Received Signal Strength Based Target Localization and Tracking Using Wireless Sensor Networks

The continuous technological upgradations in the RF (radio frequency), processors, nanotechnology, and microelectromechanical systems (MEMS) domains have fostered the growth of wireless sensor networks (WSN), which in turn allowed to develop a wide range applications based on it, for instance, the technological breakthrough in the semiconductor industry stimulated to produce low-power, low-cost, and small-sized processors with high computational capacities. Speaking in more clear words, the miniaturization of sensing and computing devices enabled the development of tiny, low-cost, and low-power sensors, controllers, and actuators. Basically WSNs consist of a large number of tiny and low-cost sensor nodes that are networked via low-power wireless communication links. These networks allow to closely observe ambient environment of interest at an economical cost much lower than other possible technological solutions. Each sensor node in WSN has sensing, communication, and computation capabilities. By exploiting appropriate advanced mesh networking protocols, these nodes form a sea of connectivity that covers the physical environmental area under observation. In WSN the transmitting node opt out possible communication paths by hopping sensed data of interest from node to node toward its destination. Although the capability of single sensor node is minimal, the composition of hundreds or thousands of such nodes offers very high new technological possibilities for wide variety of applications. The power of WSN lies in the possibility of heavy deployment of large numbers of tiny nodes, which can assemble and configure on their own. Stating in simple words, these nodes have networking capability, which facilitates coordination, cooperation, and collaboration among them to meet the requirements of the underlying application. The WSN can also provide a robust service in hostile or inaccessible environments, wherein human intervention may be too dangerous or almost not possible. This new technology is exciting with unlimited potential for numerous application areas, including environmental, medical, military, transportation, entertainment, crisis management, disaster relief operations, homeland defense, and smart spaces. It is envisioned that in the near future the WSN will be an integral as well as essential aspect of our lives.

With the recent advances in RF and MEMS field, the WSN started playing active role in target L&T applications. Basically, localization is an integral part of tracking. The prime objective of wireless-based localization is the determination of the possible locations (localization problem) of the mobile target and its trajectories (tracking problem) with the help of field measurements during the complete period of its motion. Thus, the tracking problem can be considered as the solution of a set of localization problems at successive time intervals. The target L&T in WSN can be achieved with the help of RF, infrared (IR), and ultrasound. However, the RF technology as compared to rest of the others is widely used because of its ability to penetrate smoke, nonmetallic barriers, and walls, making it a best choice for L&T applications. Although Global Positioning System (GPS)-based L&T is popular in day-to-day life, it has several drawbacks. First, the cost of the GPS receiver is approximately three times higher than the cost of a sensor node. Therefore, attaching a GPS receiver to every sensor node in the network is not at all an economical solution. Second, high power consumption is involved during the GPS operation. Third, it can only be used for outdoor environments with sufficient sky visibility. The GPS performs well for a line of sight (LOS) to several satellites; however, maintaining the LOS is generally a rare possibility especially for indoor environments. Additionally, the localization accuracy of GPS is approximately 2.5 m, which is good for navigation along the road (i.e., for outdoor environment). However, such low localization accuracy is not at all sufficient for indoor L&T applications. Therefore, the research trend is to develop innovative, low cost, and low power indoor targets tracking system, which can overcome the drawbacks of GPS-based systems. The most feasible solution in line with this trend is to use the WSN-based system for target L&T.

The RSSI-based target tracking using WSN is a very vast research domain with a wide variety of technologies involved, application of various types of Bayesian frameworks, and ANN-based implementations. In order to understand the recent research trends in these areas, the rigorous survey of existing systems is done and is split into four sections. The first section discusses the various wireless technologies useful for indoor L&T, whereas the second section is devoted to a survey of the application of Bayesian filtering in the RSSI-based L&T applications. The survey of the application of ANN as well as BLE technology in the RSSI-based L&T applications is discussed in the third and fourth sections, respectively.

One of the dominant economical approaches to L&T of the moving target with WSN is the use of RSSI. Trilateration is basically the process of obtaining the position of a target using its distances (computed using a suitable path loss model) from three anchor nodes. Three circles are formed based on these computed distances, and their intersection is used to locate the target node in space. Although the trilateration technique is not sufficient to cope up with the environmental dynamicity efficiently, it is the most basic and widely used technique in the RSSI-based target L&T domain. In this chapter, a trilateration-based L&T approach for tracking of a single mobile target, with the help of deployed WSN, is presented. There are many parameters that impact the performance of RSSI-based L&T algorithm, namely, variations in the velocity of the mobile target, anchor density, and measurement noise in the given RF environment. This chapter covers the experimentation to deal with abrupt variations in the velocity of the mobile target and uncertainties in measurement noises with the help of trilateration. During simulation experimentation the anchor density is varied from 4 to 8 in steps of 2. To understand the effect of abrupt variations in target velocity, we varied velocity abruptly in the range of −2 to 7 m/s at specific time instances. The overall target L&T performance is evaluated in terms of the localization error and RMSE. The simulation result confirms that the trilateration technique is able to track the moving target with the help of WSN, irrespective of environmental dynamicity of the given communication medium.

