Localization and angular diversity using adaptable physical layer interfaces for wireless sensor networks

Chin-Lung Yang, Purdue University

Abstract

Wireless sensor networks (WSNs) are becoming prevalent due to their diverse applications. Tiny wireless nodes equipped with different sensors can be widely spread to collect information both spatially and temporally, resulting in an interactive network which makes the world cognitive of human activities. Due to limitations of hardware and software, WSNs pose considerable challenges. One of the fundamental limitations is the radio propagation channel, the characteristics of which are not only determined by environments but also related to antenna features. An adaptable RF front end approach is proposed to overcome the impairments of wireless channels and to meet the challenges of size, cost, complexity, and power consumption in system design and implementation. The performance of WSNs can be improved by using "Arrogant" Antennas, which provide smart antenna functions on simple, limited wireless sensor nodes of low intelligence. Arrogant antennas provide better peer-to-peer link quality and overall network capacity due to the advantages of directional antennas and the nature of diversity, which provides high diversity order in a small form factor at a slightly higher cost for hardware upgrades. The impacts of semi-directional antennas on WSNs are investigated, and the benefits are demonstrated to be not merely antenna gain, but also improvement in the propagation channel statistics. Furthermore, multiple semi-directional antennas are implemented to create angularly diverse systems which provide a more reliable wireless connection and a larger margin for link budget. Performance of angular diversity enabled sensor nodes operating at 915MHz or 2.4GHz are tested both indoors and outdoors, and comparisons of different diversity techniques are obtained by indoor measurements in the aspects of diversity order, channel correlation, power imbalance, and physical feature size, which are key factors of the diversity schemes. This proposed architecture also provides extensive increased functionality to WSNs. We demonstrate the effect of using multiple estimations from independent single nodes to decrease network topology dependence on location estimation. A location determination method by using multiple semi-directional antennas is shown to give an independent estimate of location, each estimate not relying on the data from neighboring nodes as in the case of traditional triangulation. The simulation results indicate that our proposed algorithm depends significantly less on the topology (spatial arrangement) of the anchor nodes, and the result is validated by an outdoor testbed in a triangulation-adverse topology (a linear arrangement of closely spaced sensors). The position-aware sensor nodes using directional antennas can also make the routing protocols more efficient. We integrate simulation tools, including HFSS (precise antenna models), EMTerrano (realistic channels), ADS (real-world co-existent interferences), and MATLAB Simulink, which combine all components to execute network simulations. The results show that overall energy consumption is significantly reduced with a slightly longer delay and the network capacity is increased due to the interference suppression.

Degree

Ph.D.

Advisors

Chappell, Purdue University.

Subject Area

Electrical engineering

COinS