Samuel Van de Velde

Human Mobility Monitoring in Very Low Resolution Visual Sensor Network

This paper proposes an automated system for monitoring mobility patterns using a network of very low resolution visual sensors (30 × 30 pixels). The use of very low resolution sensors reduces privacy concern, cost, computation requirement and power consumption. The core of our proposed system is a robust people tracker that uses low resolution videos provided by the visual sensor network. The distributed processing architecture of our tracking system allows all image processing tasks to be done on the digital signal controller in each visual sensor. In this paper, we experimentally show that reliable tracking of people is possible using very low resolution imagery. We also compare the performance of our tracker against a state-of-the-art tracking method and show that our method outperforms. Moreover, the mobility statistics of tracks such as total distance traveled and average speed derived from trajectories are compared with those derived from ground truth given by Ultra-Wide Band sensors. The results of this comparison show that the trajectories from our system are accurate enough to obtain useful mobility statistics.

Improved Censoring and NLOS Avoidance for Wireless Localization in Dense Networks

In cooperative localization, target users take advantage of neighboring users in the network to improve their position estimates. In dense networks, the number of neighbors is high and consequently a very large amount of information is available. Using all neighbors (full cooperation) results in a considerable amount of data that is to be processed and transmitted causing high network traffic, delays and reduced battery lifetime. The goal in censoring is to limit the amount of cooperation to reduce the amount of data to be transmitted, without losing (much) in positioning accuracy compared to full cooperative localization. In this paper we propose a novel censoring technique based on the Bayesian Cramér-Rao Lower Bound (CRLB) that takes into account both the uncertainties of the neighbors and the link quality in terms of LOS/NLOS. With the use of the unscented transform and a greedy search approach, the censoring can be performed accurately and at a low computational complexity.

A low resolution multi-camera system for person tracking

The current multi-camera systems have not studied the problem of person tracking under low resolution constraints. In this paper, we propose a low resolution sensor network for person tracking. The network is composed of cameras with a resolution of 30×30 pixels. The multi-camera system is used to evaluate probability occupancy mapping and maximum likelihood trackers against ground truth collected by ultra-wideband (UWB) testbed. Performance evaluation is performed on two video sequences of 30 minutes. The experimental results show that maximum likelihood estimation based tracker outperforms the state-of-the-art on low resolution cameras.

Cooperative multipath-aided indoor localization

To obtain good position accuracy with state-of-the-art indoor localization algorithms, multiple anchors must be within radio range of the user. However, anchor placement and maintenance is expensive. By reducing the number of anchors per room, overall costs are reduced. In [1], an algorithm that can estimate the position of a user using only one anchor and a priori floor plan information. However, performance for static localization was poor due to the presence of ambiguities. This paper describes a cooperative positioning algorithm that utilizes a single anchor and provides both the position and the position uncertainty for multiple users, by letting the users exchange their position information. The problem is represented with a factor graph, and belief propagation (BP) is used to extract the positions and their accuracy. The proposed cooperative algorithm leads to a significant improvement of the positioning accuracy compared to the non-cooperative method from [1].

Wi-Fi Fingerprint Positioning Updated by Pedestrian Dead Reckoning for Mobile Phone Indoor Localization

The widespread deployment of Wi-Fi communication makes it easy to find Wi-Fi access points in the indoor environment, which enables us to use them for Wi-Fi fingerprint positioning. Although much research is devoted to this topic in the literature, the practical implementation of Wi-Fi based localization is hampered by the variations of the received signal strength (RSS) due to e.g. impediments in the channel, decreasing the positioning accuracy. In order to improve this accuracy, we integrate Pedestrian Dead Reckoning (PDR) with Wi-Fi fingerprinting: the movement distance and walking direction, obtained with the PDR algorithm, are combined with the K-Weighted Nearest Node (KWNN) algorithm to assist in selecting reference points (RPs) closer to the actual position. To illustrate and evaluate our algorithm, we collected the RSS values from 8 Wi-Fi access points inside a building to create a fingerprint database. Simulation results showed that, compared to the conventional KWNN algorithm, the positioning algorithm is improved with 17 %, corresponding to an average positioning error of 1.58 m for the proposed algorithm, while an accuracy of 1.91 m was obtained with the KWNN algorithm. The advantage of the proposed algorithm is that not only the existing Wi-Fi infrastructure and fingerprint database can be used without modification, but also that a standard mobile phone is sufficient to implement our algorithm.

