Conference PaperPDF Available

Abstract and Figures

Search and Rescue (SAR) is a critical component of disaster recovery efforts. Every second saved in the search increases the chances of finding survivors and the majority of these teams prefer using canines [5]. Our goal is to help enable SAR dog and handler teams to work together more effectively. Using a semi-structured interviews and guidance from K9-SAR experts as we iterate through designs, we develop a two-part system consisting of a wearable computer interface for working SAR dogs that communicates with their handler via a mobile application. Additionally, we discuss the system around a heuristic framework that includes dogs as active participants. Finally, we show the viability of our tool by evaluating it with feedback from three SAR experts.
Content may be subject to copyright.
Search and Rescue: Dog and Handler Collaboration
Through Wearable and Mobile Interfaces
Clint Zeagler
Georgia Institute of Technology
Ceara Byrne
Georgia Institute of Technology
Giancarlo Valentin
Georgia Institute of Technology
Larry Freil
Georgia Institute of Technology
Eric Kidder
Georgia Institute of Technology
James Crouch
Georgia Institute of Technology
Thad Starner
Georgia Institute of Technology
Melody Moore Jackson
Georgia Institute of Technology
Search and Rescue (SAR) is a critical component of
disaster recovery efforts. Every second saved in the search
increases the chances of finding survivors and the majority
of these teams prefer using canines [5]. Our goal is to help
enable SAR dog and handler teams to work together more
effectively. Using a semi-structured interviews and
guidance from K9-SAR experts as we iterate through
designs, we develop a two-part system consisting of a
wearable computer interface for working SAR dogs that
communicates with their handler via a mobile application.
Additionally, we discuss the system around a heuristic
framework that includes dogs as active participants. Finally,
we show the viability of our tool by evaluating it with
feedback from three SAR experts.
Author Keywords
Canine Interfaces, Dog to Handler Communication, Search
and Rescue System
ACM Classification Keywords
H.5.m. Information interfaces and presentation (e.g., HCI):
Recent research in animal-computer interaction has
catalyzed the creation of computer-aided systems that could
allow humans and dogs to work together more transparently
and effectively while conducting Search and Rescue (SAR)
missions. SAR dog teams are critical for locating missing
individuals in the wilderness and in the aftermath of natural
disasters or mass casualty events. Search and rescue dogs
detect human scent in conditions unfavorable to human
vision, such as low visibility environments (dark or highly
obstructed) and at far distances. Their scent-detection
capabilities, their agility at speeds greater than their human
counterparts, and their ability to hear at higher ranges, help
increase the efficiency and success rate of non-canine SAR
Although all dogs have superior olfaction compared to
humans, well-trained dog-handler pairs are essential for
success due to the strategy required for each search. SAR is
currently a highly manual job and a single person typically
works the dogs on foot. Our focus in this work is dedicated
to studying systems aimed at improving this interaction.
Technology such as GPS tracking units [6, 24, 25] are used
by handlers and command centers, called ‘incident
command,’ to track dogs in the field. This is especially
helpful when the SAR dog leaves the handler’s line of
sight. Bozkurt et al also describe a wearable computing
system for SAR dogs to monitor the dog and the
environment, where the dog could even be followed by
drones [2]. This could be a very effective way at keeping
SAR dogs safe while working. These systems and ones like
them are focused on giving the handler more information
about the dogs location and wellbeing, but does not allow
for the dog to send more information back to the handler.
The system we propose in this paper tackles this problem,
giving the trained SAR dog the ability to send a message
saying, “I’ve found something importantor “Please come
and investigate this.” This type of interaction and ability
augments the current SAR model, possibly making it more
transparent, efficient, and effective.
Search and rescue is an activity defined as “the search
for people who are in distress or imminent danger”
although in some cases, cadaver searches are also
grouped within this category [5]. Although specific
definitions vary by jurisdiction, the general field of
search and rescue includes many specialty areas typically
characterized by the type of terrain over which the search
is conducted. These areas include:
Mountain/wilderness rescue
Ground search and rescue
Permission to make digital or hard copies of all or part of this work for
personal or classroom use is granted without fee provided that copies are
not made or distributed for profit or commercial advantage and that copies
bear this notice and the full citation on the first page. Copyrights for
components of this work owned by others than ACM must be honored.
Abstracting with credit is permitted. To copy otherwise, or republish, to
post on servers or to redistribute to lists, requires prior specific permission
and/or a fee. Request permissions from
ACI '16, November 16-17, 2016, Milton Keynes, United Kingdom ©
2016 ACM. ISBN 978-1-4503-4758-7/16/11…$15.00
Urban search and rescue in cities (USAR)
Combat search and rescue on the battlefield
To establish user needs and to get feedback on our
proposed SAR system we held semi-structured
interviews with nine professional volunteer SAR
SAR Task Analysis
Task Environment
Disaster dogs are used to locate victims of catastrophic or
mass-casualty events (e.g., earthquakes, landslides, building
collapses, aviation incidents). Although SAR dogs can be
used in urban environments, such as large, controlled
settings (hospitals, factories, and airports) we will focus on
SAR dogs that work outdoors. In addition to rescue and
recovery, outdoor SAR dogs can also perform wilderness,
disaster, cadaver, avalanche, and drowning searches [1]. In
wilderness SAR applications, air-scenting dogs can be
deployed to high-probability areas. These are places where
the subject or target have the potential to be, or places
where the subject's scent may collect, or “pool”, such as in
drainage canals in the early morning. Tracking/trailing dogs
can be deployed from the subject's last known point (LKP)
or the site of a discovered clue [3]. Handlers must be
capable of bush navigation, wilderness survival techniques,
and must be self-sufficient. The dogs must be capable of
working for 48 hours without distraction (e.g., by
wildlife). Disaster dogs rely primarily on air scent, and may
be limited in mass-casualty events by their inability to
differentiate between survivors and recently deceased
In the state of Georgia, our expert contacts have stated that
most searches occur outdoors and the majority of dogs are
exclusively trained for wilderness searches. Since Georgia
doesn't have a location for specifically training urban
disaster dogs, most volunteers in the state strictly search
wilderness. Even when working in the wilderness, it is
sometimes necessary to rule out a site without buildings.
