Target Tracking With Accurate Target Computer Science Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Target tracking with accurate target matching in wireless sensor network is the most difficult issue, where the tracking should be done with the low power consumption. Since all nodes are battery powered the main issue for sensor network is the power utilization, so energy efficiency can be considered using many forms and shapes. We predict the target with low energy consumption and the performance is increased with high efficiency. In existing system tracking algorithm have the difficulty of tracking with accuracy and range coverage. It also has the problem of finding the target which is in out of range. In this project target tracking algorithm will merge the simple powerful sensor node which is high energized, reliable with the networked collection of other sensor nodes for accurate target tracking. The target will maintain a close proximity and neighbor nodes will broadcast the information for the detection without the target missing. Our approach combines a protocol for the sensor network that conserves energy by dynamically adjusting the time-to live for packets it transmits with a reactive strategy for the tracker based on its information. Implementation is presented along with experimentation. Our experimental results show that our system achieves both good tracking precision and low energy consumption.Thus our system compares energy performance using routing protocols and graph will be generated using the customized simulator.



B.Brahma Reddy and K.Kishan Rao [2012] tells about the Topology control in wireless sensor networks helps to lower node energy consumption by reducing transmission power and by restricting interference collisions and retransmissions. Decrease in node energy consumption implies probability of increasing network lifetime. In this paper, firs popular topology control algorithms are used for analyzing optimizing the power consumption in the wireless sensor network and later proposed a novel technique wherein power consumption is traded with additional relay nodes. Later relay nodes are introduced to make the network connected without increasing the transmit power. The relay node decreases the transmit power required while it may increase end-to-end delay. This project designs and analyzes an algorithm that place an almost minimum number of relay nodes required to make network connected. Greedy version of this algorithm is implemented and demonstrated in simulation that it produces a high quality link. InterAvg, InterMax MinMax, and MinTotal are used as metrics to analyze and compare various algorithms. Matlab and NS-2 are used for simulation purpose.

Aysegul Alaybeyoglu [2011] tells that new target tracking algorithm is developed for wireless sensor networks. The goal of the algorithm is to decrease power consumption of the system by decreasing the ratio of target which is missed. Then the target location is predicted by using Particle Filtering (PF) technique which aims to represent the posterior density function by a set of random samples with associated weights. Nodes are deployed according to the hexagon shaped network topology in which each of the hexagons represents a cluster with a node determined before. In order to decrease the ratio of target missed, nodes that are closer to the target's predicted location are woken up to make them ready for detecting the target. This increases the probability of detecting the target by one of the neighboring hexagons when the target makes sudden turns or unexpected movements. Tracking performance of the proposed algorithm is evaluated by comparing with K Nearest Cluster Tracking (KNCT), Wakening Based Target Tracking Algorithm (WBTA) and Generic Static Tracking Approach (GSTA) in terms of miss ratio and energy consumption metrics.

Orhan Dagdevireny, Kayhan Erciyesz, Aylin Kantarci [2011] tells that target tracking is the most important application in the wireless sensor networks. Clustering is the basic technique for a scarce network. In this clustering is developed before the target enters the system. It provides static and dynamic clustering algorithms with various mobility models. The model used here are random waypoint model, direct and gauss Markhov model. It also provides metrics for both static and dynamic approaches. This shows that dynamic method shows more accuracy whereas static provides low power consumption. It also shows that Gauss Markhov model technique provides effective accuracy tracking.

Chao Gui and Prasant Mohapatra[2010] tells about the monitoring of the target and the movements of an object which specifies the target. It includes two states of operation which are surveillance and tracking where the detection of object and there movements are tracked respectively. The power conservation and the quality are the two disturbing factors where unlimited power supply cannot be provided so that with the given area the monitoring should be done effectively using spatial and temporal also focus on the trade-off analysis between the quality and power conservation of the target tracking system.

