Flbra fuzzy logic based routing algorithm for indoor wireless sensor networks

International Journal of Computer Science & Information Technology (IJCSIT) Vol 6, No 5, October 2014
FLBRA: FUZZY LOGIC BASED ROUTING
ALGORITHM FOR INDOOR WIRELESS SENSOR
NETWORKS
Lucas Leão1, David Bianchini and Omar Branquinho
Pontifical Catholic University of Campinas, Campinas, Brazil
ABSTRACT
Considering the context of building management systems with wireless sensor networks monitoring
environmental features, this paper presents a proposal of a Fuzzy Logic Based Routing Algorithm (FLBRA)
to determine the cost of each link and the identification of the best routes for packet forwarding. We
describe the parameters (Received Signal Strength Indicator - RSSI, Standard Deviation of the RSSI and
Packet Error Rate - PER) for the cost definition of each path, the sequence of identifying best routes and
the results obtained in simulation. As expected in this proposal, the simulation results showed an increase
in the packet delivery rate compared to RSSI-based forward protocol (RBF).
KEYWORDS
Wireless Sensor Networks, WSN, RSSI, RBF, FLBRA, Fuzzy Logic, Sensors, Routing.
1. INTRODUCTION
Wireless sensor networks (WSN) have been a widely used solution for monitoring systems,
including, for example, Building Management Systems (BMS). The sensors are responsible for
monitoring aspects of the environment, capturing data such as temperature, lighting and power
consumption [1]. Through this monitoring, management strategies and control of lighting, and
HVAC (heating, ventilation, and air conditioning) are applied seeking greater efficiency in
resource use.
However, when it comes to an application of WSN in a dynamic environment, such as offices,
subject to constant physical changes, either due to the inclusion of new obstacles like furniture
and walls as well as the movement of people throughout the space, the design of the wireless
network should take into consideration the possibility of constant reconfiguration of the network.
These reconfigurations are necessary since such changes in the environment generate impacts on
the initial performance of the WSN [2].
In this scenario, it is important to identify solutions to meet the demands of reliability and to
overcome the indoor environment constraints. To do so, in order to minimize the impacts of
changes in the environment, this paper presents a strategy for network reconfiguration, by
dynamically changing the routes of communication between the sensors, aiming the continuous
communication between the sensors and the base station. The proposal makes use of fuzzy logic
to assign costs to links considering the relation between RSSI, Standard Deviation of the RSSI,
and Packet Error Rate.
DOI:10.5121/ijcsit.2014.6507 85