Examples include ensuring a lost child isn't hiding inside a
house, scanning a college campus for a specific person, or
checking barns and shelters when searching for someone
who may be lost outdoors seeking shelter. However, more
commonly, SAR teams look for hunters or hikers who are
lost in an area away from any structures. The probabilities
of indoor or outdoor work varies significantly by area and is
difficult to estimate without further research. For training
purposes, our expert consultants practice searching
buildings about 15-25% of the time.
Weather is an important factor that influences scent travel.
Scent molecules are typically denser than air, but less dense
than water. As a result, temperature can result in “lifting
and dropping” the scent, leaving "pools", and the humidity
can “bog down” scents. These phenomena can be studied
using smoke to see the impact of air flow (indoors and out)
and effects of temperatures on the terrain.
Generally, dogs are called out right before the weather
conditions are expected to deteriorate. This is because the
human search parties refuse to continue at this point and
calling the dogs is the last and only alternative.
The direction of the wind is also an important factor in
determining search plans. Teams typically start searches
downwind and “grid” across the wind to allow dogs to
perceive the scent as it's blowing. If a complete search is
performed upwind from a victim, teams are likely to end
up going past the victim and then having to backtrack
when the dogs detect the scent downwind. Dogs detect
which nostril perceives the most scent and use this to
determine which way the scent is traveling towards
them. Nevertheless, winds in the southeastern United
States change directions frequently, so teams must take
this into account.
SAR Task Scenario Without Technology
An active search and rescue could have many different
outcomes depending on many different variables. We have
produced a complete task map too detailed to be outlined
here [20], however, we will present a typical search and
rescue scenario.
Figure 1: A SAR dog leaves his handler’s sight and returns to
alert he has found someone, “ping-ponging” back and forth.
Brady&is& a&SAR&volunteer;& he&has&a&dog,&Fancy,&who&is&
trained& on& air-scent& searching.& & The& local& police&
department& calls& for& Brady’s& help& when& a& child& goes&
missing& from& a& nearby& playground.& When& Brady& arrives&
at& the& playground& he& is& given& a& map& with& an& area& the&
police& would& like& for& him& to& search.& The& police& give&
Brady& a& teddy& bear& owned& by& the& child& so& that& Fancy&
can& smell& the& child’s& unique& scent.& Brady& checks& the&
wind& direction,& gets& out& his& compass,& and& sets& off&
with&Fancy&to& search&the& nearby&woods& starting&from& a&
downwind&location.&& Fancy&runs& out& of&Brady’s& line&of&
sight& and& Brady& hears& barking& in& the& distance,& which&
Fancy’s& alert& that& she&has& found& someone.& Fancy&
returns& to& Brady& and& barks&to& confirm&she& has& found&
the& search& subject.& Brady& follows& her,& but& when& they&
arrive& at& the& location& no& one& is& there.&This& happens&
quite& often& when& searching& for& a& child,& and& Brady&
knows& that& some& children& are& afraid&of& dogs& and&
continue& to& run.& After& Fancy& has& ping-ponged& back& and&
forth& between& Brady& and& the& child,& Brady& finally& sees&
the& child& and& calms& him& enough& to& bring& him& back& to&
Current SAR Workspace
It is important to understand how a larger working system
will operate before designing a new piece of technology.
Insights from our expert SAR volunteers guided our
participatory design approach..
Interview Participants: K9-SAR Experts
We interviewed nine K9-SAR experts ranging in SAR
experience and expertise. Our participants are part of a
larger organization that handles incoming SAR calls in the
Metro-Atlanta area and reaches out to the individuals with
the relevant experience necessary.
Characteristics of Users
Through interviews with our SAR experts we established a
set of basic characteristics of the community of users we are
targeting. From this, we gathered that there are both male
and female users. The user population has a very diverse
age range, from young enthusiasts (older than 16 years) to
agile senior volunteers with their trained SAR dogs. For this
reason, it is important to keep the system informative, but
flexible for the very broad range of people who would use
the system.
One common attribute is that all of the people volunteering
in an SAR operation have to be certified in SAR through
programs such as the National Association for Search and
Rescue (NASAR) or the North American Police Work Dog
Association (NAPWDA) and many have held or continue to
hold occupations useful for on-site crises many are
Registered Nurses, Firefighters, or Paramedics. They gain
valuable information during the SAR training process, but it
also gives the administrator an opportunity to fully explain
a new SAR Handler Application. We can expect all of our
users to be able to read in English, and understand a
geophysical map. We can also expect all of our users to be
physically capable of performing the task of searching
outdoors for extended periods of time.
