A Robotic System for Home Security Enhancement

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/221338761

A Robotic System for Home Security
Enhancement
Conference Paper · June 2010
DOI: 10.1007/978-3-642-13778-5_6 · Source: DBLP

CITATIONS

READS

2

86

5 authors, including:
Andreas Gregoriades

Harris E Michail

Cyprus University of Technology

Cyprus University of Technology

53 PUBLICATIONS 313 CITATIONS

77 PUBLICATIONS 214 CITATIONS

SEE PROFILE

SEE PROFILE

Vicky Papadopoulou

Despina Michael

European University Cyprus

Cyprus University of Technology

49 PUBLICATIONS 288 CITATIONS

25 PUBLICATIONS 51 CITATIONS

SEE PROFILE

All in-text references underlined in blue are linked to publications on ResearchGate,
letting you access and read them immediately.

SEE PROFILE

Available from: Harris E Michail
Retrieved on: 30 June 2016

A robotic system for home security enhancement
Andreas Gregoriades, Samuel Obadan, Harris Michael, Vicky Papadopoulou,
Despina Michael
European University Cyprus
Nicosia, Cyprus.
University of Nicosia
Nicosia, Cyprus.
{[email protected], [email protected],
[email protected], [email protected]}

ABSTRACT: Central to smart home security is the need for adequate surrounding awareness.
Security systems have been designed for remote exploration and control, however, these still
lack the simplicity needed by elderly and disabled. The majority of elderly people find the
control of such systems laborious. This highlights the need for usable designs that take into
consideration the cognitive limitations of this category of people. This paper contributes
towards this problem through the introduction of a novel vehicular Remote Exploration
Surveillance Robot (RESBot), capable of monitoring in real time the environment in response
to events. The interaction with the system is achieved through natural language commands and
hence, provides improved usability over traditional approaches. Results from the experimental
usability evaluation of the RESBot system revealed considerable improvement over
conventional home security systems.

Keywords: Human-robot interaction, Situation awareness, Home security, Teleoperation.

1 Introduction
The worldwide population of elderly people is growing rapidly and in the coming decades the
proportion of old people will change significantly. This demographic shift will create a huge
increase in demand for domestic home security technologies. Smart homes have been credited
with saving the lives of many elderly, disabled and senior citizens [5]. The first generations of
home security systems are mainly using CCTV cameras. However, these provide limited
flexibility in maintaining sufficient situation awareness [7]. Plus, these systems are controlled
using keyboard and joystick that hence pose cognitive strain on elderly users. According to [18]
many CCTV is ineffective for day-to-day safety and security in general. Specifically, cameras
could be badly placed, broken, dirty, or with insufficient lighting and this limits their
effectiveness and efficiency. Moreover, when CCTV is used for command and control, safety
operators performance is hampered due to the large number of disparate systems and
information sources, and inefficient audio communication channels. This suggests that CCTV
for crime prevention can only be effective as part of an overall set of measures and procedures
designed to deal with specific problems. When it comes to elderly, the use of CCTV is
inappropriate for all of the above reasons along with the fact that elderly are characterized by
limited attentional resources that in effect constrain their capabilities with such systems. For
elderly and disabled people to maintain adequate situation awareness it is important to provide