The organization of this paper includes the description of related work, describing routing
solutions based on RSSI in Section 2. Section 3 describes a brief background on Fuzzy Logic.
The description of our solution, with the proposed algorithm, is shown in Section 4. Section 5
presents the discussion of the simulation results. Finally the conclusion in Section 6.
2. RELATED WORKS
Routing solutions based on RSSI are characterized by the use of RSSI measured values between
sensor nodes and the base station to define the routes to be used for packet forwarding [3]. Such
protocols may consider the physical distance between the sensor node and the base station,
choosing the closest sensor and electing it as responsible for forwarding packets to the base
station from more distant sensors [4]. Despite of that, it is risky to assume that sensor nodes with
high level of RSSI are near the base station, since shadowing effect may generate unpredicted
outcomes [3].
The proposal presented in [3,4] is a data delivery protocol that combines decisions in MAC and
routing layers in order to determine the next hop in the forwarding process. It is an event-driven
protocol, with distributed processing and state-free design. There is no need for the sender to
know the network topology or its neighbour’s location. Differently from the location-based
protocols, RBF is unaware of sink location, and the forwarding is based on the measured RSSI
received from the sink. It is assumed that the sink is capable of sending beacons, which can be
heard by all sensor nodes in the network. Based on that, sensors are expected to use the RSSI as a
parameter to determine the closest sensor to the node. The strategy goes as a competition based
selection, where the sensor node with the highest RSSI level is selected to forward the data to
next hop. Since packet delivery rate is not a concern in this strategy, it is not assured that the next
hop is the best choice and it will effectively deliver the forwarded packet (void areas).
Reference [5] presents a decentralized packet forwarding solution based on RSSI. It is a location-free
greedy forward algorithm, which makes use of broadcast messages to decide the next hop of
the forwarding task. The authors describe two different approaches, namely RSSR Election and
RSSR Selection. The first approach is based on the election of a leader, which will hold the task
of forwarding the packet. In this solution, the sink node, which is assumed to have a powerful
radio range capable of reaching the entire network, sends a query to a specific node in the form of
a broadcast message. Then, all nodes in the network saves the RSSI from the sink in a local
database. After checking if the query is addressed to itself, the node sends a broadcast message to
all its neighbours. The neighbour having the closest distance from the sink (calculated from the
RSSI) elects itself as the leader and forwards the packet in a form of a broadcast message. The
other nodes participating in the election receive the message and cancel their participation. The
second approach runs differently by not starting an election, but triggering a selection phase.
After receiving the message from the sink, nodes start exchanging advertisements, so nodes can
be aware of the distance of their neighbours from the sink. The information is stored in a local
routing table, which is then used by the inquired node to select the next hop (the neighbour
closest to the sink). Notwithstanding the gains of a distributed solution, such as implicit reaction
to network changes, the proposed solutions do not address issues related to voids in the network,
resulting in packet not being delivered.
In [6] a solution based on the characteristics of signal propagation is presented. The authors'
proposed algorithm is a centralized solution that checks the status of the links between sensor
nodes and builds a routing table, which is stored within each sensor node. Then it starts to
monitor the links, noting possible degradation in the signal. If there is a link in the routing table
whose parameters are below a desired threshold, the algorithm reconstructs the routing table,
identifying the links again with values within the required limit. The proposal uses the packet
86

delivery rate and signal to noise ratio (SNR) of each link as a parameter to evaluate the eligibility
of the route as an entry in the route table. This solution is an improvement of a previous work [7]
presenting a static routing protocol that used the average of measured RSSI and the packet
delivery ratio as parameters to construct the routing table. In both proposals, authors define a
relation between link quality and packet delivery ratio in order to classify the links of each sensor
and create a routing table. Although it is a dynamic solution, capable of adapting itself to changes
in the link quality, it is still subject of voids in the network, since it establishes thresholds for the
measured parameters. It means that occasionally sensor nodes may be out of the defined threshold
and consequently out of the routing table, creating voids in the network.
The method suggested by [8] makes use of Fuzzy Logic to select the sensor nodes that will be
used for the packet forwarding task. It considers two main variables in order to select the next
hop: sensor node angle in relation to its neighbours and number of packets forwarding to the
neighbour. Those variables are analysed through a fuzzy system and the outcome is a number that
represents the chance of a sensor node being selected as next hop. The next step consists of
selecting the neighbour node which have the highest chance to forward the packet. Although it is
a distributed solution, sink location must be known by sensor nodes in order to correctly select the
next hop. Location is obtained by comparing the angle of the node in relation to the sink. It
implies that the base station cannot have its location changed; otherwise the entire network would
be lost. Moreover, there is no controlling method to avoid nodes with high rates of packet loss.
In [9] the authors propose a decentralized algorithm based on TARF (Trust-Aware Routing
Framework) to determine the reliability of a sensor within the network and to verify if the
information from the sensor nodes are eligible for aggregation or not, thus defining the routes for
packet forwarding. The proposal makes use of Fuzzy Logic as a tool to support the decision
process. Variables such as Energy Cost, distance, signal strength and packet delivery ratio, which
after a process of fuzzification, serve as input to the decision rules.
Another proposal is presented in [10] describing a method for route selection using fuzzy logic
and considering the evaluation of three parameters: battery level, trust level and distance from the
base station considering the RSSI level. The value of the residual battery level is calculated taking
into account the energy used for transmission at a given distance. The proposed model takes into
account a decentralized solution to select nodes that should serve as a path in the routing process.
It is still considered the existence of a database, in the sensor node itself, for storing information
about the history of neighbour sensor nodes.
In [11] it is presented an estimator of the link quality between sensor nodes using Fuzzy Logic.
Parameters such as packet delivery rate, link asymmetry in uplink and downlink, the stability of
the link (changes in the measured parameters) and channel quality (signal to noise ratio) are used
as metrics for the link quality evaluation. The authors justify the use of Fuzzy Logic in the link
quality estimation due to the imprecision in the measures of the factors that affects the quality and
stability of the communication between the sensor nodes.
Our approach aims to be as simple and objective as RBF protocol described in [3,4]. However, it
is also our goal to provide more accuracy, using packet error rate as a parameter for routing
decision. Our solution also differs from [5-7] by using Fuzzy Logic to evaluate the link quality
and taking the standard deviation of the measured RSSI to determine the link stability — since
thresholds are not defined, nodes with poor link quality may still be used, but only in extreme
cases in order to avoid the existence of holes in the network. For [8-11], which are solutions using
Fuzzy Logic, we propose a simplification and the use of different parameters, in an approach
specially designed for indoor applications.
87