The current social context of Search and Rescue specialists
is heavily influenced by interactions with first responders.
In most jurisdictions in the United States, local police
officers or park rangers conduct the first attempt at SAR.
Once this approach has failed, or the human parties are
unable to continue a search, the Search and Rescue teams
are contacted for support. In some situations and
jurisdictions, it may take hours or days before the teams are
contacted. As such, Search and Rescue teams believe that
their work would benefit from closer collaboration with
local authorities.
Current Technology Use
The current use of technical systems, such as GPS, to assist
in search and rescue varies greatly with each team. Most
SAR teams are comprised of voluntary members and
technology is not provided by the organizations they belong
to. SAR teams that receive technology grants typically
obtain GPS receivers and handheld radio equipment in
addition to smartphones. The radio equipment and phones
are used to communicate with other teams in the near
vicinity to coordinate their efforts. Teams that cannot or
prefer not to use computing technology rely on paper maps
and compasses.
Our group of SAR experts also uses the TNP Terrain
Navigator Pro [21] to help mark areas they have searched
and to produce reports they are required to provide to law
enforcement once the search is complete. They use One
Call Now to call out to their SAR volunteer phone tree. The
also keep an e-copy of Lost Person Behavior [14] on their
mobile devices for reference as they search. Other useful
mobile applications they listed were:
Koredoko - provides the GPS coordinates from a
picture taken on a cell phone. If a person is lost,
but has a cellphone, they take a picture and text it
to law enforcement. The image is opened with
Koredoko and it opens a Google map with the
location of where the photo was taken, with
latitude and longitude coordinates, along with
additional useful information.
Map Tools: coordinate system conversion between
latitude/longitude, Universal Transverse Mercator,
and National Geodetic Survey coordinates.
Microsoft OneNote: note-taking and freehand
Evernote: note taking, photo integration into a
Google maps: location identification and
directions to a search area, looking at past satellite
views of an area
In our discussions before, we broached the topic of canine /
human technology mediated collaboration, many of the
SAR experts wanted to focus on methods to ease their
workflow through a SAR system combining their existing
tools. They wanted to include their Pack check list as well,
which is a checklist they use in packing their equipment
before they leave for a search. They described how helpful
it would be for this system to pre-fill a majority of the
information on the Incident Command System (ICS) Form
[10], which is a form that the SAR team has to fill out to
give law enforcement after their search. Much of the needed
information to fill out the ICS Form is capable of being
collected by mobile devices while in use.
On the canine side, common devices include a padded stick
known as a bringsel [9]. This stick is attached to the collar
of an SAR dog in a manner that allows it to swing freely.
When the target is found, the dog bites the bringsel and
holds it in his mouth until returning to the handler. In this
way, the dog effectively communicates the completion of a
successful search.
The following ten usability criteria were isolated from work
by Nielsen [18]. In addition, we have provided an
interpretation of how we might implement these criteria to a
proposed digital SAR dog and handler communication
system, based upon information gathered from our SAR
literature review and expert SAR handler interviews.
1) Visibility of system status: Handlers should be aware of
their dog’s location(s) at all times; similarly, the system
should be able to provide the dog with appropriate feedback
at the appropriate moments.
2) Match between system and the real world: Cardinal
directions should be maintained by connecting generated
maps with compasses. The system should provide
affordances appropriate for dogs. Information to the canine
and handler should be natural and in a logical order.
3) Users control and freedom: Our system should support
the ability to navigate menus quickly by incorporating undo
and redo buttons on every interface. It should also interact
well with other applications since users will need to swap
between different systems (e.g., radio) rapidly.
4) Consistency and standards: Our design should follow
Gestalt psychology principles, mimicking popular current
analogies such as Google Maps/GPS, and leverage existing
SAR standards.
5) Error prevention: Because many decisions must be made
with haste and while in motion, the transparency of actions,
such as clicking ‘yes’ or ‘no’, is important and providing
confirmation of actions, for instance, “Are you sure?” helps
manage errors.
6) Recognition rather than recall: The system should
minimize the human and canine’s memory load by making
relevant objects, actions, and options visible.
7) Flexibility and efficiency of use: The interface should
allow for both experienced and inexperienced users by
handling “accelerators” or tools that can speed up
8) Aesthetic and minimalist design: Screens should not be
cluttered with obtrusive or unnecessary information and
should maximize the screen utilization for relevant
information, such as maps. Since speed plays a pivotal role
in SAR missions, graphical quality should be sacrificed in
favor of functionality. However, decreased accuracy is the
9) Help users recognize, design, and recover from errors:
User interfaces should be sufficiently detailed such that
users do not get caught in endless loops that are costly to
efficiency and success.
10) Help and documentation: Whenever the application
fails or if the user needs assistance, access to external
resources (e.g., emergency phone numbers) should be
There is a quite a bit of current work surrounding animal-
computer interaction [1517], specifically working dogs
and their interaction with technology. Robinson and
Mancini describe a domestic system where trained
assistance dogs interact with mounted pull tabs and other
sensors to call for help in case of a home emergency [19].
