Presentation by Stephanie Reese Authors Peng Zhuang, Qingguo Wang, Yi Shang, Honchi Shi, and Bei Hua

foamyflumpΚινητά – Ασύρματες Τεχνολογίες

21 Νοε 2013 (πριν από 3 χρόνια και 7 μήνες)

45 εμφανίσεις

Presentation by Stephanie Reese

Authors
Peng

Zhuang
,
Qingguo

Wang, Yi Shang,
Honchi

Shi, and
Bei

Hua

Focus


“[T]he issues involved in applying
wireless sensor networks to search and
rescue of lost hikers in trails and focus
on the optimal placement of sensors and
access points such that the cost of
search and rescue is minimized.”


How it Relates


Similarities:


The search and rescue algorithms


Most scenarios are assumed to be “non
-
moving” accidents


Differences:


Our sensors will be scattered randomly


More broad scope

An Overview of
CenWits


C
onnection
-
less S
en
sor
-
Based Tracking
System Using
Wit
nesse
s


Hikers wear sensors that have
communication and GPS capabilities


Access Points (AP) are strategically
placed around the trail


When any of these sensors come into
range of another, their information is
recorded as “witness”


Provides constant, dynamic information about
the hiker and their movement

Finding a Probable Location


“The lost case is assumed to be [a] non
-
moving accident, such as being injured,
sick, or stuck along the trail.”


The range of the hiker is established by
the witness information held by APs


A probable path is determined



This section of the implementation is mostly
irrelevant due to the fact that there are no
trails in our scenarios and therefore no paths

Search and Rescue


There are four types of search and
rescue (
SaR
) to consider:


Single Ground
SaR

Agent (S
-
GSA)


Multiple Ground
SaR

Agents (M
-
GSA)


Single Air
SaR

Agent (S
-
ASA)


Multiple Air
SaR

Agent (M
-
ASA)

Single Ground
SaR

Agent (S
-
GSA)


Minimizing the worst case scenario:




Where:


c
M

is maximum cost


G
i

is a trail segment


c`(P) is the cost to travel along
G
i

on the
shortest path P


n(e) is the number of times an edge is visited


c(e) is the cost to search each edge

Single Ground
SaR

Agent (S
-
GSA),
con’t


Minimizing cost for the expected scenario:





Where:


c
E

is the expected cost


t is the specific numbered tour segment (path whose edges have
not been visited before)


j is the specific numbered redundant segment (path whose
edges have been visited at least once)


n is the number of total paths


l
t

is a list of all tour segments


r
j

is a list of all redundant segments


p(
l
t
) is the probability of a hiker getting lost in segment (
l
t
)


w(
l
t
) is the weight of tour segment
l
t


w(
r
j
) is the weight of redundant segment
r
j


w
i

is the total weight of the number of edges

Multiple Ground
SaR

Agents (M
-
GSA)


Minimizing the search effort of each
agent:





When:


k is the number of agents


E
i
T
(x)

is the set of edges travelled by agent x
in
G
i

Multiple Ground
SaR

Agents (M
-
GSA),
con’t


Minimizing cost for the expected
scenario:




Where:


l
t
x

are the tour segments by agent x


r
j
x

are the redundant segments by agent x

Difference Between Air and Ground Rescue


Ground:


Strictly stick to the paths as defined


Air:


Can cross from one trail to another


Calls for an insertion of “dummy” edges in
order to follow previous standard of defining
paths


**crossing can only happen at two vertices**

Single Air
SaR

Agent (S
-
ASA)


Minimizing search cost:




When:


E
i
T

is the set of all edges (including dummy
edges) traveled

Single Air
SaR

Agent (S
-
ASA),
con’t


Minimizing the expected search cost:






Similar to expected search cost of S
-
GSA

Multiple Air
SaR

Agents (M
-
ASA)


Use the same
Gi

as the S
-
ASA equation


Maximal and expected cost are the same as
M
-
GSA