3. FUZZY LOGIC
Traditional logic, based on assumptions and conclusions, cannot represent naturally imprecise
phenomena [12], because the propositions come as absolutely true or absolutely false, with no
margin for abstract conclusions.
As an alternative to treating problems with inaccurate information or based on incomplete data,
we can mention the techniques of Soft Computing. According to [13], Soft Computing seeks to
address the uncertainties and inaccuracies inherent in the world we live in order to facilitate the
acquisition of more accurate results. Among the techniques of Soft Computing, we can mention
Fuzzy Logic, which aims the abstraction of concrete concepts in order to eliminate uncertainties,
and evaluate the proposals according to rules based on experience.
In [12] it is stated that the Fuzzy Logic, unlike traditional logic, allows different levels of
certainty. This means that it is possible to characterize variables across membership functions that
overlap and are influenced by modifiers. Thus, it is possible to deal with abstract concepts similar
to human thinking, not considering absolute values, but rather the gradual perception of a
particular information.
Therefore, taking into account the uncertainties of the factors influencing the quality of
communication in a dynamic environment, we can apply the Fuzzy Logic to verify a link quality
index based on the average RSSI, Standard Deviation of the RSSI, and the Packet Loss in order to
determine the best routes for packet forwarding.
4. PROPOSAL
The proposed algorithm uses Fuzzy Logic to evaluate quality parameters from each link in the
WSN. It is analysed the average RSSI level, the standard deviation of the measured RSSI, and the
packet error rate in order to classify the link for the routing decision. Once the cost of each link is
determined, it uses Dijkstra's algorithm [14] to perform the search of best routes and thereby
reconfigure the network.
The base station (sink) acts as a coordinating entity, collecting the RSSI measures, standard
deviation, and Packet Error Rate (PER) from each sensor node. Each sensor is responsible for
storing the requested data in its own memory and carry out of the necessary measures. However,
measurements occur during normal network traffic, not hampering the network overall
performance with excessive control packets. The base station is also endowed with a better
computational power and radio coverage.
The operating steps of the FLBRA are described as follows:
Setup Phase - the base station sends a beacon message to all sensors in the network. The sensors
use this message to evaluate their RSSI with the base station. Each sensor node near the base
station, and with better RSSI, answers the message. In order to avoid collisions, sensor nodes wait
for a time interval based on their RSSI with the base station before sending the answer.
Neighbour nodes listen to the answers and keep the RSSI values from other sensors in a local
database. The base station waits for the answer of each node in the network until a timeout is
expired. After that, the base station sends a new broadcast message asking for RSSI information
from the neighbours of each sensor. The sensors answer the base station with the information of
their neighbours. At this moment, the base station acknowledges the existence of other sensors,
for which the replies could not reach the sink in a single hop. By using the received data, the base
88