Zeagler et al’s motivation for studying a dog’s ability to
interact with an IR touchscreen is quite similar. They were
also interested in understanding how to better create
Touchscreen systems for canine interaction [26]. Within the
domain of SAR dogs, Ferworn et al augmented dogs with
video capture for monitoring while searching areas unsafe
for humans [7, 8], and Tran also works to collect
information from the environment around the SAR dog
[22]. In any case, assistance dogs are being trained to
effectively interact with technology to support the dog’s
working life. On-dog-body interfaces for dogs to use while
wearing their service vest, designed by Jackson et al, have
been tested and found to be quite effective [11]. One such
sensor, the capacitive bite sensor [12], was chosen to use as
our dog interface for communicating to the human handler.
The working SAR, wearable computer communication dog
vest components include a Capacitive Bite sensor [12]
(figure 4) in order to detect when the dog bites.
Additionally, a central hub with Bluetooth Radio is used to
broadcast sensor activation feedback tones to the dog and
activate alerts to the cellular phone. The cellular phone
tracks the precise GPS location of the dog, and send the
alerts from the Capacitive Bite sensor to the backend server
API. The components are sealed in an OtterBox
protective case to protect the electronics from water (figure
5). Furthermore, the cell phone is an S4 Active, which is
designed to be water resistant.
The vest also has an additional capacitive sensor chest
strap, which was designed to prevent false-positives if and
when the dog traveled through water. Since the dog would
not intentionally activate the chest strap, it works to detect
if the dog is in water, thereby deactivating the side-mounted
capacitive sensor used for alerts. Shortly after the dog has
left the water, the chest strap deactivates and re-enables the
alert sensor.
The construction of the vest and the materials chosen reflect
the type of activity, the wearer, and the environment. A
Julius K9 powerharness [13] is augmented with extra foam
padding to stabilize the vest while in use and to take
pressure off the dogs spine (figure 6). The location of the
on-body dog interaction was informed by previous research
in canine reachability of snout-based wearable inputs [23].
Figure 3: Capacitive sensor operation[12]
Figure 4: SAR working wearable computer vest
Figure 5: SAR working wearable computer vest
Figure 6: SAR working wearable computer vest
Dog Activation
During a typical SAR scenario, the working dog is trained
to alert the handler that he has found something of interest.
In this case, the dog would bite the capacitive sensor at his
side (Figure 7). Notice the bite sensor is the same shape and
size of a bringsel, which follows our second heuristic Match
between system and the real world. Upon activation, the
dog’s system beeps a tone, notifying the dog that he has
made a successful activation (the dog is trained to
understand this tone as a reward marker, and as the
completion of a task). Also upon activation, the dog’s vest
would send a signal through the cell phone to the handler’s
smart phone, including the GPS data and activation
Figure 7: Border collie activating a capacitive sensor [12]
Figure 8: SAR Handler Interface showing dog location,
handler location, wind direction and compass.
Figure 9: SAR Handler Interface showing a dog activation,
and also a note entered by the handler.
Figure 10: Handler Interface icons and meanings
The human interface for monitoring the dog’s location and
activities is an application that runs on an Android platform.
The application works on all GPS-enabled Android cell
phones and tablets, but currently requires a device that can
pick up cellular service while in the field. The SAR handler
application (Figure 8 & 9) shows the location of the dog
with respect to the handler’s location on a map. The
application also shows a compass and general wind
direction, as gathered from local weather data. What differs
about the program from other GPS tracking systems is the
ability of the system to mark searched locations and receive
input from the SAR dog.
As the SAR dog moves, a path is drawn on the map
showing where the dog has searched (Figure 9). The path is
set to a width advised by our SAR experts. After we
inquired the distance from the dog the experts would feel
comfortable saying the dog had searched, the width of the
path search was set to a total of 10 meters wide. This
distance from the dog is also a conservative translation of
an air-scent trained SAR dog’s ability when working in
hilly forested terrain. If an SAR dog finds something he
believes he is looking for, he will be trained to activate his
capacitive bite sensor. Upon activation, the handler
application will mark the point of activation and continue to
follow the activated dog’s path. This allows the dog to
continue without returning to the handler (which is helpful
in our earlier scenario of a frightened child). The path is
displayed as red until the dog returns to the handler. As the
handler follows the dog, the handler can leave GPS-tagged
notes on any points of interest or they can document context
clues seen by the handler along the path (Figure 9).
The menu buttons along the bottom of the screen (from left
to right) allow the map to be toggled through
street/satellite/hybrid views, allow the ability to create a
note, allow the handler to re-center the map on themselves,
allow them to stop the program, and allow them to clear the
map. Adding a note, stopping, and clearing the map all have
error prevention pop up confirmation screens.
To test our SAR system, we asked one of the volunteer
SAR handlers to train a dog using our vest. We conducted
the prototype feasibility test on a 72-acre, fenced property
containing fields, paths, and wooded areas. We had one of
the members of our team hide, and the SAR volunteer used
the vest and dog to search. We were able to follow the
dog’s path as expected. When the dog activated the system,
it responded as designed. As this was a prototype test of the
system, our goal was to gain handler feedback, and as such,
when found, the team member prompted the dog to activate
the sensor. We were not trying to test the dog’s capability
of using the system at this time, although the dog’s ability
for sensor activation was tested in a previous study [12].
During the course of the prototype test the handler was able
to clearly understand the mobile interface.
