Jorge Juan Robles

3D fingerprint-based localization for wireless sensor networks

The position information of the nodes in a WSN has become crucial for many advanced features such as routing, clustering, and context-based applications. If the localization algorithm is distributed, the sensor nodes are able to calculate the position on their own without the help of a central computer. In this case the design of localization algorithms is a challenge due to the fact that the nodes are limited in relation to the memory capacity, processing power and energy consumption. In this paper we analyze the suitability of a distributed fingerprint-based localization algorithm, where sensor nodes store relevant information called fingerprints in order to use this knowledge during the position calculation. The localization algorithm has been evaluated in a three dimensional scenario using the framework MIXIM of the simulator tool OMNET++.

Extended Min-Max algorithm for position estimation in sensor networks

A localization algorithm that is often used by sensor nodes is Min-Max [1] [2]. This algorithm can be easily executed due to the fact that it principally consists of few additions, subtractions and logical comparisons. However, Min-Max provides a coarse position estimation. In our proposal we improve the accuracy of Min-Max by including simple extra operations. We compare the accuracy of our extended Min-Max (E-Min-Max) with other algorithms by using simulation.

A low-power scheme for localization in wireless sensor networks

One of the most challenging issues in the design of system for Wireless Sensor Networks (WSN) is to keep the energy consumption of the sensor nodes as low as possible. Many localization systems require that the nodes keep the transceiver active during a long time consuming energy. In this work we propose a scheme to reduce the energy consumption of the mobile nodes that need to know their positions. Our strategy consists of decreasing the idle listening and an optimized allocation of the localization tasks on the nodes. Thus, the nodes that are externally powered calculate the position for the resource-constrained nodes. The scheme is based on a low-power IEEE 802.15.4 nonbeacon enabled network.

Enabling low-power localization for mobile sensor nodes

One of the most challenging issues in the design of localization system is to maximize the battery lifetime of the mobile nodes as much as possible. Several localization algorithms exploit the received signal strength indicator (RSSI) for determining the position. In this paper the RSSI samples acquisition is investigated from an energetic point of view. For enabling energy efficient localization two novel protocols are proposed: Low Power Localization (LPL) and Optimized Listening Period (OLP). A non-beacon enabled IEEE 802.15.4 sensor network was built in order to take real measurements and allow a comparison.

Performance evaluation of an indoor localization protocol in a 802.15.4 sensor network

Nowadays there are many indoor localization systems, which use a sensor network as general platform. In this work, the protocol for localization “Highly Configurable Protocol” (HCP) [1] was implemented in a 802.15.4 test-bed forming a tree topology. Our goal was to empirically investigate how the limitations of a IEEE 802.15.4 sensor network affect the performance of HCP in the localization process, principally in the transmission of localization information across the network. The obtained results allowed to identify not only the causes of packet losses in the network, but also possible improvements to increase the efficiency of HCP and the scalability of the localization system.

A high configurable protocol for indoor localization systems

We present a novel protocol for IEEE 802.15.4 beacon-less sensor networks, which provides support to RSSI-based localization algorithms to estimate the position of mobile nodes. The proposed protocol is designed to allow the application to change the mobile node operation mode on demand. In this way it is possible to manage the trade-off between energy consumption and achieved position accuracy. First results related to the parameterization of the protocol and the expected mobile node energy consumption are presented.

Considerations in the design of indoor localization systems for wireless sensor networks

The design of an indoor sensor network to support localization-based services is a challenging issue in that the protocols and localization algorithms have to be adapted to the capabilities of resources-constrained sensor nodes. Furthermore, the achieved position accuracy depends on many factors like the amount of signaling and quality of inter-node measurements, as well as the mathematical algorithm used for estimating the position. In this paper we present experiences and considerations in the design of such systems for IEEE 802.15.4 indoor sensor networks.

Adaptive selection of Anchors in the Extended Kalman Filter tracking algorithm

In many localization systems, the Mobile Node (MN) takes distance measurements with reference nodes called Anchors (ANs) in order to estimate its position. In general, the MN can obtain a better estimation when it takes measurements with multiple ANs. Unfortunately, this can lead to consume more energy and generate more traffic in the network. In this paper, we present an adaptive mechanism for the localization algorithm Extended Kalman Filter. Here, the MN decides the number of ANs to use according to measurable error indicators, which can be used to have an idea about the MNs position error. In this way, if the error indicator suggests that the position error was high in previous periods, then our Selective Extended Kalman Filter (S-EKF) will take measurements with more ANs in the next periods to improve the position accuracy.

Node degree improved localization algorithms for ad-hoc networks

In this paper we improve the well-known localization algorithms Lateration, Weighted Centroid Localization and Min-Max by using a improved distance estimation. It does not only consider the hop count between two nodes, but also the neighbor degree information. Simulation studies show the performance improvements.