station runs the fuzzy logic procedure and the Dijkstra algorithm in order to create the routing
table. The base station sends the routing information for each sensor node, which starts
forwarding packets according to the new rules. A new broadcast message asking for RSSI
information from neighbours is sent in order to identify new sensors. This process is repeated
until the base station detects no new sensors in the network.
The Algorithm 1 represents the overall procedures of the setup phase.
Algorithm 1 NETWORK SETUP
1: Function setup_phase()
2: send_broadcast(beacon_message) //sends the beacon message to the sensor nodes
3: timeout = set_timeout(null) //starts a timeout waiting for sensor nodes replies
4: [net_info, count_sensors , neighbour_sensors] = wait_response(timeout) //receives
89
information from sensor nodes
5: send_broadcast(rssi_request, count_sensors) //sends the request for rssi readings
6: timeout = set_timeout(count_sensors) //starts a timeout waiting for sensor nodes
replies based on the number of sensor nodes
8: route_definition(net_info, count_sensors) //defines the routes based on the received
information
9: while neighbour_sensors<>null do
10: send_broadcast(rssi_request, neighbour_sensors) //sends the request for rssi
readings from neighbours
11: timeout = set_timeout(neighbour_sensors) //starts a timeout waiting for sensor
nodes replies based on the number of sensor nodes
13: route_definition(net_info, count_sensors) //defines the routes based on the received
information
14: end
15: return (null)
16:
17: Function wait_response(timeout)
18: while NOT expired(timeout) do //keep listening sensor while timout is not
expired
19: sensor_info = receive_answer() //receive the answers from each sensor node
20: timeout = update_timeout() //update timeout for each new detected sensor
21: [net_info, count_sensors , neighbour_sensors] = new_sensor_check(sensor_info)
//number of sensor nodes that can reach the base
22: end
23: return (net_info, count_sensors , neighbour_sensors)
24:
25: Function route_definition(net_info, count_sensors)
26: path_info = fuzzy_system(net_info) //runs the fuzzy system
27: route_table = dijkstra(path_info) //runs the Dijkstra algorithm
28: For each sensor in count_sensors //changes route to each sensor node in the network
29: change_route(sensor, route_table) //send the message to change routes
30: end
31: return(null)

Operation Phase – once the initial routing table is defined, the network starts the operation phase.
The sensors forward their data to the sink following the rules defined in the routing table. At the
same time, RSSI information from neighbours and PER are collected and stored for a future
check. After a predefined time, the base station sends a new broadcast message asking for new
readings. The sensors forward the stored information from their neighbours and the base station
analyse if a change in the routing table is required.
The Algorithm 2 represents the operation phase of the FLBRA protocol.
Algorithm 2 NETWORK OPERATION
1: Function operation_phase()
2: network_status = OPERATIONAL
3: timeout = THRESHOLD
4: while network_status = OPERATIONAL do
5: [net_info, data] = wait_sensor_data(timeout) //waits for data from sensor nodes
90
until timeout expires
6: if data <> null then //forwards data to application layer if any
7: application_layer(data)
8: end
9: if expired(timeout) then //checks if timeout expired for network status check
10: network_status = network_check()
11: if network_status = FAULTY then //checks if network is working fine
12: setup_phase() //in case network needs reconfiguration
13: end
14: timeout = THRESHOLD //updates the timeout threshold
15: end
16: end
17: return (null)
18:
19: Function network_check()
20: send_broadcast(rssi_request, count_sensors) //sends the request for rssi readings
21: timeout = set_timeout(count_sensors) //starts a timeout waiting for sensor nodes
replies based on the number of sensor nodes
23: if neighbour_sensors > 0 then //check if news sensor nodes were detected/added
24: return(FAULTY)
25: end
26: new_path_info = fuzzy_system(net_info)
27: if new_path_info <> path_info then
28: return(FAULTY)
29: end
30: return(OPERATIONAL)
The process to evaluate the cost of each link makes use of fuzzy logic. Hence, it is necessary to
define the membership functions of each variable to be analysed (RSSI, Standard Deviation and
PER). One must also specify an output variable named "Cost", which is the abstraction of the
intervals that are used to determine the cost of the link, where the pertinence of the resulting
assessment of the rules will be identified. The sets that comprise the variable "Cost" are: High,
Medium and Low as represented in Figure 1. RSSI values were also divided in a membership
function and categorized into: Weak, Average and Strong as described in Figure 2. Likewise, the