The handler also directed the dog to run/swim across a
small lake, after which the system again responded as
designed. The capacitive chest strap stopped the bite sensor
from sending false positive signals, and everything began to
work like normal once dry. The more compact version of
the SAR vest, as seen in Figures 4-6, is not waterproof, the
lake test was performed with an earlier and larger, but
waterproof vest with the same capabilities.
The first version tested was quite larger in scale. The
OtterBox was double in size and the bite sensor was
larger and hung lower on the dog’s torso. After feedback
from the SAR handler we knew we had to make the system
more compact for it to be viable. The handler’s main
concern was the vest catching when the dog had to run
through dense brush. The current model is now much
smaller and compact, but our goal is to continue to reduce
its size.
Additionally, to evaluate our system, we shared both the
handler and canine interfaces with three expert SAR
handlers having extensive experience in dog tracking
(approximately 9 years). All three SAR handlers, who shall
be referred to as E1, E2, and E3, received the system with
positive reviews and provided constructive feedback on
how to make the system more robust for fieldwork. In the
words of E3, “overall, very impressed with the product.”
SAR Dog Vest
While we show through our testing and iterations that the
system was effective for canine interactions, E1 and E3
recommended improvements for durability, visibility, and
connectivity to the handler.
In the words of E3, it’s incredibly important that “all
equipment, cabling, handles, straps, and material must be
Georgia-briar-patch-proof, water/humidity-proof, and heat-
proof (110+ degrees F), otherwise it won't last a summer.
E1 recommended making the vest lighter and more
breathable, saying that their canines were not accustomed to
working with vests on. While this is not the case for all
SAR handlers, the vest should not overheat the canine while
they are working.
Furthermore, all three experts were worried about the
footprint of the vest and the possibility of it snagging on the
brush while their dogs were working. While the Capacitive
Bite sensor doesn’t need to be tugged, E3 mentioned that
QASM (Quick Attach Surface Mount) connectors should
be able to withstand up to 200lbs of pulling force -- I've
seen trailing dogs nearly pull an adult male off his feet
when starting a search.” E3 also recommending building
the Styrofoam pads into the vest so that the structure of the
vest remains, but that they don’t come off while working.
E3 requested that the vest be recognized as a working dog
vest and, as such, come in colors related to safety work,
such as safety orange or green. Additionally, it is important
for handlers to be able to put appropriate and easily
identifiable Service Dog badging on their vests.
E3 had a few concerns regarding how the vest connected to
and relayed information to the handler’s device. Our
decision to use Bluetooth limits us to approximately 100
feet, and E3 recommended using longer range protocols,
such as cellular. Additionally, E3 requested that the SAR
dog vest record information locally in the event that
connections were not stable so that their paths were still
logged, but so that they could be pushed to the handler’s
device once connectivity was restored.
Mobile Human Interface
The mobile human interface also received positive
feedback, however they provided constructive criticism
regarding mapping, iconography, and annotation.
All the experts highlighted the fact that the mapping
functionality was the most important aspect of the handler
interface. “Mapping is critical to the search and anything
that obscures the minimal space we have on the screen will
detract from the benefit of having that screen,” said E3.
Additionally, terrain maps were preferred over Google’s
street-optimized view.
Secondly, there were concerns using latitude and longitude
as the primary means of displaying coordinates. E3
mentioned “Lat/Long is good, but the US government has
standardized on USNG as the preferred coordinate system
with WGS84 as the datum for SAR mapping.” The
additional perk of switching to USNG is that it is a shorter
notational method and will therefore reduce screen real
estate coverage.
The experts were extremely happy with the functionality
and readability of the icons. However, as the map is the
primary functionality of the handler’s device, E3
recommended that we reduce the size of the icons or
decrease the opacity when they’re not in use.
All three experts expressed enthusiasm for the ability to
attach notes to specific incidents and GPS coordinates.
However, they requested further functionality. In particular,
E3 noted that most handlers work with gloves on and that
they may not be able to have easy access to typing. The
ability to use voice-to-text or attach verbal messages to a
location would help facilitate the process.
After building feedback into our system, our next step is to
create a user study to determine how well our prototype will
work during extensive use in the field. There might also be
an opportunity to combine some of Bozkurt’s wearable dog
and environmental monitoring [2] and remote human to dog
communication. By using research from Britt [3], combined
with newer data on a dogs ability to feel and understand
haptic input [4] we could move towards creating a system
where the handler could guide the dog remotely.
SAR Task Scenario With Technology
We present a projected scenario of a SAR team using the
SAR system described in this paper.
We would like to express our gratitude to the National
Science Foundation for supporting this work under award
IIS-1525937. We would also like to thank Barbara Currier
and Gabby Gammans for their help with dog training.
[1] American Rescue Dog 2002. Search and rescue
dogs: Training the k-9 hero. Howell Book House.
[2] Bozkurt, A., Roberts, D.L., Sherman, B.L.,
Brugarolas, R. and Mealin, S. 2014. Toward Cyber-
Enhanced Working Dogs for Search and Rescue.
IEEE Intelegent Systems. (2014).
[3] Britt, W. 2009. A software and hardware system for
the autonomous control and navigations of a
trained canine. Auburn University.
[4] Byrne, C., Kerwin, R., Zuerndorfer, J., Gilliland, S.,
Guo, Z., Jackson, M. and Starner, T.E. 2014. Two-
way communication between working dogs and
their handlers. IEEE Pervasive Computing. 13, 2
(2014), 8083.
