Target tracking and target localization

Target tracking and target localization

ABSTRACT

Target tracking and target localization are important applications in wireless sensor networks. The aim of this project is to obtain the accurate location information of the target. The sensor network runs a target localization application, where the objective of the network is to provide accurate location information of the target. The coordinates of sensors are known, and the location of the target is estimated based on the measurements and coordinates of nearby sensors. The coverage problem for localization of target is done as a conventional disk coverage problem, where the sensing area is a disk centered at the sensor. Sector coverage model is proposed using a distributed sector coverage algorithm. This requires less number of sensors to locate the target than disk model. The energy required by the active sensors to detect and locate the target is also more in disk model. Simulation results show that sector model locates the target with less consumption of energy when compared to disk model.

I. INTRODUCTION

Wireless Sensor Network

A Wireless Sensor Network (WSN) consists of spatially distributed autonomous sensors to cooperatively monitor physical or environmental conditions such as temperature, vibration, sound, pressure, motion or pollutants. They involve deploying a large number of small nodes. The nodes sense environmental changes and report them to other nodes over flexible network architecture.

Area monitoring is a common application of WSNs. In area monitoring, the WSN is deployed over a region where some phenomenon is to be monitored.

For example, a large quantity of sensor nodes could be deployed over a battlefield to detect enemy intrusion instead of using landmines. When the sensors detect the event being monitored (heat, pressure, sound, light, electro-magnetic field, vibration, etc), the event needs to be reported to one of the base stations, which can take appropriate action (e.g., send a message on the internet or to a satellite).

Depending on the exact application, different objective functions will require different data-propagation strategies, depending on things such as need for real-time response, redundancy of the data (which can be tackled via data aggregation and information fusion techniques), need for security, etc.

A wireless sensor network consists of hundreds or thousands of nodes, which could either, have a fixed location or randomly deployed to monitor the environment. Because of their small size, they have a number of limitations. Each node in a sensor network is typically equipped with a radio transceiver or other wireless communications device, a small microcontroller, and an energy source, usually a battery. The cost of sensor nodes is similarly variable, ranging from hundreds of dollars to a few pennies, depending on the size of the sensor network and the complexity required of individual sensor nodes.

II. RELATED WORK

The coverage concept is subject to a wide range of interpretations due to a variety of sensors and their applications. In [1] the minimal exposure path is determined which provides valuable information about the worst-case exposure- based coverage in sensor networks. In [2] a novel localization algorithm called Area-based Point-In-Triangulation Test (APIT) is presented that is range-free. The APIT scheme performs best when an irregular radio pattern and random node placement are considered, and low communication overhead is desired.

In [3] the coverage problem is formulated as a decision problem, whose goal is to determine whether every point in the service area of the sensor network is covered by at least k sensors, where k is a predefined value. The sensing ranges of sensors can be unit disks or non-unit disks. In [4] maintaining sensing coverage and connectivity is done by keeping a minimal number of sensor nodes in the active mode in wireless sensor networks. A decentralized and localized density control algorithm, Optimal Geographical Density Control (OGDC) is devised for density control in large scale sensor networks.

In [5] the sequential Monte Carlo Localization method is used and argued that it can exploit mobility to improve the accuracy and precision of localization.

III. PROBLEM MOTIVATION AND FORMULATION

A. PROBLEM MOTIVATION

The following two characteristics of sensor networks lend importance to the energy efficient target localization problem.

  • Spatial Queries
  • Limited Battery Power

B. PROBLEM FORMULATION

Given a query in a sensor network, small number of sensors must be selected that are sufficient to localize the target in the query area.

One fundamental issue in application of wireless sensor networks is to provide proper coverage of their deployment regions, In Wireless Sensor Networks, the coverage problem reflects how well an area is monitored or tracked by sensors. Since sensors may be spread in an arbitrary manner, one of the fundamental issues in a wireless sensor network is the coverage problem. Sensors are usually powered by batteries; so the on-duty time of sensors should be properly scheduled to conserve energy. If some nodes share the common sensing region and task, then we can turn off some of them to conserve energy and thus extend the lifetime of the network. This is feasible if turning off such a node still provides the same “coverage”

The goal is to determine the location of the target in a field with sound sensors scattered in that field. The WSN's target detection performance is directly related to the placement of the sensors within the field of interest. The main objective is to place the minimal number sensors such that coverage points in the field can be covered by at least k sensors. The coverage concept is a measure of the Quality of Service of the sensing function and is subject to a wide range of interpretations due to a large variety of sensors and applications. The goal is to have each location in the physical space of interest within the sensing range of at least one sensor. Sensor nodes will be randomly deployed and sensor nodes do not move once they are deployed. Sensors have omni directional antenna and can monitor a disk centered at the sensor's location, whose radius equals the sensing range to sense the target around its sensing range.

IV. COVERAGE FOR TARGET LOCALIZATION

The goal of target localization is to provide the accurate location information of the target (i.e. target localization). The coordinates of sensors are assumed to be known, and the location of the target is estimated based on the measurements and coordinates of nearby sensors.

The network formed by sensors can measure their distance to the target, e.g., Time Difference Of Arrival (TDOA) and Received Signal Strength Indicator (RSSI) sensors. The sensors can provide distance measurements only when the target is within its detection range of its radii.