standard deviation sets were also defined: Good, Average and Bad also represented in Figure 3;
and PER variable were represented with Low, Medium and High sets as described in Figure 4.
Importantly, the values for the ranges of the variables RSSI and Standard Deviation may vary
according to the minimum acceptable power of the radio being used. For this work, we defined a
sensitivity of -90 dBm for the purpose of simulation, since it is a practical limit.
91
Figure 1 Cost Membership Function
Figure 2. RSSI Membership Function
Figure 3.Standard Deviation Membership Function

92
Figure 4.Packet Error Rate Membership Function
Once defined the membership functions of the variables to be analysed, it is necessary to identify
the relevance of each value obtained from the measurements in relation to the fuzzy sets. The
pertinence defined for each measure of each fuzzy set is used as input for the execution of the
defined rules. Table 1 represents the fuzzy rules defined for this work. The execution of the rules
allows the identification of subsets in the membership function of the variable "Cost"; in other
words, it is possible to verify the areas of influence of each rule for each parameter in the link
cost definition. The definition of fuzzy rules is an important step of the process. It reflects the
empirical analysis of the studied system. It is possible to check in [15] the Mamdani inference
method that was used in this work, which is basically the execution of a IF…THEN process.
Table 1.Fuzzy Rules
Parameter 1 Parameter 2 Parameter 3 Consequence
High PER High PER High PER High Cost
Medium PER Medium PER Medium PER High Cost
Low PER Weak RSSI Bad Standard Deviation High Cost
Low PER Weak RSSI Average Standard Deviation Medium Cost
Low PER Weak RSSI Good Standard Deviation Low Cost
Low PER Average RSSI Bad Standard Deviation High Cost
Low PER Average RSSI Average Standard Deviation Medium Cost
Low PER Average RSSI Good Standard Deviation Low Cost
Low PER Strong RSSI Bad Standard Deviation High Cost
Low PER Strong RSSI Average Standard Deviation Low Cost
Low PER Strong RSSI Good Standard Deviation Low Cost
To determine an absolute value (called crisp) for the effective cost of the link it is necessary to
implement a defuzzification strategy. The centroid method (centre of area) is used, since it
presents lower mean square errors in comparison to other methods, such as mean of maximum
method [16]. Once identified the crisp value of the link cost to each sensor node, Dijkstra's
algorithm is run in order to find the best routes for packet forwarding. At this point, the routing
table of each sensor is defined, and it is identified the sequence to be followed for sending packets
from the base station to a specific sensor node and vice versa.

5. RESULTS
To evaluate the performance of the proposed algorithm in comparison to the RBF protocol, it was
developed a simulator using SciLab [17]. RSSI values were generated taking into account the
distance between the sensor nodes in each scenario. For this process, we adapted the model
described in [18]. The standard deviation was calculated based on the generated RSSI, and the
PER was randomly attributed. The network was designed identically for both FLBRA and RBF
solutions.
Six scenarios were defined only varying the total amount of sensor nodes, and the total size of the
environment. The distance between the sensor nodes was kept as a constant and it was set at 3 m.
Table 2 shows the characteristics of each scenario.
93
Table 2.Test Scenarios
Scenario Quantity of Sensor Nodes Room size
S01 8 36 m²
S02 24 144 m²
S03 48 324 m²
S04 80 576 m²
S05 120 900 m²
S06 160 1296 m²
To establish a comparison factor between the FLBRA and the RBF protocol, we defined the
process presented in (1):

F = Ʃn
i=1(SFLBRA i – SRBF i)/n , [–1 ≤ F ≤ 1] (1)
Where:
F: single comparison parameter
N: number of sensor nodes
SFLBRA: Success rate based on the proposed fuzzy solution
SRBF: Success rate based on the RBF solution
Success rates can be found as follows in (2.a) and (2.b):
SFLBRA = 1– PERFLBRA , [0 ≤ PERFLBRA ≤ 1] (2.a)
SRBF = 1– PERRBF , [0 ≤ PERRBF ≤ 1] (2.b)