[5] Chiu, W., Arnold, J., Shih, Y.-T., Hsiung, K.-H.,
Chi, H.-Y., Chiu, C.-H., Tsai, W.-C. and Huang,
W.C. 2002. A survey of international urban search-
and-rescue teams following the Ji Ji earthquake.
Disasters. 26, 1 (2002), 8594.
[6] eliteK9:
bundle/productinfo/EG0320/. Accessed: 2016-11-
[7] Ferworn, a., Sadeghian, a., Barnum, K., Rahnama,
H., Pham, H., Erickson, C., Ostrom, D. and
Dell’Agnese, L. 2006. Urban search and rescue
with canine augmentation technology. 2006
IEEE/SMC International Conference on System of
Systems Engineering. April (2006), 334338.
[8] Ferworn, A., Ostrom, D., Sadeghian, A., Rahnama,
H., Barnum, K. and Woungang, I. 2007. Rubble
search with canine augmentation technology. 2007
IEEE International Conference on System of
Systems Engineering, SOSE. (2007).
[9] Gerritsen, R. and Haak, R. 2014. K9 Search and
Rescue: A Manual for Training the Natural Way,
2e. Brush Education.
[10] ICS Form:
[11] Jackson, M., Zeagler, C. and Valentin, G. 2013.
FIDO-facilitating interactions for dogs with
occupations: wearable dog-activated interfaces.
Proceedings International Symposium on Wearable
Computer. (2013), 8188.
[12] Jackson, M.M., Valentin, G., Freil, L., Burkeen, L.,
Zeagler, C., Gilliland, S., Currier, B. and Starner, T.
2014. FIDOFacilitating interactions for dogs with
occupations: wearable communication interfaces
for working dogs. Personal and Ubiquitous
Computing. (Oct. 2014).
[13] Julius K9 Powerharnes:
harness-category. Accessed: 2016-12-04.
[14] Koester, R.J. 2008. Lost person behavior: a search
and rescue guide on where to look for land, air, and
water. dbS Productions LLC.
[15] Mancini, C., Keynes, M., Juhlin, O., Cheock, A.D.,
Linden, J. Van Der and Lawson, S. 2014. Animal-
Computer Interaction ( ACI ): Pushing Boundaries
Beyond “ Human .” (2014), 833836.
[16] Mancini, C., Keynes, M., Lawson, S., Linden, J.
Van Der, Noz, F., Francisco, S., Wingrave, C. and
Juhlin, O. 2012. Animal-Computer Interaction SIG.
(2012), 12331236.
[17] Mancini, C., Linden, J. Van Der, Bryan, J. and
Stuart, A. 2012. Exploring Interspecies
Sensemaking: Dog Tracking Semiotics and
Multispecies Ethnography. (2012), 143152.
[18] Nielsen, J. 1994. Heuristic Evaluation. Usability
Inspection Methods. 2562.
[19] Robinson, C., Mancini, C., van der Linden, J.,
Guest, C., Swanson, L., Marsden, H., Valencia, J.
and Aengenheister, B. 2015. Designing an
emergency communication system for human and
assistance dog partnerships. Proceedings of the
2015 ACM International Joint Conference on
Pervasive and Ubiquitous Computing - UbiComp
’15. (2015), 337347.
[20] Task Map Flow Chart:
Accessed: 2016-11-04.
[21] Terrain Navigator Pro:
[22] Tran, J., Street, V., Gerdzhev, M. and Ferworn, A.
2010. Continuing Progress in Augmenting Urban
Search and Rescue Dogs. Technology. (2010), 784
[23] Valentin, G., Alcaidinho, J., Freil, L., Zeagler, C.,
Jackson, M. and Starner, T. 2014. Canine
reachability of snout-based wearable inputs.
Proceedings of the 2014 ACM International
Symposium on Wearable Computers - ISWC ’14.
(2014), 141142.
[24] viewranger:
more-than-50-search-rescue-teams. Accessed:
[25] Weilenmann, A. 2011. Understanding People and
Animals: The Use of a Positioning System in
Ordinary Human-Canine Interaction. (2011), 2631
[26] Zeagler, C., Gilliland, S., Freil, L., Starner, T. and
Jackson, M.M. 2014. Going to the Dogs: Towards
an Interactive Touchscreen Interface for Working
Dogs. (2014), 497507.
... AAR attracted significant attention as it can provide significant insights about the behavior, health condition, and location of the observing animal [1]. In addition, if a proper network implementation is considered (e.g., with the proper devices, software, and communication protocol) the monitoring of the animal can be performed in real-time to allow exploitation of AAR for various purposes, e.g., study of the interaction between different animals, search and rescue missions [2], protection of animals from poaching and theft, etc. [3]. To perform this, the use of inertial sensors is mandated, such as accelerometers, gyroscopes, and magnetometers as well as a Machine Learning (ML) method, which after the proper training can accurately classify the animal activity [4]. ...
... The authors in [2] developed a two-part system consisting of a wearable computer interface for working SaR dogs communicating with their handler via a mobile application. The wearable comprised a bite sensor and a GPS to display the K9s location in the mobile application. ...
... Similar to the first layer, this is followed by a ReLU activation function, a (1,2) strided max-pooling operation and a dropout probability equal to 0.5. • layer 3: thirty-two convolutional filters with a size of (2, 11), i.e., W 3 has shape (2,11,24,32). ...
