different combinations

of his hand gestures along with the robot tasks which

are stored in the system to construct more

complicated production routines. Accordingly, the

proposed solution can increase the manufacturing

system reliability by adapting to the production

requirements, reduce the final production lead time

and cost, and improve the worker productivity and the

robot efficiency.

As the coordination between two or more than

worker and industrial robot is out of context of this

research, a staff holon has not been implemented.

However, in the future work the proposed solution

can be scaled over more than one worker and

industrial robot. Therefore the staff holon can be

taken into consideration. Also during the future work,

more hand gestures can be investigated and studied,

thus it can be offered to the worker to build more

sophisticated products.

REFERENCES

Bellifemine, F., Caire, G., Greenwood, D., Developing

multi-agent systems with JADE, Vol 7 ed. John Wiley

& Sons, 2007.

Berci, N., Szolgay. P., Vision based human–machine

interface via hand gestures, 18th European conference

on circuit theory and design (ECCTD 2007), pp 496–

499, 2007.

Caire, I., JADE Tutorial: JADE Programming for

Beginners, TILAB & West Sussex, 2009.

Elmaraghy, Hoda A., Flexible and reconﬁgurable

manufacturing systems paradigms, International

Journal of Flexible Manufacturing Systems, Volume

17, Issue 4, pp 261-276, October 2005.

FIPA, Fipa.org, 2016. [Online]. Available:

http://www.fipa.org/. [Accessed: 06- Oct- 2016].

ICAART 2017 - 9th International Conference on Agents and Artiﬁcial Intelligence

198

Fontana, M., the Body Extender: A Full-Body Exoskeleton

for the Transport and Handling of Heavy Loads, IEEE

Robotics & Automation Magazine, Vol. 21, pp. 34-44,

2014.

Giret, A., Botti, V. Holonic Manufacturing Systems,

ANEMONA - A Multi-agent Methodology for Holonic

Manufacturing Systems, Springer Series in Advanced

Manufacturing, pp. 7-20, 2008.

Goodrich, M.A., Schultz, A.C., Human–robot interaction: a

survey, Foundations and Trends in Human–Computer

Interaction Vol. 1, pp. 203–275, 2007.

JADE | Java Agent DEvelopment Framework,

Jade.tilab.com. [Online]. Available:

http://jade.tilab.com/. [Accessed: 1- Oct- 2016].

Jennings, N.R., Wooldridge, M.J., Applications of

intelligent agents, Agent Technology: Foundations,

Applications, and Markets, Springer, pp. 3–28, 1998.

Koestler, A., The Ghost in the Machine, Hutchinson, 1967.

Koren, Y., Heisel, U., Jovane, F., Moriwaki, T., Pritschow,

G., Ulsoy, G., Brussel H., 1999, Reconfigurable

Manufacturing Systems, CIRP Annals, Vol 48, pp. 527-

540.

Kruger, K., Basson, A., 2015, Implementation of an Erlang-

Based Resource Holon for a Holonic Manufacturing

Cell, Studies in Computational Intelligence, Vol 594,

pp. 49-58.

Lasota, A., Rossano, F., Shah, A., Safe Close-Proximity

Human-Robot Interaction with Standard Industrial

Robots, IEEE International Conference on Automation

Science and Engineering (CASE), 2014.

Leap Motion, Leapmotion.com, [Online]. Available:

https://www.leapmotion.com/. [Accessed: 03- Oct-

2016].

Leitao, P., Restivo, F., ADACOR: A holonic architecture

for agile and adaptive manufacturing control,

Computers in Industry, Vol 57, pp. 121-130 , 2006.

Leitao, P., Restivo, F., Implementation of a Holonic Control

System in a Flexible Manufacturing, IEEE

TRANSACTIONS ON SYSTEMS, MAN, AND

CYBERNETICS—PART C: APPLICATIONS AND

REVIEWS, Vol. 38, pp. 699-709, 2008.

Lozano, J., Teba, J., Larrañeta, L., Álvarez, P., Dynamic

part-routing in a Flexible Manufacturing, Journal of

Operations Reaserch, Statistics and Computer Science,

Vol. 34, pp. 16-28, 1994.

Parker, L.E., Draper, J.V, Robotics Applications in

Maintenance and Repair, Handbook of Industrial

Robotics, 2nd Edition, J. Wiley, 1998.

Pinker, S., How the Mind Works, Annals of the New York

Academy of Sciences, Vol. 882, pp. 119–127, 1999.

Potter, L., Araullo, J., and Carter, L., "The Leap Motion

controller: A view on sign language," in 25th Australian

Computer-Human Interaction Conference, 2013.

Poslad, S., Specifying Protocols for Multi-Agent Systems

Interaction, ACM Transactions on Autonomous and

Adaptive Systems,

Vol 2, 2007.

Radu, F., Frank, F., Development and applications of

holonic manufacturing systems: a survey, Journal of

Intelligent Manufacturing, Vol 17, pp. 111-113, 2006.

Rethink Robotics | Advanced Robotics Technology |

Collaborative Robots, Rethink Robotics, [Online].

Available: http://www.rethinkrobotics.com/.

[Accessed: 02- Oct- 2016].

Shen, W., Hao Q., Yoon, H. J., Norrie, D. H., Applications

of agent - based systems in intelligent manufacturing:

An updated review, Advanced Engineering Informatics,

Vol 20, pp. 415-431, 2006.

Su, J., Component-based intelligent control architecture for

reconfigurable manufacturing systems, Ph.D. Thesis,

Virginia Polytechnic Institute and State University,

2007.

Teahan, W., Artificial Intelligence – Agent Behaviour,

BookBoon, 2010.

Van Brussel, H., Wyns, J., Valckenaers, P., Bongaerts, L.,

Petters, P. Reference Architecture for Holonic

Manufacturing Systems: PROSA. Computers in

Industry, Vol. 37, pp. 95–108, 2003.

Applying the PROSA Reference Architecture to Enable the Interaction between the Worker and the Industrial Robot - Case Study: One

Worker Interaction with a Dual-Arm Industrial Robot

199