The KF is the one of the most widely used approaches to solve the problem of target L&T using RSSI measurements. However, very few existing KF-based research works address the issues such as abrupt changes in the target velocity and variation in the target trajectories. The existing KF-based L&T algorithms though perform well for one type of target track; it is guaranteed to perform if the target track changes or the monitoring area is enlarged. It is very difficult to achieve high tracking accuracy with the help of traditional techniques such as trilateration alone in this context. Hence, the combination of the trilateration and KF can be a very good option to deal with the abovementioned issues. This chapter presents two range-based KF algorithms, namely, trilateration+KF and trilateration+UKF, to deal with these important issues. The chapter discusses the performance of the trilateration+KF and trilateration+UKF L&T algorithms for the changes in the target velocity trajectory, target trajectory, as well as WSN monitoring area. The proposed techniques are tested for linear as well as nonlinear target trajectories. The WSN monitoring (simulation) area is varied from 100 m × 100 m to 200 m × 200 m. To understand the effect of abrupt variations in the target velocity, we varied velocity abruptly in the range of −2 to 7 m/s at specific time instances in all the three cases. The overall target L&T performance is evaluated in terms of the localization error and RMSE. The results confirmed that the proposed target L&T algorithms are able to track the moving target with the help of WSN, irrespective of the dynamicity of the given RF environment.

The traditional RSSI-based moving target L&T using WSN generally employs trilateration technique. Although being a very simple technique, it creates significant errors in localization estimations due to nonlinear relationship between RSSI and distance. To deal with such a highly nonlinear mapping between an input and an output, a suitable artificial neural network (ANN) technique can be the better alternative to achieve a high target tracking accuracy. The generalized regression neural network (GRNN) is a one-pass learning algorithm, which is well-known for its ability to get trained quickly with very few training samples. Once trained with the RSSI measurements and associated locations in the off-line phase, it can learn the dynamicity of any given indoor environment quickly to give location estimates of the mobile target in the online phase. This chapter presents an application of GRNN to solve the problem of target L&T. The GRNN can estimate the location of mobile target moving in WSN, which can be then further smoothed using KF framework. Utilizing this idea to improve target tracking accuracy GRNN+KF and GRNN+UKF algorithms is presented in this chapter. The GRNN is trained with the RSSI measurements received at mobile target from anchor nodes and the corresponding actual target 2-D locations. Extensive simulation experiments are carried out to prove the efficacy of these proposed algorithms. In Case I, the performance of GRNN-based L&T algorithm is compared with the traditional trilateration-based localization technique. In Case II, the GRNN+KF and GRNN+UKF algorithms are compared with trilateration technique, whereas in Case III, the efficacy of GRNN+KF and GRNN+UKF algorithms is compared with the previously proposed trilateration+KF and trilateration+UKF algorithms. The proposed GRNN- and KF-based target L&T algorithms demonstrate a superior target tracking performance (tracking accuracy in the scale of few centimeters) irrespective of abrupt variations in the target velocity, environmental dynamicity as well as nonlinear system dynamics.

As discussed in the previous chapters, being an interconnection of artificial neurons, the ANN is capable to mimic the behavior of biological neurons with the help of activation functions. The connections between neurons are through appropriate weights, which get automatically adjusted during the off-line ANN training step. This is called as supervised learning. In off-line training step, the training vector is necessary to train the ANN. This training vector includes a set of inputs and corresponding outputs. Once trained in off-line phase, the ANN becomes ready to estimate (predict) the system output for random input vector in the online phase. The GRNN is one of the important supervised learning architecture, which we discussed and applied to target L&T domain in the previous chapter. In this chapter, we are going to discuss the applications of other important supervised learning architectures such as feed-forward neural network (FFNT), radial basis function neural network (RBFNN), and multilayer perceptron (MLP). All of these architectures will be trained with RSSI measurements and the corresponding 2-D locations of the mobile target in off-line phase. Two cases are considered during experimentation. In Case I, the impact of various training functions on FFNT-based target L&T system is analyzed in the context of localization accuracy, whereas in Case II, the target localization performance comparison is made between traditional trilateration-, FFNT-, MLP-, GRNN-, and RBFN-based target L&T system. In order to accomplish fair comparison of all of the supervised learning architectures in both the cases, all of them are fed with four RSSI measurements, and they are supposed to estimate 2-D target location, corresponding to this input vector of four RSSI measurements. In both of the cases, locations of anchor nodes as well as target locations to be estimated are kept fixed.