TARGET LOCALIZATION USING DISK COVERAGE

The algorithm used to perform the localization if the target using disk coverage is Centralized k-coverage algorithm. This algorithm calculates a special drowsiness factor for each sensor and sends to sleep a subset of sensors, depending on their drowsiness factor. The algorithm is as follows:

  1. Run the network for some time period.
  2. Initially all sensors will be in the awake mode.
  3. Compute drowsiness factor for each node.
  4. Select the node with the largest positive drowsiness factor. Send that node to sleep.
  5. Repeat steps 3-4 while possible (i.e. there is at least one node with positive drowsiness factor).

The drowsiness factor of a node s with current energy Esis defined as, where and α is a positive constant (e.g α =2) and Φr is the coverage ratio of region ‘r' and r is defined as,

where cris the degree of node rin G. The coverage ratio Φris positive if the region is over-covered, i.e. more than k sensors cover region r. Φris negative if region ris not over covered: in this case the operation of all sensors covering r is essential.

The drowsiness factor DStakes into account the energy of sensor s: the smaller the energy of a sensor the larger its drowsiness. Negative drowsiness indicates that the sensor is not allowed to sleep.

When disk coverage algorithm is performed in the area selected, the sensor density (i.e. sensor coverage) for the sensors available in the area selected is displayed. Then the drowsiness factor DS is calculated for these sensors alone. The sensors with largest drowsiness factor value (i.e.) the sensors away from the target are send to sleep state. Now the sensors that sense the target alone remain in awake state. In these sensors also the sensors having battery energy less than 50 are made to sleep and the other sensing nodes locate the target using their distance.

TARGET LOCALIZATION USING SECTOR COVERAGE

In high density wireless sensor networks, sensors need to periodically switch to the sleep state to save energy. To ensure that the field is well monitored, the awake sensors should provide coverage over the whole field. In other words, a sensor can only go to sleep when there is no coverage hole in its sensing area. Thus, a distributed coverage algorithm is needed to check whether there are coverage holes in the field when the positions of the waking sensors are known.

For the disk coverage model, a sensor can determine whether its sensing region is k-covered by checking the intersection points of the sensing boundary of its neighbors. This gives an algorithm of complexity O(n3), where n is the number of waking neighbors within range of 2r. However, in the sector coverage model, a sensor needs to check whether there are sector voids around it, which requires more computation.

For a fixed orientation, the sector coverage can be checked using a similar method as in the disk coverage. If all intersection points of sectors of different sensors can be covered, then the field is fully covered. For the sector-shaped sensing region with the same orientation, the border of the sensing region of two sensors can intersect at most on 4 different points.

Calculating the intersection points for two sensors can be done in constant time. When there are n sensors within distance of 2r, there are O(n2) intersection points to be checked. For each intersection point, we need to check whether it is within the sensing region of the other n − 1 sensors. The overall computation complexity for our approximate sector coverage algorithm is O(n3/δ).

The algorithm proceeds as follows:

  1. First, the sensor needs to know the position of all waking sensors within 2r distance from it.
  2. If the disk with radius 0.867r around it can be fully covered by disks of radius rdcentered at other waking sensors, it can go to sleep and the algorithm terminates.
  3. If the disk with radius r around it cannot be fully covered by disks of radius r centered at other waking sensors, it cannot go to sleep and the algorithm terminates.

Compared to disk coverage algorithms, the sector coverage algorithm has computation complexity of O(n3/δ) rather than O(n3). So, sector coverage requires more computation when δis small. The sector coverage algorithm only needs to know the information of neighboring sensors within range of 2r (their location and sleep-wake state). Such information is also required by disk coverage algorithms. So, the information exchange protocols for disk coverage can also be used in the sector coverage algorithm. This algorithm can use fewer sensors to cover the network, while it cannot provide guarantees on the network resolution.

V. SIMULATION

The described coverage algorithm for disk and sector are simulated in Java. Sensor nodes are randomly placed in a rectangular region. All sensor nodes have a circular sensing region of radius r associated with them. The environment is simulated with 25, 50, 75 and 100 sensors. The following parameters are assumed for the simulation:

  • Random positioning of sensor nodes and a single target, which is static.
  • Random assignment of energy levels to each sensor nodes.
  • Uniform sensing Radius & Communication Radius.
  • Communication Radius should be greater than or equal to twice the sensing radius.
  • Every Sensor knows about itself, its neighbor with the global array of node positions.
  • Protocol is operated at the query area selected from the whole set of sensors.
  • Query area is rectangular

PERFORMANCE MEASURES

The performance measure for a sample simulation is given below. It represents the number of active sensors need to localize the target and the distribution energy needed for disk coverage.

Fig 5.3 Performance analysis

The required energy needed to locate the target using disk model or sector model is computed as,

Disk/ Sector Coverage Distribution Energy = Number of active sensors x Coverage angle x Radius of the sensor.

The energy required is more when using disk model since it uses 360-degree coverage for each sensor but reduced in sector model since the disk is partitioned into sectors of small orientations. The sensor name, its drowsiness factor value and the battery energy of the sensors in the selected region is displayed in the command prompt.

VI. CONCLUSION:

Using the disk and sector coverage model the location of the target is obtained. The disk model provides a simple way to estimate the sensor density but with more number of sensors when compared to sector coverage model. Total energy consumption is reduced in sector coverage when compared to disk coverage model. Thus sector coverage is proved to be the better model for target localization since it uses few numbers of sensors and lesser energy.

Please be aware that the free essay that you were just reading was not written by us. This essay, and all of the others available to view on the website, were provided to us by students in exchange for services that we offer. This relationship helps our students to get an even better deal while also contributing to the biggest free essay resource in the UK!