Vinay Kolar

Dynamic localization control for mobile sensor networks

Localization is a fundamental operation in mobile and self-configuring networks such as sensor networks and mobile ad hoc networks. For example, sensor location is often critical for data interpretation. Existing research focuses on localization mechanisms: algorithms and infrastructure designed to allow the sensors to determine their location. In a mobile environment, the underlying localization mechanism must be invoked repeatedly to maintain accurate location information. We propose and investigate adaptive and predictive protocols that control the frequency of localization based on sensor mobility behavior to reduce the energy requirements for localization while bounding the localization error. In addition, we evaluate the energy-accuracy tradeoffs. Our results indicate that the proposed protocols reduce the localization energy significantly without sacrificing accuracy.

A multi-commodity flow approach for globally aware routing in multi-hop wireless networks

Routing in multi-hop wireless networks is typically greedy, with every connection attempting to establish a path that minimizes its number of hops. However, interference plays a major role in limiting the capacity of such networks; this effect is ignored by most existing protocols. It is likely that approaches that coordinate routing to account for mutual interference would be able to achieve better performance than traditional approaches. Modeling routing with interference constraints is a complex non-linear optimization problem. We approach the problem using a multi commodity flow (MCF) formulation. We analyze the interaction of multiple routes and propose effective objective functions which attempt to maximize interference separation while limiting path inflation. Initial experimental results show significant improvement in performance over a traditional routing protocol. We evaluate the formulation against routes obtained using DSR under several scenarios and show that better performance is achieved in terms of throughput, goodput, and end-to-end delay

Enhancing cognitive radios with spatial statistics: From radio environment maps to topology engine

Radio environment maps are a promising architectural concept for storing environmental information for use in cognitive wireless networks. However, if not applied carefully their use can lead to large amounts of measurement data communicated over wireless links, causing substantial overhead. We propose enhancing the basic radio environment map concept by spatial statistics and probabilistic models, enabling applications to benefit from environment data while reducing overhead. In this paper we discuss the development of a topology engine, an agent in the CWN collecting and processing spatial information about the environment for storage in the REM. We discuss both technical and architectural issues in enabling such an approach, and outline some of the potential application scenarios for the topology engine.

Avoiding head of line blocking in directional antenna [MAC protocol]

In existing directional MAC protocols a single queue is used at the MAC layer; this is inherited from omnidirectional implementations. However, while there is a single channel state in omnidirectional transmission (either the channel is busy or not), the state of the channel varies with the desired direction of transmission in directional antennas. Thus, existing implementations which use a single FIFO queue potentially lead to head of line blocking if the medium is busy in the direction of the packet at the top of the queue but is available in other directions. We propose a new queuing organization which could take advantage of the channel more effectively using the underlying antenna system by eliminating head of line blocking. We also identify a problem with the directional virtual carrier sense implementation due to side-lobes and provide a solution to it. Our results indicate that by using a greedy approach, to schedule the packet which has the least wait time, increases the overall throughput and reduces end-to-end delay considerably, especially under heavy loads.

Measurement and Analysis of Link Quality in Wireless Networks: An Application Perspective

Estimating the quality of wireless link is vital to optimize several protocols and applications in wireless networks. In realistic wireless networks, link quality is generally predicted by measuring received signal strength and error rates. Understanding the temporal properties of these parameters is essential for the measured values to be representative, and for accurate prediction of performance of the system. In this paper, we analyze the received signal strength and error rates in an IEEE 802.11 indoor wireless mesh network, with special focus to understand its utility to measurement based protocols. We show that statistical distribution and memory properties vary across different links, but are predictable. Our experimental measurements also show that, due to the effect of fading, the packet error rates do not always monotonically decrease as the transmission rate is reduced. This has serious implications on many measurementbased protocols such as rate-adaptation algorithms. Finally, we describe real-time measurement framework that enables several applications on wireless testbed, and discuss the results from example applications that utilize measurement of signal strength and error rates