Where:
PERFLBRA: packet error rate of a specific link for the solution based on fuzzy logic
PERRBF: packet error rate of a specific link for the solution based on RBF
The parameter F represents a factor of comparison between FLBRA and RBF. F values between
[-1,0[ demonstrate that the RBF solution is more efficient. For F = 0, we have that the solutions
are similar and break even at the same results. As for F values between ]0,1], we have that
FLBRA has a better performance compared to RBF.

For each defined scenario, it was executed 100 iterations of the simulation in order to generate
values of the F parameter and evaluate the consistency of the proposed routing solution. Figure 5
shows a graph of the performance based on the evaluation of the F parameter. The θ1 and θ2
curves are respectively the lower and higher values of the confidence interval based on a
confidence level of 95% and FM is the mean of the of the F parameter after 100 iterations.
94
Figure 5. Performance curve (values displayed only for the FM parameter)
It is possible to see that the proposed algorithm always performs better than the strategy based
solely on the RSSI. This means that the links created by evaluating the RSSI, standard deviation
and packet error rate imply in greater network reliability. It is also observed that as they increased
in the number of sensor nodes in the network, we obtain even better results for FLBRA. However,
when the number of sensor nodes rises above 100 devices, there is a drop in the performance for
both solutions. This is due to the need for more sensor nodes to go on multi hop routes until they
can reach the base station. Since there are more sensor nodes and the total area of the test scenario
was increased, it is natural that the accumulated performance decreases. The increase of hops
works negatively in the packet error probability, as described in (3):
i=1(1 – PER i)] (3)

PEP = [1– Πn
Where:
PEP: final packet error probability
N: number of links
PER: packet error rate of a specific link
FLBRA also performs better in terms of total number of hops for the farthest sensor node. As
evidenced by Figure 6, FLBRA uses less hops to reach the base station. This is also observed in
Figure 7 which represents the average number of hops in each scenario for both FLBRA and RBF
solutions.

95
Figure 6 Number of hops for the farthest sensor node
Figure 7.Average Number of Hops
Thus, one can say that FLBRA that is based on the evaluation of RSSI, standard deviation, and
packet error rate parameters using fuzzy logic is presented as more appropriate solution for
defining routes with better performance, while keeping the simplicity of the implementation.
6. CONCLUSION
This paper proposed the use of fuzzy logic as a tool to determine the cost of each link based on
signal and communication quality attributes, seeking to identify the best routes in wireless sensor
networks, aiming at improving the overall performance of the network in relation to packet error
rate.
It can be observed from the simulation that the proposed algorithm presents an improved success
rate of packet delivery in relation to the RSSI-based solution. It is also visible that as the amount
of sensor nodes in the network raises, the proposal based on fuzzy logic becomes even more
advantageous, with larger differences from the RSSI-based routing.
As future work, we intend to modify the algorithm so that it can dynamically vary the intervals of
the RSSI, Standard Deviation and PER membership functions in order to adapt to different
environments at runtime, while following specified quality criteria. We also intend to develop the
algorithm on a real wireless sensor network, using Radiuino [19] platform, and conduct tests in
buildings with constant physical variations.