Murrieta-Cid explains the method which consists of the motion planning problem for target tracking. The case for predictable targets is presented in which it describes an algorithm that computes numerical and optimal solutions for problem so flow-dimensional configuration spaces. However, the assumption that the motion of the target is known in advance is a very limiting constraint. The algorithm relies on a heavy discretization of the environment in order to apply recursion based on the dynamic programming principle. The game theory is proposed as a framework to formulate the tracking problem. The main contribution of this work is a technique that periodically commands the observer to move into a region that has no localization uncertainty in order to delocalize and better track the target afterwards. A technique is proposed to track a target with-out the need of a prior map. Instead, a range sensor is used to construct a local map of the environment, and a combinatorial algorithm is then used to compute a differential motion for the observer teaches iteration. The advantage of this technique is that no explicit self-localization mechanism is required. Thus, the implementation of the tracking system becomes simpler.

Jiyan Pan, Bo Hu, and Jian Qiu Zhang [2006] explained that object tracking, the most complex background is that local maxima is formed to distract the target avoid such issues or risk Kalman filter is adapted to predict the initial search with the coordinate transform so that both reliability in tracking and simplicity in computation is improved. Noise is reduced using the Kalman filter and estimation is done without any artificial assumption which makes the target simulation and step size without any manual interventions. In this the simulation analysis shows the effectiveness of the filter using the prediction algorithms.

Boyoon Jung1 and Gaurav S. Sukhatme tell this approach is not directly applicable to a real world system. The most critical limitation is that the size of the state space increases exponentially as the number of target increasing. Since the evaluation time problem is exponential in the size of the state space then the problem becomes intractable. Therefore, for scalability, a distributed solution is preferable to the centralized, optimal solution. Another limitation is that the optimal policy needs to be re -computed whenever the system configuration changes (examples include adding or removing robots at runtime, or adding/removing targets at run time) which implies that the policy computation should be done in real-time. Where the problem has been overcome by each robot broadcasts its location and the locations of currently tracked targets. Based on this information and similar information gathered from other robots, each robot independently maintains an estimate of two density distributions the robot density and the target density. The control law for each robot is generated by using these density estimates. Communication among robots is the key enabler for multi-robot coordination, so the effect of communication range was analysed. It observed the performance change as the communication range varies. The simulation results show that the proposed algorithm is efficient and robust.


Wireless Sensor Network is a technique which can facilitate communication without any connected links and real time data processing in a complex environment. A wireless network sensor consists of large number of sensors which are interconnected to each other and all of them will communicate with the base station. Sensor nodes have limited processing feasibility, storage energy and bandwidth when you compare them to traditional desktop computers. A network that is formed when a set of small sensor devices that are deployed in an "ad hoc fashion" with no predefined routes, and cooperate for sensing a physical phenomenon. A Wireless Sensor Network (WSN) consists of base stations and a number of wireless sensors. It is simple, tiny, inexpensive, and battery-powered. Research in Materials Science has resulted in novel sensing materials for many Physical, Chemical and biological sensing tasks.Transceivers are becoming smaller, less expensive, and less power hungry in wireless devices. Power source improvements in battery as well as passive power sources such as vibration or solar energy are expanding application options for the sample autocorrelation matrix, or batch singular value decomposition (SVD) of the data matrix, where both are computationally too expensive for adaptive applications. Modern subspace tracking algorithms are recursive in nature and update the subspaces in a sample by sample fashion. An adaptive invisible multiuser detector can be based on subspace tracking by sequentially estimating the subspace components and forming the closed form detector based on these estimates. Tracking problems for mobile nodes have received substantial attention in recent years. In these problems, a target tracker seeks to maintain close proximity to an unpredictable target. Best target tracking algorithms have many important applications including monitoring and security. Another potential application is that the tracker needs to carry cargo for the target and has to 'follow' the target in real time. We note that this type of target tracking problems is different from the sensor network target tracking problem in which the main goal is to identify the moving trajectory of the target. It focuses on the target tracking problem. Algorithms have been developed to solve this problem with mobile targets under various constraints and sensor models.