them with only the salient cues from the environmental context. Too much or too little
information may have the opposite effect with regards to situation awareness. Therefore,
central to smart homes is the need for technologies that sufficiently address these requirements.
Prerequisite for adequate situation assessment is effective interaction with the technology.
Despite their success stories, smart homes have their limitations. It has been reported that smart
home technology will be helpful only if it‟s tailored to meet individual needs [2]. This currently
poses a problem as many of the interface control console designed do not take into
consideration non-functional limitations associated with the elderly. In the same vein, there is a
fundamental problem in making IT system user friendly for the elderly and disabled [19]. This
work is motivated by the need to improve the safety living conditions of elderly/disabled by
addressing two important factors that interest this category of people, namely, usability and
cost. The former prerequisites usable interfaces designed that improve users‟ task performance
through reduced cognitive effort. This requirement strive engineers in investigations for best fit
between man and machine by considering the constraints and capabilities of elderly people.
The later factor we address through a simple generic robotic platform.
Given that elderly and disabled are characterized by reduced memory and attention, the home
security system proposed in this study uses an interaction metaphor that minimizes user effort
and improves situation awareness. To that end, we adopted a human-centered approach for the
human-robot interaction through simple natural language instructions. The underlying
technology of the interaction metaphor is a mobile phone. The ubiquitous property and low
weight of mobile phones overcomes fundamental problem during interaction. The proposed
implementation utilizes a vehicular robotic surveillance system capable of indoor/outdoor
remote observation in response to natural language commands. This constitutes an
improvement over traditional approaches to home security management, characterized by
dexterous manipulation of joy stick and keyboard interfaces that obstruct the users from the
primary task. The research question that we address in this work is: “Does voice activated
robotic surveillance system provides improved situation awareness for elderly/disabled
compared to existing home security systems?”
The paper is organized as follows. Next section covers the literature on developments of human
centered robotics especially in smart home security technologies. Following this is a detailed
breakdown of the „RESBot‟ architecture, design methodology and software implementation.
Next, follows a usability evaluation of the proposed system and an interpretation of the results.
The paper concludes with future directions.
2 Literature Review
The last 10 years witnessed robots becoming increasingly common in non-industrial
applications, such as homes, hospitals, and service areas. These robots are often referred to as
“human-centred” or “human-friendly” robotic systems due to the way the robot interfaces with
the human users [15]. This close interaction can include contact-free sharing of a common
workspace or direct physical human-machine contact. Contrasts to industrial robots where
specific tasks are performed repetitively, human-centred robots are implemented on a totally
different set of requirement [16]. These include: safety, flexibility, mechanical compliance,
gentleness, and adaptability towards user, ease of use, communicative skills, and sometimes
possession of humanoid appearance and behaviour. According to Heinzmann and Zelinsky [17]
human centred robotics should have natural communication channels that involve not only
language but also facial gestures and expressions along with providing high level of
functionality and pleasure. Some successful human-centred robotic implementations include:
Rhino the museum tour guide robot that was assigned tasks via internet teleoperation technique
[20]. Among their other roles, robotic systems have been designed to aid home, industrial and
business security. An example of robot security management system is the Mobile Detection

2

Assessment and Response System (MDARS) [21] deployed by the department of defense.
MDARS simultaneously control multiple autonomous robots that provide automated intrusion
detection and warehouse inventory assessment. Similarly, robots are becoming popular in home
security and telecare [1-4]. Telecare is defined as the use of a combination of communications
technology and sensing technologies to provide a means of manually or automatically signaling
a local need to a remote service centre, which can then deliver or arrange an appropriate care
response to the telecare service user. AVENUE is an example telecare robot [23]. However,
current telecare and home security robots require dexterous manipulations [19] and hence,
failed to take into consideration the human-requirements of elderly and disabled. Efforts have
been reported that aim in easing the complexity of the interaction between elderly users and the
robot. Phonebot, is an example that uses cell phone communication between user and robot
[25]. Phonebot responds to calls via, a ring detector circuit, which establishes a connection
between the user and the robot. The main limitation of Phonebot is the lack of a voice activated
control mechanism and this creates a communication overload that obstructs the user from
achieving his/her goals [25]. This one of the limitations we address in this research.
An additional issue that has emerged in the field of human-robot usability is the notion of
enjoyment during interaction as reported by [26]. Specifically, [26] identified that there is high
correlation between enjoyment and intention to use of a robotic system in a usability analysis
conducted with elderly people. This indicates the need for enjoyment to be part of robot design.
This constitutes another motivating factor of the research conducted and described in here.
3 The Remote Exploration Surveillance Robotic System (RESBot).
RESBot is a roving maneuverable vehicular surveillance system, which projects in real time
contextual information of the environment for enhanced situation awareness of elderly and
disabled people. The RESBot is a voice activated control mechanism that enables remote
command and control for home security purposes. The main components of the system are
described below:

User navigation component

Autonomous obstacle avoidance component

Onboard pinhole surveillance camera.

