Benjamin R. Hamilton

ACES: adaptive clock estimation and synchronization using Kalman filtering

Clock synchronization across a network is essential for a large number of applications ranging from wired network measurements to data fusion in sensor networks. Earlier techniques are either limited to undesirable accuracy or rely on specific hardware characteristics that may not be available for certain systems. In this work, we examine the clock synchronization problem in resource-constrained networks such as wireless sensor networks where nodes have limited energy and bandwidth, and also lack the high accuracy oscillators or programmable network interfaces some previous protocols depend on. This paper derives a general model for clock offset and skew and demonstrates its applicability. We design efficient algorithms based on this model to achieve high synchronization accuracy given limited resources. These algorithms apply the Kalman filter to track the clock offset and skew, and adaptively adjust the synchronization interval so that the desired error bounds are achieved. We demonstrate the performance advantages of our schemes through extensive simulations obeying real-world constraints.

Tracking low-precision clocks with time-varying drifts using kalman filtering

Clock synchronization is essential for a large number of applications ranging from performance measurements in wired networks to data fusion in sensor networks. Existing techniques are either limited to undesirable accuracy or rely on specific hardware characteristics that may not be available in certain applications. In this paper, we examine the clock synchronization problem in networks where nodes lack the high-accuracy oscillators or programmable network interfaces some previous protocols depend on. This paper derives a general model for clock offset and skew and demonstrates its application to real clock oscillators. We design an efficient algorithm based on this model to achieve high synchronization accuracy. This algorithm applies the Kalman filter to track the clock offset and skew. We demonstrate the performance advantages of our schemes through extensive simulations and real clock oscillator measurements.

Propagation Modeling for Radio Frequency Tomography in Wireless Networks

Non-cooperative localization is an important class of applications with the goal of detecting and tracking objects without their explicit participation. Recognizing that these objects will attenuate any radio signals that pass through them, recent works have proposed using the received signal strength (RSS) measurements from networks of portable wireless devices to detect, localize, and track objects through radio frequency (RF) tomography. Because the effects of these objects on radio signals can be relatively small, accurate radio propagation models are essential to the accuracy and reliability of these methods. Despite their critical importance, only limited work has been done to develop and test such models. In this paper, we describe the RF tomography problem in detail, comparing both existing and novel shadowing models analytically. We design an RF tomography testbed and describe the techniques used to conduct field test measurements. We present the experimental results from our RF tomography testbed comparing several different shadowing models. Further, we present the results of applying the different shadowing models to RSS measurements obtained in a field test and evaluate how well each model approximates the radio propagation experienced during our test.

OFDM Pilot Design for Channel Estimation with Null Edge Subcarriers

Wireless communication systems are increasingly adopting orthogonal frequency division multiplexing (OFDM) to enable efficient high-data rate transmissions. Such systems often employ pilot symbols to estimate wireless channels, and null subcarriers on band edges to reduce adjacent channel interference. Recent work has focused on designing the pilot sequence to improve channel estimation performance. In this paper, we present a new pilot design for OFDM systems based on a arbitrary-order polynomial parameterization of the pilot subcarrier indices. We show our design achieves better performance than existing methods. Theoretical models and simulated performance demonstrate the increased accuracy of our design.

Radio frequency tomography in mobile networks

Networks of portable wireless devices have shown enormous promise in improving communications and sensing capabilities in a large number of important applications. While much work has focused on using specialized sensors to obtain information about the physical environment, we examine the problem of exploiting the received signal strength (RSS) measurements already being used for ad-hoc networking. We describe how information about the shadowing environment is encoded into RSS measurements and present the novel RF Exploitation for Tomographic Imaging and Non-cooperative Analysis (RETINA) algorithm to detect stationary obstacles and track moving objects. We compare this algorithm with existing methods through analysis and simulation and show that RETINA achieves significantly higher accuracy.

Noncooperative Routing with Cooperative Diversity

Routing in wireless networks confronts more diverse and more rapidly varying characteristics associated with wireless links. Several recent protocols have tried to exploit cooperative diversity to deal with these channel characteristics, but they suffer from large overhead costs due to the need for collaboration among candidate nodes of next hop. We propose a novel routing protocol that is able to realize cooperative diversity without exchanging information among candidate nodes. Our protocol divides the routing decision between the transmitter and the receivers: the transmitter decides the direction and an angle spread to broadcast the packet; each receiver, without communicating each other, decides whether to start a timer whose duration depends on the strength of the channel state information. The node with shortest timer transmits first and thus becomes the transmitter for the next hop. Thus, a multi-hop route is found in a noncooperative way where the communication overhead is minimal, and we show that cooperative diversity is collected. The performance characteristics and design parameters of this protocol are first theoretically analyzed, and then examined in a realistic network setting through simulations.