However, these existing methods for target tracking are hampered by two primary limitations. Existing tracking methods generally rely on sensors, which by nature only provide information about the target's location when the target is nearby. This limitation is particularly problematic in cases where the tracker starts with little or no knowledge of the target's location or the tracker loses contact with the target during its execution. To recover from these situations using only local information is a challenging problem, requiring extensive search in the worst case. The tracking task can be divided into two parts: sensing the target and following its movements. As such, we decouple these parts and delegate the sensing task to a stationary sensor network. The mobile tracker then follows the target using only the observations made by these sensor nodes. Such an arrangement has several advantages. Firstly, it eliminates the need for complex sensors on the tracker, which relaxes the hardware requirements of the tracker and simplifies the sensing data processing algorithm that is needed to filter out the difference of consecutive readings caused by the tracker's movement. Secondly, it provides a means for delivering nonlocal information to the tracker, which is critical to help the tracker recover when it loses contact with the target .We devise two energy-efficient, low-maintenance, and robust routing protocols that can forward information towards a mobile tracker. Both routing protocols leverage cross layer information to reduce routing overheads. Neither protocol requires periodic beacon exchanges between neighbors, nor are both robust to a wide range of network dynamics, e.g., sensor nodes may fail or even move. Through our extensive simulation study, we demonstrate that for such systems, both tracking performance and energy-efficient routing protocols can be achieved simultaneously. We also demonstrate that the addition of greater computing power and memory to the sensor nodes enables smarter algorithms that improve the performance even more in certain circumstances. Distributed shortest-path routing protocols for wired networks either describe the entire topology of a network. They continually update the state describing the topology at all routers as the topology changes to find correct routes for all destinations. Hence to find the strong routes, they generate routing protocol message traffic proportional to the product of the number of routers in the network and the rate of topological change in the network. Current ad-hoc routing protocols, designed specifically for mobile, wireless networks, exhibit similar scaling properties. It is the relation of these routing protocols on state concerning all links in the network, or all links on a path between a source and destination, that is responsible for their poor scaling. We present Greedy Perimeter Stateless Routing (GPSR) novel routing protocol for wireless datagram networks that uses the positions of routers and a packet's destination to make packet forwarding decisions. GPSR makes greedy forwarding decisions using only information about a router's immediate neighbours in the network topology. When a packet reaches where greedy forwarding is impossible, the algorithm recovers by routing around the perimeter of the region. By keeping state only about the local topology, GPSR scales better inner-router state than shortest-path and ad-hoc routing protocol as the number of network destinations increases. Under mobility's frequent topology changes, GPSR can use local topology information to find correct new routes quickly. We describe the GPSR protocol, and use extensive simulation of mobile wireless networks to compare its performance with that of Dynamic Source Routing. Our simulations demonstrate GPSR's scalability on densely deployed wireless networks.


In chapter 2 work done in phase I is explained. It starts with system architecture design of phase followed by description of modules done in phase I. In chapter 3 overall system architecture designs is present. Chapter 4 detailed descriptions of various modules and algorithm involved in this project details are given.




Fig 2.1.1: Wireless Sensor Node Architecture





Collection of information from the neighbour node

Usage of routing protocols for sensor

Close proximity is maintained

Detection of target is done


Analysis of energy consumption is generated

Target position is initialised



Robotic target tracking

Sensor network target tracking

Routing in sensor network

Data aggregation

Cooperative tracking

Analysis of energy consumption

: Robotic target tracking:

In target tracking the main problem dealt with the visibility of the tracker in the system. Sometime the tracker will go out of range and the detection of the target will become difficult. So to avoid such problem of stealth tracking the boundary will be fixed and then the target privacy will be formulated. Algorithms are known for planning the tracker's motions using reactive, dynamic programming, sampling-based and approaches Thus the problem will be overcome by finding the mobile agent whenever they are out of coverage away from the environment.