Robotic vehicle
User navigation is achieved through a voice recognition component. This can be trained in any
language of choice and provides natural means for interacting with the system. This is of major
importance to elderly and disabled people that are characterized by low cognitive and motor
capabilities. Through this technology, we built a robotic system by incorporating existing
technology into a unified robotic surveillance system.
The robot‟s on-board controlled camera provides 360 degrees visual coverage and thereby
provides its users with the necessary contextual cues for improved situation awareness. For
brevity, the user is capable of interacting with the robotic system using voice commands over
GSM mobile phone. The visual feedback display enables the user to carry out real time
surveillance of a remote environment. In order to limit the scope of our work, we concentrated
on vehicle teleoperation that enable remote task execution.
The other main component of the system is the autonomous obstacle detection and avoidance
system that provide effortless navigation of the robot. Finally, the automatic intrusion detection
component provides real time recognition through a motion sensor. This component is
responsible for the aromatic orientation of the camera and the notification of the user of an
intrusion event. An overview of the system‟s architecture is presented in Figure 1.

3

Figure 1: Outline of an operational RESBot system

Figure 2: The RESBot implementation.
4. Design and Implementation of RESBot
The RESBot is a novel implementation of a voice activated robotic system leveraging the use
of modern GSM technology. The parallax Board of education (BOE) was used as the platform
for its development. The robot is a small, skid-steered wheeled vehicle capable of limited
outdoor/indoor work. The robot is equipped with a ring of infrared proximity sensors, power
monitoring, and voice recognition decoder. An onboard pinhole camera provides forward
video. Digital/Analogue transmitters are used for video and data communication. An on-board
micro-controller processor performs navigation and obstacle avoidance.
The communication between the user and the robot is achieved through GSM digital
transmission. GSM digital signals can pass through an arbitrary number of regenerators with no
loss in signal and thus travel long distances with no information loss [8]. Users perceive
situational cues from the environment through the onboard video feedback and accordingly
respond to them by engage the robot in a voice command. The command and control
mechanism is based on a user-centred interface design methodology in which visual displays
provide information for decision-making and control [9]. Through this we achieve easy robotuser interaction for adequate contextual awareness [10]. A wireless video camera (Figure 2) is
used to enhance motion planning. An onboard infra-red sensor provides situational cues that are
analyses by the central system and subsequently inform the user through an automatic events
generation.
The user can request status updated by invoking the robot at any time. The voice recognition
decoder is responsible for recognizing the user input and accordingly converting this into an
instruction for the robot‟s actuators. This component can be trained to recognize up to 40
instructions. After training, the user can instruct the robot using the specified language corpus.
During training the user can validate the system‟s knowledge by repeating a trained word into
the microphone. During the validation of this component, the index of the corresponding
human instruction was displayed on a digital display. This helped to verify that the right
instructions are associated with the right indexes. Output from the voice recognition circuit is
fed into the robot‟s microcontroller for navigation. To achieve a more robust interaction
between user and the robot, words with the same meaning were associated with the same action

