Autor(es): Santos, João Carlos Bernardo Oliveira Alves
Data: 2016
Identificador Persistente: http://hdl.handle.net/10362/31871
Origem: Repositório Institucional da UNL
Assunto(s): AoA; GTRS; RSS; SDP; wireless localization; WSN; Domínio/Área Científica::Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática
This dissertation addresses one of the most important issues present in Wireless Sensor Networks (WSNs), which is the sensor’s localization problem in non-cooperative and cooperative 3-D WSNs, for both cases of known and unknown source transmit power PT . The localization of sensor nodes in a network is essential data. There exists a large number of applications for WSNs and the fact that sensors are robust, low cost and do not require maintenance, makes these types of networks an optimal asset to study or manage harsh and remote environments. The main objective of these networks is to collect different types of data such as temperature, humidity, or any other data type, depending on the intended application. The knowledge of the sensors’ locations is a key feature for many applications; knowing where the data originates from, allows to take particular type of actions that are suitable for each case. To face this localization problem a hybrid system fusing distance and angle measurements is employed. The measurements are assumed to be collected through received signal strength indicator and from antennas, extracting the received signal strength (RSS) and angle of arrival (AoA) information. For non-cooperativeWSN, it resorts to these measurements models and, following the least squares (LS) criteria, a non-convex estimator is developed. Next, it is shown that by following the square range (SR) approach, the estimator can be transformed into a general trust region subproblem (GTRS) framework. For cooperative WSN it resorts also to the measurement models mentioned above and it is shown that the estimator can be converted into a convex problem using semidefinite programming (SDP) relaxation techniques.It is also shown that the proposed estimators have a straightforward generalization from the known PT case to the unknown PT case. This generalization is done by making use of the maximum likelihood (ML) estimator to compute the value of the PT . The results obtained from simulations demonstrate a good estimation accuracy, thus validating the exceptional performance of the considered approaches for this hybrid localization system.