4.2: Sensor Network target tracking:

Wireless sensor networks (WSNs) have been deployed to track the positions of moving vehicles humans and other moving targets. Those scrutiny systems influence stationary sensor networks in which each node collects measurements using on-board sensors and reports the measurements to the sink via multi-hop routing. On the other hand, to keep track of locations of targets sensors are attached to the moving targets. Whether stationary or attached to targets, sensor nodes passively collect measurements and rely on multi-hop communication to deliver data to the sink for further analysis. As a result, the communication can become expensive when the network size is large. The tracking architecture proposed in this work addresses such communication issues by having a mobile tracker follow a target and collect the information from the target in its vicinity. the mobile tracker query the target location by flooding the entire networks. Then a 'near-node' responds after discovering a route to the tracker. Unlike their work, our routing protocols can deliver information to the mobile tracker without route discovery or neighbor awareness. The use of mobile sensor networks, in which individual nodes have both sensing and motion capability, has also been proposed as a means to track moving targets. The primary concern is to maintain the connectivity while tracking is established.

4.3: Routing in sensor network:

Routing is a basic building block of networking so it has been studied extensively. In the area of sensor networks spanning tree based routing builds a routing tree rooted at the sink past to data delivery. Such protocols work well with stationary sinks, but are inappropriate for a mobile receiver. In the area of wireless sensor networks with mobile sinks, anticipated to influence mobility prediction to figure fresh routes to a mobile sink before old routes become useless. Their networks are formed in a grid structure and messages are first routed between cells then flooded inside the cell that contains the mobile sink. In our network, sensors are not required to predict the mobility of the sink, nor to form a hierarchical multi-tier structure. Another category of routing in sensor networks is called data-centric routing, where data is stored or searched based on their name rather than the network addresses of nodes. Essentially the data-centric routing is a query problem, whereby route discovery is initiated by queries, and then data of interest will be collected via the discovered route our robotic tracking problem does not involve queries, but focuses on delivering messages to a mobile tracker efficiently. In the area of mobile ad hoc networks (MANET), Optimized Link State Routing (OLSR) and Ad hoc On-demand Distance Vector (AODV) are two well-known routing protocols. AODV works reactively, and it discovers routes only if they are needed, while OLSR proactively maintains and updates path selection regardless of Wireless Network being used or not. Those protocols either impose high latency for the initial path setup or lead to a wasteful overhead of routing traffic, and therefore are not suitable for a network with a highly mobile receiver and battery-operated sensors. Additionally, several routing algorithms exploiting geographic information have been proposed. Those geographic routing algorithms refer to destinations by their location, and forward messages greedily, when possible, towards the destination, to a node that is closest to the destination. Such routing algorithms cannot be applied to the robotic tracking problem directly, because they require the precise location information of destinations and focus on delivering messages to the destination location. However, the goal of our routing algorithm is to deliver messages to the mobile tracker instead of a specific router.

4.4: Data Aggregation:

The goal of in-network data aggregation in WSNs is to reduce expensive transmission. Typically, the data aggregation algorithm, such as TAG , routes the aggregated values up towards the root of the routing tree with partial data aggregated at internal tree nodes. That algorithm relies on the routing tree that is established prior to the data aggregation process. The smart network concept presented in this paper is different in the sense that the in-network state computation is used to assist route selection, so that messages are delivered efficiently between the sensors near the target and the ones near the tracker.

4.5: Cooperative robot tracking:

The idea of combining WSNs with mobile robots has been studied. In particular, mobile robots are used for sensor network deployment with the goal of achieving good sensor coverage . Our work complements theirs in the sense that we focus on the tracking application after the deployment is done. WSNs are also proposed to assist mobile robots to track targets, using the sensors that can supply precise location information to the robot. We take a different viewpoint: we design the tracking algorithms by considering issues associated with both sensor networks and mobile robots; and thus, achieve good tracking performance at reasonable operational cost for the network while using simple sensor devices and robots.