Full-text available
Search and Rescue (SaR) dogs are important assets in the hands of first responders, as they have the ability to locate the victim even in cases where the vision and or the sound is limited, due to their inherent talents in olfactory and auditory senses. In this work, we propose a deep-learning-assisted implementation incorporating a wearable device, a base station, a mobile application, and a cloud-based infrastructure that can first monitor in real-time the activity, the audio signals, and the location of a SaR dog, and second, recognize and alert the rescuing team whenever the SaR dog spots a victim. For this purpose, we employed deep Convolutional Neural Networks (CNN) both for the activity recognition and the sound classification, which are trained using data from inertial sensors, such as 3-axial accelerometer and gyroscope and from the wearable’s microphone, respectively. The developed deep learning models were deployed on the wearable device, while the overall proposed implementation was validated in two discrete search and rescue scenarios, managing to successfully spot the victim (i.e., obtained F1-score more than 99%) and inform the rescue team in real-time for both scenarios.
... It seems expectable that the longer a SAR teams spends in the field, the more their locomotor activity and physical performance level are likely to change. In general, SAR dogs must be able to work between 4 and 8 h without distraction (Zeagler et al., 2016); however, in real situation various factors can negatively affect dog's psychophysical condition (Rovira et al., 2008a(Rovira et al., , 2008b, and, therefore, the probability of finding a missing person can decrease as search times increase (see Greatbatch et al., 2015). When observing dogs during SAR work, it is clear that they tend to move more cautiously in difficult terrain (A. ...
Full-text available
Ground SAR work is a demanding activity that requires trained dogs to work reliably for hours in various terrain and environmental conditions. Factors that would affect a dog’s physical performance in conditions imitating a real SAR mission remain poorly studied. We tested a total of 50 shepherd dogs between 2 and 10 years of age. The group comprised dogs that were both certified and uncertified in SAR work. Testing took place during 5 simulated all-day search sessions. We wanted to know how terrain slope, vegetation cover, ambient temperature, and the number of search rounds would affect locomotor activity of both certified and uncertified dogs. The locomotor activity of dogs was described by vertical and horizontal speed, duration of the search as well as the ratio of time spent in vertical (i.e. uphill) locomotion to total time of locomotion (this ratio is herein referred to as “effectiveness”), and duration of the search. We found out that SAR certification was associated with an increased effectiveness; certified dogs spent more time in energy-conserving horizontal locomotion along contour lines. Terrain and environmental factors influenced both certified and uncertified dogs; a slope of 11° or greater increased vertical speed and duration of searching but decreased effectiveness. Thick vegetation slowed horizontal and vertical locomotion, impaired effectiveness and prolonged search times. Relatively high temperatures (>20 °C) also contributed to longer search times. In order to make quick decisions and attain success during SAR operations, authorities responsible for SAR missions must have a sound knowledge of the locomotor characteristics and skill level (certification level) of SAR dogs in relation to terrain and environmental conditions.
... Frawley and Dyson (2014) used animal personas to represent the perspectives of chickens in participatory design processes with humans. Several projects (Avila 2017;Bos et al. 2009;Isokawa et al. 2016;Zeagler et al. 2016) have studied nonhuman stakeholders in their habitat and through scientific information, but have chosen not to involve them as direct participants. Meanwhile, others (Jørgensen and Wirman 2016;Mankoff et al. 2005;Westerlaken and Gualeni 2016) have involved animals as the direct participants of participatory processes, predominantly through play. ...
... A two-part system consisting of a wearable computer interface for working search and rescue dogs that communicates with their handler via a mobile application, as developed in [12], is an example of such effort. The viability of the tool was evaluated with feedback from three experts. ...
Full-text available
Search and rescue operations can range from small, confined spaces, such as collapsed buildings, to large area searches during missing person operations. K9 units are tasked with intervening in such emergencies and assist in the optimal way to ensure a successful outcome for the mission. They are required to operate in unknown situations were the lives of the K9 handler and the canine companion are threatened as they operate with limited situational awareness. Within the context of the INGENIOUS project, we developed a K9 vest for the canine companion of the unit, aiming to increase the unit’s safety while operating in the field, assist the K9 handler in better monitoring the location and the environment of the K9 and increase the information provided to the Command and Control Center during the operation.
... Dogs are commonly recruited as participants in Animal-Computer Interaction (ACI) studies in a variety of contexts. For working dogs, a growing number devices have been developed including tech embedded wearables [19,20,8,2,44,50], tactile interaction systems [41,27,9], and touchscreen interfaces [51]. Studies focusing on companion dogs have explored wearable tracking [28,42,31,24,39] and technology supported interactions [40,4]. ...
Conference Paper
Full-text available
Outdoor enrichment has a variety of potential uses for increasing physical activity and strengthening companion animal bonds. In this pilot protocol, we aim to explore a species-centric evaluation of drone flying patterns and distances for enrichment purposes. This includes presenting participants with introductory and training phases to encourage positive familiarization before evaluating flight patterns, with the ability to walk away at any time. Sessions will be repeated within participants to explore any novelty or familiarization effects, as well as collect guardian perceptions of impact. From this pilot, we aim to explore our evaluation methods, characterize pet guardian perspectives, and narrow preferred movement patterns and distances for a future deploy.