Comparison of message passing algorithms for cooperative localization under NLOS conditions

For large wireless networks, there is a need for accurate localization in a distributed manner. Several algorithms have been developed in order to achieve this goal. However, comparing different algorithms is hard because of the use of different network topologies and measurement models. In this paper two promising message passing algorithms, called SPAWN and SLEEP, are compared in terms of accuracy, complexity, and network traffic. To enable a fair comparison, several alterations are made to SLEEP resulting in ASLEEP with reduced network traffic and the incorporation of reference nodes. Simulations, using measurement models from real ultra-wideband equipment, show that ASLEEP is able to achieve similar estimation quality as SPAWN at much lower complexity and network traffic.

Frame Theory and Optimal Anchor Geometries in Wireless Localization

We revisit the problem of describing optimal anchor geometries that result in the minimum achievable MSE by employing the Cramer Rao Lower bound. Our main contribution is to show that this problem can be cast onto the whelm of modern Frame Theory, which not only provides new insights, but also allows the straightforward generalization of various classical results on the anchor placement problem. For example, by employing the frame potential for single-target localization we see that the directions of the anchors, as seen from the target, should optimally be as orthogonal as possible and that the existence of an optimal geometry for an arbitrary number of anchors is governed by a fundamental inequality. Furthermore, the frame-theoretic approach allows for simple derivation of some properties on optimal anchor placement that prove to be useful in a tractable approach for the more complex, multi-target anchor placement problem. In a more general sense, the paper builds a refreshing bridge between the classical problem of wireless localization and the powerful domain of Frame Theory, with far-reaching potential.

Cooperative hybrid localization using Gaussian processes and belief propagation

Wireless localization using signal strength has been very popular in commercial applications due to the wide availability of 802.11 WiFi networks. However, signal strength information alone provides very rough location estimates. In this paper we consider supplementing the receiver of each user with a ranging unit required for accurate positioning. By allowing range-based cooperation between the users, it becomes possible to increase the positioning accuracy without the need of a fully deployed network of ranging anchors. To this end, we propose a fully distributed localization algorithm that uses belief propagation for fusing signal strength and ranging information. Extensive simulations, using 3D ray tracing to provide accurate radio maps, show that the proposed fusion of measurements results in a very scalable localization solution, where the localization performance smoothly transitions in accuracy, depending on the available infrastructure.

CUPID algorithm for cooperative indoor multipath-aided localization

The main drawback today for range-based indoor localization is the requirement of a sufficient amount of fixed reference nodes within radio range of the user. However, these reference nodes, called anchors, are expensive and require professional maintenance. Using ultra-wideband in an indoor environment, the number of anchors can be reduced to one when reflections are taken into account. With the help of a floorplan it is possible to obtain a set of virtual anchors that can be associated with the reflections. In this paper, a low-complex two-step algorithm is proposed that is able to accurately estimate the user positions using a single anchor. In a first step, the algorithm tries to estimate a number of rigid structures using the noisy inter-node distances and tries to fit this structure in the room by exploiting the measured reflections. It is shown that the presented algorithm can provide positioning accuracy similar to multi-anchor localization algorithms, even in scenarios with many unwanted scatterers and non-line-of-sight.

On the effect of movement on the localization accuracy in practical wireless networks

The problem of localization involves estimating the position of a user from a number of noisy sensor measurements. In a practical wireless network, these sensor measurements cannot be collected instantaneously and arrive after a certain delay. In a dynamic scenario where the users move around, this delay will render some measurements out-dated and, if not taken into account, have a negative effect on the localization performance. This paper consists of two parts, in the first part we investigate the effect of user movement on the measurement models for ranging. In the second part we use these models to analyze the impact of movement on the accuracy of the position estimate by means of the Cramer-Rao lower bound (CRLB) which bounds the performance of the estimation.