Research On Connectivity-based Node Localization In Wireless Sensor Networks

Wireless sensor networks(WSNs), which consist of a large number of small-sized sensor nodes each capable of sensing, processing and transmitting environment information, have many applications in various domains, such as target tracking, environmental monitoring, battle field surveillance and etc. In these applications, the sensors' measurements may become useless without the information of nodes' position. Furthermore, location information is also important in many networking protocols, such as deployment, coverage, routing and etc.Localization algorithms for multi-hop wireless sensor networks can be divided into two categories: range-based and range-free approaches. In range-based localization, each node is assumed to be able to measure the distances between itself and its neighboring nodes based on some ranging technique. These techniques, however, often require sophisticated hardware for measurements, or suffer from inaccurate transmission models. On the other hand, in range-free localization, nodes generally have no ability to measure the distances to their neighbors, but only with the connectivity information. Compared with the range-based localization, range-free localization methods are suitable for large scale wireless ad hoc and sensor networks due to their less-demanding hardware requirements.This paper mainly focuses on the range-free localization algorithm and first reviews some recent advances of range-free localization for different network scenarios. And it then presents some improvement approaches for the connectivity-based node localization algorithm, including the DV-Hop and DV-RND algorithm, and finally evaluates their localization performance via simulation. The main research work of this paper is as follows:1) This paper proposes an anchor selection algorithm based on minimum residual distance to improve the original localization methods. It discusses the impact of the number of selected anchors and examines the proposed anchor selection method for two classical connectivity-based localization algorithms, the DV-RND and DV-Hop. This paper then evaluates the localization performance via simulation in three different networks, uniform deployment, non- uniform deployment and uniform network with coverage hole. The simulation results show that it can in general increase the localization accuracy.2) The original DV-RND algorithm uses the number of neighbor nodes to estimate the distance. While for a boundary node, the number of its neighbor nodes may decrease, which would increase the estimation error of the RND. To reduce the impact of boundary effect, this paper first presents a boundary compensation algorithm based on the probability of valid area to fix the estimation of RND. This paper also presents the BCAS algorithm, the combination of boundary compensation and anchor selection. The simulation results show that the BCAS algorithm can achieve great localization performance.3) The ARND approach uses the Packet Reception Ratio(PPR) to define the neighbor nodes. It needs to transmit more packets to get an accurate PPR, which on the other hand increase the energy cost. This paper analyses the impact of the transmitting power to the PPR and RND, and proposes a localization method based on RND under limited energy supply. For constant total energy consumption, an appropriate transmitting power selection algorithm is presented to obtain more precious estimated position. Simulation results show that it can improve the localization accuracy.