4

in the language corpus. Central to the robot is the microcontroller that processes human
instructions for command and control. This component is responsible for the rotation of the on
board camera during surveillance mode. Recognition of an event raises an alert to the user that
accordingly navigates the robot and orients the camera. The robot is also equipped with an
obstacle detection and avoidance algorithm that enables its autonomous navigation. To achieve
this feature, three infrared sensors are used. Intrusion detection is achieved through an on board
motion detector sensor. For the rotation of the camera and the wheeling of the robot sstandard
and rotational servos are used.
5 Usability evaluation of the RESBot
Core to the success of the proposed home security system is adequate usability, given that the
indented users have special needs and attentional constraints. Therefore, it was essential to
primarily understand prospective users‟ relevant skills and mental models and accordingly
develop evaluation criteria. However, robotics systems differ significantly from desktop user
interfaces and hence, the use of empirically defined set of heuristics such as Nielsen are not
suitable. To that end, it was decided to use an experimental usability evaluation approach in
controlled settings. Typical HCI evaluations use efficiency, effectiveness, learnability and user
satisfaction as evaluation metrics. Efficiency is a measure of performance time; effectiveness is
a measure of task performance; learnability is a measure of how easy is to learn a system by a
novice user and satisfaction a measure of pleasure. Learnability is considered as a key indicator
for usability in human-robot interaction with elderly and disabled since it significantly affect
system acceptance. Learnability incorporates several principles like familiarity, consistency,
generalizability, predictability, and simplicity. These four measures were deemed appropriate
for the experimental evaluation of the RESBot system and its comparison against the most
popular home security system, namely CCTV.
To benchmark the proposed system against available low budget home security systems such as
CCTV, we opted for a comparative study using two groups of participants. Each group was
composed of 10 participants of age 60 and over. Throughout the study, CCTV and RESBot
systems were referred to as System A and B respectively. Prior to the experiment each
participant had to complete a consent form. Participants of each group were given instructions
and demonstration on how to use each system through example scenarios. During training,
participants were encouraged to ask questions. The evaluations involved two experiments
using system A and B accordingly. The scenario of each experiment, involved participants
independently investigating an alarm caused by a human intruder. The experiments took place
in the participants‟ homes and a TV was used for video feedback. For system A, a static camera
was mounted in the participant‟s garden. The CCTV camera was also equipped with a motion
sensor that raised an alarm whenever it sensed motion within the covered area.
The first experiment was conducted with the use of the CCTV system and group A. The
evaluation scenario required users to locate the hiding position of an intruder that was allowed
to move around at will. In this experiment, it was assumed that intruder was not aware of the
position of the CCTV camera. During the experiment, participants were asked to locate the
intruder and specify his hiding position. The second experiment was conducted with the use of
the RESBot system and group B. This group did not participate in the first experiment. Prior to
the use of the RESBot system, participants were expected to train its instructions corpus.
During this experiment, the RESBot was placed in the surroundings of each participant‟s home
and the goal was to locate an intruder. Participants were informed of the intrusion through the
system alarm. The manipulation of the RESBot was performed with the aid of a mobile console
using voice commands. Visual feedback from the robot was projected on the TV. In both
experiment the scenario was terminated with the recognition of the intruder‟s position.

5

Throughout the experiments, participants‟ actions and mistakes were recorded by the
researchers along with their tasks and task‟s completion times. To assess users‟ situation
awareness, we opted for the SAGAT approach described by Endsley[6]. Hense, during the
scenarios, participants were asked to designate at different intervals, where they though the
intruder was located. The actual and perceived location of the intruder enables the
quantification of their situation awareness, that in turn guided their consequent
actions/instructions with the system. With the completion of both experiments, each participant
was asked to complete a questionnaire. This included constructs relating users‟ perceptions of:
ease of use, usefulness, attitudes, and intention to use of the system. In addition, behavioral
information regarding satisfaction and user experience with both systems, were also collected.
Each participant‟s response was associated with a unique identification number to avoid bias
during data analysis.
The evaluation of the two systems’ was based on the level of user-acceptance, the assessment
of which was based on the Technology Acceptance Model (TAM) [12]. TAM has been
successfully used to study user’s acceptance of IT systems using quantifications of users
attitude that define the positive or negative feelings toward the IT system. In its most basic
form it states that perceived usefulness and perceived ease of use determine the behavioral
intention to use a system that can predict actual use. Specifically:

Perceived ease of use (P1): defines as the degree to which an individual believes that
learning to use a technology will require little effort. The participants’ perceptions
that system (A or B) was easy to use were captured with 12, five-point Likert-scale
questions.