A compact Planar Inverted-F Antenna design for off-body communications, in the ISM band (2.4–2.5 GHz), is presented. Slots are properly introduced in the radiating element of the PIFA in order to reduce the resonant frequency, thus achieving compactness. The proposed design is based on a textile substrate, making it suitable for integration into garments and clothing items. A lateral coaxial feed mechanism is also employed to reduce the antenna profile. Simulation analysis in terms of antenna return loss and gain is presented as validation.
Full-text available
Computer-mediated interaction for working dogs is an important new domain for interaction research. In domestic settings, touchscreens could provide a way for dogs to communicate critical information to humans. In this paper we explore how a dog might interact with a touchscreen interface. We observe dogs' touchscreen interactions and record difficulties against what is expected of humans' touchscreen interactions. We also solve hardware issues through screen adaptations and projection styles to make a touchscreen usable for a canine's nose touch interactions. We also compare our canine touch data to humans' touch data on the same system. Our goal is to understand the affordances needed to make touchscreen interfaces usable for canines and help the future design of touchscreen interfaces for assistive dogs in the home.
Full-text available
Working dogs have improved the lives of thousands of people throughout history. However, communication between human and canine partners is currently limited. The main goal of the FIDO project is to research fundamental aspects of wearable technologies to support communication between working dogs and their handlers. In this study, the FIDO team investigated on-body interfaces for dogs in the form of wearable technology integrated into assistance dog vests. We created five different sensors that dogs could activate based on natural dog behaviors such as biting, tugging, and nose touches. We then tested the sensors on-body with eight dogs previously trained for a variety of occupations and compared their effectiveness in several dimensions. We were able to demonstrate that it is possible to create wearable sensors that dogs can reliably activate on command, and to determine cognitive and physical factors that affect dogs’ success with body–worn interaction technology.
Conference Paper
Full-text available
We designed an experiment with the goal of assessing wear-able reachability for canines. We investigated the effect of placement on the ability of dogs to reach on-body interfaces with their snouts. In our pilot study, seven placements along the front legs, rib cage, hip and chest are tested with six dogs. The results showed that the front leg placements are reachable with the least amount of training and are also the most invariant to small changes in location. With training, the lower half of the rib cage area had the fastest access times across subjects. We hope that these results may be useful in mapping the constraint space of placements for snout interactions.
Full-text available
Working dogs have improved the lives of thousands of people. However, communication between human and canine partners is currently limited. The main goal of the FIDO project is to research fundamental aspects of wearable technologies to support communication between working dogs and their handlers. In this pilot study, the FIDO team investigated on-body interfaces for assistance dogs in the form of wearable technology integrated into assistance dog vests. We created four different sensors that dogs could activate (based on biting, tugging, and nose gestures) and tested them on-body with three assistance-trained dogs. We were able to demonstrate that it is possible to create wearable sensors that dogs can reliably activate on command.
Full-text available
The Facilitating Interactions for Dogs with Occupations (FIDO) project in the Animal Interaction Lab at Georgia Tech aims to facilitate communication between working dogs and their handlers. Here, the authors discuss their research on testing a working dog's ability to perform distinct tasks in response to vibrations at different points on their body.
Conference Paper
In this research we developed an alarm system that enables assistance dogs to call for help on behalf of their vulnerable owners in an emergency, involving the end users (both assistance dogs and their owners) directly in the entire design process. Here we present a high-fidelity prototype of a user-friendly canine alarm system. In developing the system, we sought to understand the level of support required for a canine user to successfully interact with an interface, finding that the type of emergency a dog is faced with may vary widely and that consequently dogs may have to act on behalf of their assisted owners with varying degrees of autonomy. We also explored the process of conducting usability testing with both canine and human participants, seeking to identify where requirements of one species may overlap with, or diverge from, the other.
Through short presentations, collaborative design exercises and plenary discussions, this one-day workshop aims to explore questions and possibilities for the development of ACI as a discipline.
The authors introduce the fundamental building blocks for a cyber-enabled, computer-mediated communication platform to connect human and canine intelligence to achieve a new generation of Cyber-Enhanced Working Dog (CEWD). The use of monitoring technologies provides handlers with real-time information about the behavior and emotional state of their CEWDs and the environments they're working in for a more intelligent canine-human collaboration. From handler to dog, haptic feedback and auditory cues are integrated to provide remote command and feedback delivery. From dog to handler, multiple inertial measurement units strategically located on a harness are used to accurately detect posture and behavior, and concurrent noninvasive photoplethysmogram and electrocardiogram for physiological monitoring. The authors also discuss how CEWDs would be incorporated with a variety of other robotic and autonomous technologies to create next-generation intelligent emergency response systems. Using cyber-physical systems to supplement and augment the two-way information exchange between human handlers and dogs would amplify the remarkable sensory capacities of search and rescue dogs and help them save more lives.
User-computer interaction research is demonstrating growing interest in the relation between animals and technology (e.g., computer-mediated interspecies interactions and animal-computer interfaces). However, as a research area, this topic is still underexplored and fragmented, and researchers lack opportunities to exchange ideas, identify resources, form collaborations and co-operatively develop a coherent research agenda. The Animal-Computer Interaction (ACI) SIG meeting aims to provide such an opportunity, promoting the development of ACI as a distinct area of research which is relevant to both animals and humans.