Modeling and analysis of two-flow interactions in wireless networks

Interference plays a complex and often defining role in the performance of wireless networks, especially in multi-hop scenarios. In the presence of interference, Carrier sense multiple access MAC protocols are known to suffer from the hidden terminal and exposed terminal problems, which can cause poor performance and unfairness. In this paper, we examine the possible interference modes arising among two interfering one-hop connections under a two-disc model of interference. We classify the large set of resulting configurations into five categories and develop closed form expressions to compute their probability of occurrence. The analysis exposes two new categories, whose occurrence is common, and whose behavior differs significantly from the three known interference categories. Further, the frequency of occurrence of the categories differ significantly from existing results (obtained with a simpler unit disc model of interference). We develop throughput estimation models for the different categories and validate them using simulation.

How do wireless chains behave?: the impact of MAC interactions

In a Multi-hop Wireless Networks (MHWN), packets are routed between source and destination using a chain of intermediate nodes; chains are a fundamental communication structure in MHWNs whose behavior must be understood to enable building effective protocols. The behavior of chains is determined by a number of complex and interdependent processes that arise as the sources of different chain hops compete to transmit their packets on the shared medium. In this paper, we show that MAC level interactions play the primary role in determining the behavior of chains. We evaluate the types of chains that occur based on the MAC interactions between different links using realistic propagation and packet forwarding models. We discover that the presence of destructive interactions, due to different forms of hidden terminals, does not impact the throughput of an isolated chain significantly. However, due to the increased number of retransmissions required, the amount of band-width consumed is significantly higher in chains exhibiting destructive interactions, substantially influencing the over-all network performance. These results are validated by testbed experiments. We finally study how different types of chains interfere with each other and discover that well behaved chains in terms of self-interference are more resilient to interference from other chains.

Coverage in visual sensor networks with Pan-Tilt-Zoom cameras: The MaxFoV problem

We consider the problem of target coverage in visual sensor networks with Pan-Tilt-Zoom (PTZ) cameras. The finely controllable movement in PTZ dimensions creates a large number of possible Field-of-View (FoV) settings, making it prohibitively expensive to consider them all in coverage algorithms. However, these FoVs are redundant as each group of targets is generally covered by many FoVs. Thus, an important problem is, how to identify FoVs that cover all maximal subsets of targets (MaxFoV) efficiently? We show that MaxFoV is an instance of generating all maximal cliques, which is NP-hard in general but polynomial if the number of cliques is polynomial. We construct an optimal algorithm to solve the problem with a worst case complexity of O(n3). Simulation and testbed experiments show that the algorithm drastically reduces the number of FoVs allowing multi-camera coverage to scale without sacrificing coverage quality.

Link quality analysis and measurement in wireless mesh networks

Protocols and applications in wireless mesh networks often optimize their performance by measuring the quality of wireless links. However, measuring and characterizing link-quality is a challenging task due to the nature of wireless channel and device-specific properties of radios. The paper proposes two aspects of link-quality measurement and estimation in realistic networks that benefit higher-layer protocols. First, we analyze the statistical properties of link-quality metrics, such as received signal strength and packet error rates, in an indoor IEEE 802.11 mesh network. We show that the statistical distribution and memory properties vary across different links, but are predictable. The next contribution of the paper is a real-time measurement framework that enables higher-level protocols in wireless mesh networks. We discuss the architectural requirements and our implementation experiences of a measurement framework. In addition, we provide three concrete applications that use the measured link-quality and statistical inference to better adapt their behavior.

Cognitive Radio for Home Networking

Cognitive Radios have emerged as one the most promising methods to increase wireless system efficiency through dynamic spectrum access combined with other cross-layer optimization methods. Most of the research prototypes and demonstrations have so far focused on either general platforms or scenarios that are predominantly taken from military or emergency communications domain. In this demonstration we show the prototype environment that is build around realistic home networking scenarios. The demonstration has two purposes. First, it demonstrates how a set of different implemented and integrated components can achieve local area optimization both in frequency allocation and other domains. Second, it shows the viability and attractiveness of cognitive radio methods for future commercial home networking devices. The demonstration showcases dynamic spectrum allocation and policy based behavioral changes in a home environment, where several multimedia stream and data communication connections are competing against each other.