Perceived usefulness (P2): examines individual believes that use of the technology
will improve performance. The participants’ perceptions of systems’ usefulness were
captured in 8, five-point Likert-scale questions

Attitude (P3): feeling or emotion about using the technology. The participants’
attitudes towards using the system (A or B) were assessed using 5, five-point Likerttype questions.

Intention to use (P4): the likelihood that an individual will use the technology in the
future. The participants’ intention to use each system was assessed using 4 five-point
Likert-type questions.
Questionnaires were used as the main instrument for measuring these influences. Each
component of the TAM model was expressed in a number of questions. Items of the instrument
that measured the same influence were grouped together to form generic constructs. In addition
to the core TAM constructs, an extra set of questions regarding user satisfaction were also
incorporated in the research instrument.
5.2 Analysis and Results
Data collected from the experiments was analyzed using the statistical package SPSS.
Comparative analysis between the two systems was performed using a 2-tailed t-test for each of
the constructs of TAM. Since we have no strong prior theory to suggest any relationship
between the TAM components of system A and B, we opted for the 2-tailed t-test. The analysis
performed a paired samples t-test to check the difference between the scores in each dimension.
The difference between the two paired mean scores for TAM constructs ranged from P1-P4. P1
compared perceived usefulness of the two systems. P2 compared perceived ease of use, while
P3 and P4 attitude towards use and intention to use respectively.
The collated results of table 1 highlight that for each of the TAM constructs, the mean scores
between systems A and B differ significantly. In particular, perceived usefulness of system B is
significantly higher than system A. Similarly, perceived ease of use, attitude and intention to

6

use is significantly higher for system B. The 2-tailed test indicates that the difference between
the two systems is significant and the results are not due to chance.
Table 1: Paired sample t-test
Pair
P1
P2
P3
P4

Mean
-12.800
-12.600
-12.100
-3.700

Std Dev
2.616
5.037
2.282
2.057

t
-15.472
-7.909
-16.762
-5.687

df
9
9
9
9

Sig
.000
.000
.000
.000

The results shows that the robotic system provides better usability and control over the CCTV
system by better addressing the needs and capabilities of the elderly and disabled.
Additional data regarding the usability of the robotic system revealed that 90% of participants
could use and interpret the feedback from the visual display with ease. Moreover, 90% of
participants feel that the alarm produced by the system is suitable for attracting attention.
Overall, the results yielded from the experiments demonstrated that:
 80% of participants were able to navigate the robot with the voice recognition system.
 70% of the participants spotted the intruders‟ changing position faster with the RESBot
than the CCTV.
 Participants were not able to get the system to respond to the „survey‟ and „hold-on‟
commands using RESBot. Nevertheless, the other commands were sufficient for
effective navigational surveillance.
 90% of participants found customizing the commands with easy to remember words in
their preferred language very satisfying
 Throughout the experiment 20% of participants complained about issuing a command
more than once over the control console, for the robotic response.
 100% of the participants were satisfied with the latency (reaction time) between a
recognized command and robotic response.
 90% of the participants agreed they had an enjoyable, engaging and satisfying
experience using the RESBot.
 80% of participants managed to locate the intruder in less time using the RESBot.
 70% of participants had better situation awareness with the RESBot.
Conclusion/Future direction
As smart homes implementation gradually become widely adapted due to increasing aging
population, the need for improving the safety living conditions of the elderly/disabled increases
along with it. Similarly, there is a stressing need for usable human-robot interface designs that
meet the needs and constraints of elderly and disabled [6]. In this research, we demonstrate the
implementation of a vehicular robotic surveillance system (RESBot) capable of remote
surveillance of the environment in response to natural language commands. This is an
improvement over keyboard and joystick controls and hence, provides a more usable
communication metaphor between elderly users and robotic systems. The study also
highlighted an important issue in the literature by examining the effects of robotic systems
operated by elderly people for enhanced home security through improved contextual
awareness. The current research results provide key information to educators and commercial
industries in providing a more robust security implementation of smart homes.
Future direction will be geared towards incorporating audio processing capabilities to address
the needs of users with speech impediment. A possible implementation could also be the
combination of visual and voice interface in a single control console such as PDAs.

