Tarun Bansal

Symphony: cooperative packet recovery over the wired backbone in enterprise WLANs

In this paper, we propose Symphony, a packet recovery architecture that encourages collisions among transmitters, and utilizes the unused capacity in the backbone to transmit recovered data packets and coordinate the efficient recovery of collided packets. Symphony improves the wireless throughput while incurring a low overhead on the typically under-utilized wired backbone. In Symphony, upon receiving the collided transmissions, the APs carefully suppress a subset of the transmissions. Realizing this idea in practice entails several challenges including identification of clients that have data to transmit and ensuring that the algorithm works despite imperfect time-synchronization and non-zero latency among APs. We present Symphony that addresses these challenges and show how it leverages Successive Interference Cancellation (SIC) to further increase the network throughput. Experiments performed on an USRP testbed shows that on an average, Symphony provides 43% and 187% higher throughput over Omniscient TDMA and IEEE 802.11, respectively. ns-3 based simulation results show that on an average, Symphony provides a throughput of up to 1.63x compared to omniscient TDMA and 5.6x compared to IEEE 802.11.

Achieving User-Level Fairness in Open-Access Femtocell-Based Architecture

Femtocell-based architectures have the potential to position the cellular service providers to compete head-on with the WiFi market. However, significant interference can happen due to unplanned deployments. Current use of hard partitioning approaches for resource allocation, and lack of guidelines for configuring the femtocells, makes it difficult to obtain significant performance gains over traditional cellular networks. In this paper, we study the dynamic OFDMA subchannel assignment problem while jointly considering power assignment and association control to provide maxmin fairness. Toward this objective, we first consider a noninterfering model (NINT model), which disallows interfering femtocells in the solution. A more general interfering model (INT model) is then considered under which the problem is transformed into the partition coloring problem. We then show the NP-hardness of the problems and design centralized approximation algorithms with provable bounds and distributed solutions. Through extensive simulations in realistic settings we show that, compared to previous work, our solutions under the NINT model can achieve two times the minimum throughput, and under the INT model the minimum throughput can be up to three times the baseline algorithms.

MARVEL: multiple antenna based relative vehicle localizer

Access to relative location of nearby vehicles on the local roads or on the freeways is useful for providing critical alerts to the drivers, thereby enhancing their driving experience as well as reducing the chances of accidents. The problem of determining the relative location of two vehicles can be broken into two smaller subproblems: (i) Relative lane localization, where a vehicle determines if the other vehicle is in left lane, same lane or right lane with respect to it, and (ii) Relative front-rear localization where it needs to be determined which of the two vehicles is ahead of the other on the road. In this paper, we propose a novel antenna diversity based solution, MARVEL, that solves the two problems of determining the relative location of two vehicles. MARVEL has two components: (i) a smartphone; and (ii) four wireless radios. Unlike exisiting technologies, MARVEL can also determine relative location of vehicles that are not in the immediate neighborhood, thereby providing the driver with more time to react. Further, due to minimal hardware requirements, the deployment cost of MARVEL is low and it can be easily installed on newer as well as existing vehicles. Using results from our real driving tests, we show that MARVEL is able to determine the relative lane location of two vehicles with 96.8% accuracy. Through trace-driven simulations, we also show that by aggregating information across different vehicles, MARVEL is able to increase the localization accuracy to 98%.

R2D2: Embracing Device-to-Device Communication in Next Generation Cellular Networks

Device-to-device (D2D) communications is being pursued as an important feature in next generation cellular networks. D2D can improve resource utilization in two ways: Offloading cellular traffic to D2D, and Reuse of resources used by conventional cellular transmissions for D2D communication. In this paper, we show that in multi-cell environments that employ FFR (Fractional Frequency Reuse), the benefits from D2D toward reuse are limited. We then propose R2D2- a holistic approach to efficient offloading with D2D traffic. R2D2 leverages the flexible nature of D2D traffic (in using downlink/uplink resources) to cater effectively to the spatial and temporal asymmetry in traffic load both across and within cells. R2D2 incorporates a two time-scale solution: a coarse time-scale dynamic FFR scheme that leverages D2D traffic to determine the FFR patterns for downlink and uplink jointly among interfering sectors; and a fine time-scale scheduling solution that intelligently schedules cellular and D2D traffic jointly across DL (Downlink) and UL (Uplink) resources. We establish the hardness of the scheduling problem and present efficient and low complexity algorithms with approximation guarantees. Through extensive evaluations, we confirm that R2D2 delivers the offloading benefits of D2D, with its proposed algorithms performing very close to the optimal.