G-STAR: Geometric STAteless Routing for 3-D wireless sensor networks

3-D aerial and underwater sensor networks have found various applications in natural habitat monitoring, weather/earthquake forecast, terrorist intrusion detection, and homeland security. The resource-constrained and dynamic nature of such networks has made the stateless routing protocol with only local information a preferable choice. However, most of the existing routing protocols require sensor nodes to either proactively maintain the state information or flood the network from time to time. The existing stateless geometric routing protocols either fail to work in 3-D environments or have tendency to produce lengthy paths. In this paper, we propose a novel routing protocol, namely Geometric STAteless Routing (G-STAR) for 3-D networks. The main idea of G-STAR is to distributively build a location-based tree and find a path dynamically. G-STAR not only generalizes the notion of geographic routing from two modes to one mode, but also guarantees packet delivery even when the location information of some nodes is either inaccurate or simply unavailable regardless of the uses of virtual coordinates. In addition, we develop a light-weight path pruning algorithm, namely Branch Pruning (BP), that can be combined with G-STAR to enhance the routing performance with very little overhead. The extensive simulation results by ns-2 have shown that the proposed routing protocols perform significantly better than the existing 3-D geometric routing protocols in terms of delivery rate with competitive hop stretch. We conclude that the proposed protocols serve as a strong candidate for future high-dimensional sensor networks.

Node localization and tracking using distance and acceleration measurements

Advances in miniaturized wireless and sensing technologies have enabled the construction of cheap, low-powered, portable wireless devices capable of forming ad hoc networks. While these networks have shown enormous potential in applications such as remote sensing and target tracking, these applications require the devices to determine their own location. Additionally, devices capable of self-localization can also be used to implement location-based services or to improve coordination between first-responders to disaster sites or infantry in tactical situations. Existing techniques such as GPS may not be available due to design or environmental constraints, so other methods need to be devised. Previous works have proposed methods for wireless devices to self-localize based on received signal strength (RSS), but these methods offer limited accuracy due to the large error in RSS measurements. Recognizing the trend for these portable wireless devices to contain acceleration sensors, we propose an algorithm to combine these acceleration measurements with RSS readings to achieve accurate localization. We apply a distributed extended Kalman filter to track position based on these two measurements and a kinematic node movement model. This algorithm is able to take advantage of correlations between successive location estimates to improve estimation accuracy. We calculate the posterior Cramér-Rao bound for this algorithm and analyze it through simulation. Our analysis shows that by utilizing the acceleration information, the network is able to self-localize despite the large inaccuracy in RSS readings.

Efficient node self-localization in large ad-hoc wireless networks using interlaced particle filters

Large self-organizing networks of wireless devices have shown potential in many emerging applications. While knowing the physical location of devices in enhances and enables many such applications, the locations of some or most of the wireless devices may be unknown and the devices may be mobile. Several recent works have examined using distance or RSS measurements to estimate the position of such wireless devices. In this work we examine methods to efficiently estimate and track the positions of a subset of mobile wireless nodes in the network and consider how well performance of these algorithms scale with network size. We consider position estimate error and convergence rate of these methods as the size of the network increases. We use simulations to show that a new localization method using interlaced particle filters converges faster and has higher estimation accuracy than existing algorithms.

SAL: Shadowing Assisted Localization

Ad hoc networks of inexpensive, low-powered, portable wireless devices have shown enormous potential in emerging applications. Such applications benefit greatly if the nodes can determine their own location. Traditional localization techniques such as GPS may not be available due to design or environmental constraints. Recent works have proposed using received signal strength (RSS) measurements for localization. These techniques suffer from limited accuracy since the RSS measurements are affected by fading and shadowing of the wireless channel. In this paper we propose the Shadowing Assisted Localization (SAL) algorithm, which employs a model of the shadowing environment to improve localization accuracy. This protocol is unique since it has the flexibility for localization of multi-hop networks in shadowing environments with mobile reference nodes. We show that, in shadowing environments, SAL significantly reduces the mean-squared error (MSE) of node position estimates compared to algorithms that do not consider shadowing.