7

REFERENCES
[1]. Kinsella A. Home telecare in the United States. J Telemed Telecare 1998; 4: 195-200.
[2]. Doughty K, Cameron K, Garner P. Three generations of telecare of the elderly. J
Telemed Telecare 1996; 2: 71-80.
[3]. Ruggiero C, Sacile R, Giacomini M. Home telecare. J
Telemed Telecare 1999; 5:
11-17.
[4]. Stoecke JD, Lorch S. Why go see the doctor? -- Care goes from office to home as
technology divorces function from geography. Int J Technol Assess Health Care
1997; 13: 537-546.
[5]. Cheek, Penny. (2005). “Aging Well With Smart Technology”. Nursing
Administration Quarterly. Vol. 29, No. 4: 329-338.
[6]. M. R. Endsley, “Design and evaluation for situation awareness enhancement,” Proc
Human Factors Society 32nd Annual Meeting, Santa Monica, CA, 1988.
[7]. T. Sobh, R. Mihali, G. Gosine, P. Batra, A. Singh, S. Pathak, T. Vitulskis and A.
Rosca, S. Grodzinsky, L. Hmurcik. Web-Controlled Devices and Manipulation: Case
Studies In Remote Learning for Automation and Robotics. World Automation
Congress, 2002
[8]. Terry Fong. Collaborative Control: A Robot-Centric Model for Vehicle
Teleoperation. The Robotics Institute Carnegie Mellon University Pittsburgh,
Pennsylvania 23 January 1998

[9]. Newman, W. and Lamming, M., “Interactive System Design”, AddisonWesley, Reading, MA, 1995.
[10].Robert Riener, Dr-Ing, Lars Lünenburger, Dr rer nat, Gery Colombo PhD Humancentered robotics applied to gait training and assessment. Journal of rehabilitation
research and development. Volume 43, number 5, pages 679- 694. August/September
2006
[11].Zinn M, Roth B, Khatib O, Salisbury JK. A new actuation approach for human
friendly robot design. Int J Robot Res. 2004;23(4–5):379–98.
[12].Heinzmann J, Zelinsky J. A safe-control paradigm for human-robot interaction. J
Intell Robot Syst. 1999;25(4): 295–310.
[13].Toprani, K. Smart houses become reality. The Triangle Online. December 10 2004
[14].Wolfram Burgard, Sebastin Thrun, Dieter fox, Armin B. Cremers, Dirk Harnel,
Gerhard Lakemeyer, Dirk Schulz, Walter Steiner, Experiences with an interactive
museum tour-guide robot. Artificial intelligence 1999, pp. 1-53
[15].Zeng Dehuai; Xie Cunxi; Li Xuemei. Design and Implementation of a security and
patrol robot system. Mechatronics and Automation, IEEE International Conference
Volume 4, Issue, 2005 Vol.4 Digital Object Identifier.
[16].P. Allen, I. Stamos, A. Gueorguiev, E. Gold, and P. Blaer. AVENUE: Automated
site modeling in urban environments. Proceedings of 3rd Conference on Digital
Imaging and Modeling in Quebec City, Canada,May 2001, pp. 357.364.
[17].Marcel Heerink, Ben Kröse, Bob Wielinga, and Vanessa Evers Enjoyment, Intention
to Use And Actual Use of a Conversational Robot by Elderly People. Hogeschool
van Amsterdam, Instituut voor Information Engineering Almere, Netherlands year
2007
[18].M Sasse, Not seeing the crime from the cameras?, Communications of the ACM,
53(2), 2010.

8