IEEE 2016 - 2017 EMBEDDED ACCURATE WIRELESS SENSOR LOCALIZATION TECHNIQUE BASED ON HYBRID Hybrid PSO-ANN ALGORITHM FOR INDOOR AND OUTDOOR TRACK CYCLING

1
ACCURATE WIRELESS SENSOR LOCALIZATION TECHNIQUE
BASED ON HYBRID PSO-ANN ALGORITHM FOR INDOOR AND
OUTDOOR TRACK CYCLING
2

• Abstract
• LOCATION knowledge in some potential
  applications of wireless sensor networks (WSNs)
  is crucial, e.g., target tracking, person
  tracking, monitoring, healthcare, agriculture,
  disasters, and environment management.
  Localization accuracy is one of the main
  challenges in WSN localization. Several
  approaches are used in WSN localization including
  (i) range-based and (ii) range-free. This study
  aims to determine the distance between the mobile
  sensor node (i.e., bicycle) and the anchor node
  (i.e., coach) in outdoor and indoor environments.
  Two approaches were considered to estimate such a
  distance.

3

• The first approach was based on the traditional
  channel propagation model that used the
  log-normal shadowing model (LNSM), while the
  second approach was based on a proposed hybrid
  particle swarm optimization artificial neural
  network (PSOANN) algorithm to improve the
  distance estimation accuracy of the mobile node.
  The first method estimated the distance according
  to the LNSM and the measured received signal
  strength indicator (RSSI) of the anchor node,
  which in turn used the ZigBee wireless protocol.
  The LNSM parameters were measured based on the
  RSSI measurements in both outdoor and indoor
  environments.

4

• Existing system
• Hop-counts are an easy to carry out
  and can be implemented for a large network, but
  the localization error is also consequently
  increased. The pattern matching method, also
  called the Fingerprint algorithm, involves two
  phases. The first phase is called radio map,
  whereby the received signals at the predefined
  location are recorded in an offline database. The
  second stage uses the pattern matching algorithm
  whereby the observed signal online matches with
  stored data to determine the position of the
  unknown node.
• Disadvantage
• Because the range-free method is aware of the
  position not the distance of the sensor node, it
  is out of the scope of our study.

5
BLOCK DIAGRAM
POWER SUPPLY
MICRO CONTROLLER
ZIGBEE
LCD
POWER SUPPLY
KEY
ZIGBEE
MICRO CONTROLLER
6

• Proposed system
• Our system is based on the wireless
  personal area network (WPAN). In the first
  approach, the system consists of one anchor node
  and one router node and one sensor node. On the
  other hand, the ANN algorithm requires more data
  samples in the training and testing process to
  give a minimum error. The anchor nodes are
  distributed in a fixed position in the test area.
  The mobile nodes comprise one router node and one
  sensor node that are mounted on and move along
  with the bicycle. The sensor nodes measure the
  cycling parameters. All nodes are connected
  wirelessly in a ZigBee network. The sensor node
  communicate with the coach (one of the anchor
  nodes) via a router node to transfer their
  collected bicycle parameter data. The router node
  is used because the sensor nodes are unable to
  forward the bicycle data directly to the coach.

7

• The anchor nodes and router node
  consist of a ZigBee transceiver and an Arduino
  Atmega 328p microcontroller, while the sensor
  node consists of a ZigBee transceiver, an Arduino
  Atmega 328p microcontroller, and sensor element.
  The anchor nodes are mounted on a surface 1.5 m
  above the ground, while the router node is fixed
  under the bicycle seat, which is 0.85 m above the
  ground. The router node, which moves along with
  the bicycle, is called the mobile node
  hereafter. The mobile node is battery powered,
  while the coach node is powered by the laptop of
  the coach via a USB cable. The X-CTU software is
  used to configure the XBees of the mobile and
  anchor node. In addition, it is used to measure
  the RSSI values of the anchor node at a mobile
  node as well as the received successful data
  packets.

8

• Advantages
• The distance error of the mobile node was
  subsequently improved.
• The LNSM parameters, such as path loss exponent
  and standard deviation, were also estimated.

Conclusion Two WSN distance
estimation approaches were presented in this
paper for outdoor and indoor environments. These
methods aimed to determine the distance between
the bicycle location on the track and the coach.
The first method was based on traditional LNSM,
whereas the second approach was an adopted LM ANN
algorithm.
9

• The ANN algorithm was improved by combining the
  PSO and ANN algorithms to select the optimum
  number of neurons in the two hidden layers and to
  achieve the optimum learning rate of ANN. The
  distance error of the mobile node was
  subsequently improved. The propagation channel
  was modeled based on LNSM. The LNSM parameters,
  such as path loss exponent and standard
  deviation, were also estimated. The LNSM-based
  traditional error estimation method was also
  compared with the hybrid PSO ANN algorithm,
  which in turn was compared with the algorithms
  that were employed in previous studies. The
  comparison results disclosed that the MAE and
  RMSE of the hybrid PSOANN algorithm were
  significantly better than those of the LNSM-based
  traditional method, with the latter having a
  significant distance error. The hybrid PSOANN
  algorithm also outperformed similar systems in
  terms of average distance or localization error.
  Therefore, ANN can be conveniently used in both
  outdoor and indoor environments and can be
  applied to any mobile or static node or to any
  WSN.