96
REFERENCES
[1] D. De Guglielmo and G. Anastasi, "Wireless sensor and actuator networks for energy efficiency in
buildings," in Sustainable Internet and ICT for Sustainability (SustainIT), 2012, 2012, pp. 1-3.
[2] M. S. Aslam, A. Guinard, A. McGibney, S. Rea and D. Pesch, "Wi-design, wi-manage, why bother?"
in Integrated Network Management (IM), 2011 IFIP/IEEE International Symposium on, 2011, pp.
730-744.
[3] A. Awang, X. Lagrange and D. Ros, "RSSI-based forwarding for multihop wireless sensor networks,"
in The Internet of the FutureAnonymous Springer, 2009, pp. 138-147.
[4] A. Awang, X. Lagrange and D. Ros, "A cross-layer medium access control and routing protocol for
wireless sensor networks," Proc.10emes Journées Doctorales En Informatique Et Réseaux (JDIR
2009), 2009.
[5] A. Boukerche, H. A. B. F. Oliveira, E. F. Nakamura and A. A. F. Loureiro, "A novel location-free
greedy forward algorithm for wireless sensor networks," in Communications, 2008. ICC '08. IEEE
International Conference on, 2008, pp. 2096-2101.
[6] A. Kanzaki, T. Hara, S. Nishio and Y. Nose, "A dynamic route construction method based on
measured characteristics of radio propagation in wireless sensor networks," in Advanced Information
Networking and Applications (AINA), 2011 IEEE International Conference on, 2011, pp. 30-37.
[7] Y. Nose, A. Kanzaki, T. Hara, & S. Nishio, "Route Construction Based on Received Signal Strength
in Wireless Sensor Networks," (2010) International Conference on Computers and Their Applications
(CATA 2010), USA, pp. 127-130.S.
[8] J. Dastgheib, M. R. S. , H. Oulia, S. J. Mirabedini. A new method for flat routing in wireless sensor
networks using fuzzy logic. In Computer Science and Network Technology (ICCSNT), 2011
International Conference on, IEEE Vol. 3, pp. 2112-2116.
[9] I. Sakthidevi and E. Srievidhyajanani, "Secured fuzzy based routing framework for dynamic wireless
sensor networks," in Circuits, Power and Computing Technologies (ICCPCT), 2013 International
Conference on, 2013, pp. 1041-1046.
[10] S. S. Babu, A. Raha, M. K. Naskar, O. Alfandi and D. Hogrefe, "Fuzzy logic election of node for
routing in WSNs," in Trust, Security and Privacy in Computing and Communications (TrustCom),
2012 IEEE 11th International Conference on, 2012, pp. 1279-1284.
[11] N. Baccour, A. Koubâa, H. Youssef, M. B. Jamâa, D. Do Rosário, M. Alves and L. B. Becker, "F-lqe:
A fuzzy link quality estimator for wireless sensor networks," in Wireless Sensor
NetworksAnonymous Springer, 2010, pp. 240-255.
[12] L. A. Zadeh, "A summary and update of “Fuzzy logic”," in Granular Computing (GrC), 2010 IEEE
International Conference on, 2010, pp. 42-44.
[13] L. A. Zadeh, "Soft computing and fuzzy logic," Software, IEEE, vol. 11, pp. 48-56, 1994.
[14] A. S. Tanenbaum, "Computer Networks 4th Edition◦." 2003.
[15] E. H. Mamdani, "Application of fuzzy algorithms for control of simple dynamic plant," Electrical
Engineers, Proceedings of the Institution of, vol. 121, pp. 1585-1588, 1974.
[16] C. Lee, "Fuzzy logic in control systems: fuzzy logic controller. II," Systems, Man and Cybernetics,
IEEE Transactions on, vol. 20, pp. 419-435, 1990.
[17] S. Enterprises, "Scilab: Free and Open Source software for numerical computation," Scilab
Enterprises, Orsay, France, 2012.
[18] C. Papamanthou, F. P. Preparata, R. Tamassia. "Algorithms for location estimation based on rssi
sampling." Algorithmic Aspects of Wireless Sensor Networks. Springer Berlin Heidelberg, 2008. 72-
86.
[19] O. Branquinho, "Plataforma Radiuino para Estudos em Redes de Sensores Sem Fio,"
http://www.radiuino.cc, Em, pp. 30-09, 2011.

Flbra fuzzy logic based routing algorithm for indoor wireless sensor networks

More Related Content

What's hot (19)

Viewers also liked (20)

Similar to Flbra fuzzy logic based routing algorithm for indoor wireless sensor networks (20)

Recently uploaded (20)

Flbra fuzzy logic based routing algorithm for indoor wireless sensor networks