Cognitive Indoor Positioning and Tracking using Multipath Channel Information

This thesis presents a robust and accurate positioning system that, like the visual brain, adapts its behavior to the surrounding environment by filtering out clutter and focusing on relevant activity and information. In indoor environments, which are characterized by harsh multipath propagation, achieving the necessary level of accuracy while meeting reasonable infrastructural needs remains challenging. In such environments, it is essential to distinguish relevant information from irrelevant information and develop an appropriate uncertainty model for positioning measurements. This thesis aims to achieve this objective by incorporating four basic principles of human cognition?the perception-action cycle (PAC memory, attention, and intelligence?into positioning systems. To incorporate these principles, multipath-assisted indoor navigation and tracking (MINT) concepts are integrated with cognitive dynamic systems (CDS) principles developed by Simon Haykin and his colleagues. MINT exploits specular multipath components (MPCs), which can be associated with local geometry using a known floor plan. Thus, MPCs can be viewed as signals from virtual sources, or virtual anchors (VAs), which are mirror images of physical anchors with respect to features of a floor plan. Thus, additional position-related information contained in the radio signals is exploited. This information is quantified using the Cramer-Rao lower bound (CRLB) of the position error in a geometry-based stochastic channel model (GSCM), which accounts for geometry-dependent MPCs and stochastically modeled diffuse/dense multipath (DM). This demonstrates that the signal-to-interference-plus-noise ratio (SINR) quantifies the amount of position-related information. However, inaccuracies in the floor plan, and the resulting uncertainties in the VAs, are not considered at this stage. Therefore, this thesis introduces probabilistic MINT, which aims to (i) remove the requirement of a precisely known a priori floor plan and (ii) cope with uncertainties in environmental representation. In probabilistic MINT, the VAs are included in a geometry-based probabilistic environment model (GPEM). In the next step, this algorithm is extended to a probabilistic, multipath-assisted, feature-based simultaneous localization and mapping (SLAM) approach that can operate without prior knowledge of the floor plan. The GSCM and GPEM represent the built-in memory of the developed cognitive positioning system. In contrast, the algorithm itself executes attention by enabling separation between relevant and irrelevant information and focusing on memorized model parameters. Closing the PAC with transmit waveform adaptation based on a cognitive controller supports this separation process. It also facilitates (i) gaining new position-related information from the surrounding environment and (ii) suppressing additional noise. The interplay of these characteristics is key to the intelligent behavior of the cognitive positioning algorithm.