4.6: Analysis of energy consumption:

Based on the tracking of the target an analysis report will be generated where different routing protocols like AODV OLSR etc. is used and then by the path selection the energy will be calculated. This value is used for calculation and then the performance analysis will be generated. the Dynamic TTL algorithm is simple to implement and provides nice tracking performance in a steady state, whereas the smart local algorithm reaches this steady state faster, and works well in complex, highly nonconvex environments.





NS-2 is an event driven packet level network simulator with C++/OTCL integration feature. Version 2 included a scripting language called Object oriented Tcl (OTCl). It is an open source software package available for both Windows 32 and Linux platforms. NS-2 has many and expanding uses included.

To evaluate that performance of existing network protocols

To evaluate new network protocols before use.

To run large scale experiments not possible in real experiments

To simulate a variety of ip networks.

NS -2 is an object oriented discrete event simulator. Simulator maintains list of events and executes one event after another. Single thread of control: no locking or race conditions Back end is C++ event scheduler.



In the simulation, there are the two tools are used.

NAM(Network Animator)


NAM (Network Animator):

NAM provides a visual interpretation of the network topology created. The application was developed as part of the VINT project. Its feature is as follows.

Provides a visual interpretation of the network created

Can be executed directly from a Tcl script

Controls include play; stop fast forward, rewind, pause, a display speed controller button and a packet monitor facility.

Presented information such as throughput, number packets on each link

X Graph:

X- Graph is an X-Window application that includes:

Interactive plotting and graphing Animated and derivatives To use Graph in NS-2 the executable can be called within a TCL script. This will then load a graph displaying the information visually displaying the information of the file produced from the simulation. The output is a graph of size 800 x 400 displaying information on the traffic flow and time.


NS2 are often growing to include new protocols. LANs need to be updated for new wired/wireless support. ns are an object oriented simulator, written in C++, with an OTCl interpreter as a front-end. The simulator supports a class hierarchy in C++ and a similar class hierarchy within the OTCl interpreter (also called the interpreted hierarchy). The two hierarchies are closely related to each other; from the user's perspective, there is a one-to-one correspondence between classes in the interpreted.

NS2 uses two languages because simulator has two different kinds of things it needs to do. On one hand, detailed simulations of protocols require a systems programming language which can efficiently manipulate bytes, packet headers, and implement algorithms that run over large data sets. For these tasks run-time speed is important and turn-around time (run simulation, find bug, fix bug, recompile, re-run) is less important.

On the other hand, a large part of network research involves slightly varying parameters or configurations, or quickly exploring a number of scenarios. In these cases, iteration time (change the model and re-run) is more important. Since configuration runs once (at the beginning of the simulation), run-time of this part of the task is less important. Ns meets both of these needs with two languages, C++ and OTCl. C++ is fast to run but slower to change, making it suitable for detailed protocol implementation. OTCl runs much slower but can be changed very quickly (and interactively), making it ideal for simulation configuration. NS (via tcl) provides glue to make objects and variables appear on both languages.




We presented a target tracking algorithm that uses collaboration between a sensorless robot and a network of unreliable sensor nodes. Simulations demonstrate that this algorithm has good performance in balancing energy efficiency with tracking accuracy, even in the presence of false positive sensor errors. However based on the multihop routing,the path selection is done effectively and hece the enery consumption is reduced to low level.It also makes the target coverage within the range so that missing is reduced.


The contribution of this paper is to propose and evaluate the tracking technique for a robot cooperating with a sensor network with the following features. Nothing other than a maximum speed is known about the target's motion. The tracking robot has no sensors that directly provide information about the target. The sensor nodes detect only when the target is nearby, but do not provide any precise location information, and unpredictable failures. Each sensor node has a limited energy budget for making transmissions. It is very fast, low-complexity transformation of the input data into accurate