FastProbe: Malicious user detection in Cognitive Radio Networks through active transmissions

Sensing white space channels to detect whether a particular channel is free or not is very crucial to the operation of Cognitive Radio Networks (CRNs). Cooperative sensing has been shown to improve the performance of channel sensing. However, cooperative sensing is susceptible to malicious users that may not faithfully follow sensing instructions to save energy and/or time, or to launch denial of service attacks against the network. In this paper, we propose a novel active transmissions based algorithm, FastProbe for detecting malicious users. FastProbe proactively detects malicious users before the CRN causes any interference to the Primary Users. Further, using active transmissions, FastProbe achieves higher detection accuracy while maintaining lower overheads when compared with existing algorithms. Simulations and experiments show that in the presence of malicious nodes operating under 2 different attack models, FastProbe reduces the throughput loss due to sensing by as much as 65% compared to existing algorithms.

BBN: throughput scaling in dense enterprise WLANs with Bind Beamforming and Nulling

Todays Enterprise Wireless LANs are comprised of densely deployed access points. This paper proposes BBN, an interference nulling scheme that leverages the high density of access points to enable multiple mobile devices to transmit simultaneously to multiple access points (APs), all within a single collision domain. BBN also leverages the capability of the APs to communicate with each other on the wired backbone to migrate most of the decoding complexity to the APs, while keeping the design at the mobile clients simple. Finally, we leverage the static nature of the access points to make BBN more practical in networks where the mobility of clients inhibit the use of traditional interference alignment schemes. We implement a prototype of BBN on USRP testbed showing its feasibility. The experiment results show that BBN provides a throughput gain of 1.48X over omniscient TDMA. Results from our trace-driven simulations show that BBN obtains a throughput of up to 5.6X over omniscient TDMA.

Opportunistic Channel Sharingin Cognitive Radio Networks

Licensed white space channels can now be used opportunistically by unlicensed users, provided the channels are relinquished when needed by the primary users. In order to maximize their potential, these channels need to be assigned to the secondary users in an efficient manner. The protocols to enable such an assignment need to simultaneously aim for fairness, high throughput, low overhead, and low rate of channel reconfigurations. One way of channel assignment is to allow neighboring access points (APs) to operate on the same channel. However, if not done properly, this may increase the number of collisions resulting in lower throughput. In this paper, we present a new channel assignment algorithm that performs controlled channel sharing among neighboring APs that increases not only the fairness but also the total throughput of the APs. Controlled sharing and assignment of channels leads to a new problem that we call as the Shared Coloring Problem. We design a protocol based on a centralized algorithm, called Share, and its localized version, lShare that work together to meet the objectives. The algorithm has tight bounds on fairness and it provides high system throughput. We also show how the 802.22 MAC layer protocol for wireless regional area networks (WRANs) can be modified considering the typical case of low degree of interference resulting from the operations of Share and lShare. Results from extensive ns-3 simulations based on data traces show that our protocol increases the minimum throughput among all APs by at least 58 percent when compared to the baseline algorithms.

ForeSight: Mapping vehicles in visual domain and electronic domain

Using broadcast in vehicular applications such as autonomous cruise control and collaborative driving can disturb unrelated drivers and fail to convey the message due to unspecified receiver, resulting in increased risk of accidents. For supporting the unicast communication primitive, it is important to know the electronic identities (EIDs), e.g., the IP addresses and the relative positions of the nearby vehicles. We show that the estimated GPS coordinates alone are not accurate enough to uniquely identify the intended vehicle. On the other hand, there is an increasing array of devices, such as on-board camera, RADAR, and DSRC radio that are becoming available in newer vehicles. These heterogeneously deployed devices provide information sources that have varying levels of accuracy and potentially different coverage regions, making it challenging to accurately identify the vehicle. As a first step, we design ForeSight, a system that dynamically integrates the information observed in the visual domain (e.g., from camera) and the electronic domain (e.g., WiFi radio) to match the vehicles observed in these two domains with high accuracy. The experiment and simulation results show that ForeSight is able to significantly improve the vehicle identification accuracy compared to using GPS or other algorithms. In our case study, ForeSight reduces disturbing messages by 14 × as compared to the number of a GPS-based communication method.

RobinHood: sharing the happiness in a wireless jungle

Todays Enterprise Wireless LANs are comprised of densely deployed access points. This paper proposes RobinHood, an interference nulling scheme that leverages the high density of the access points to enable multiple mobile devices to transmit simultaneously to multiple access points (APs), all within a single collision domain. RobinHood also leverages the capability of the APs to communicate with each other on the wired backbone to migrate most of the complexity to the APs, while keeping the design at the mobile clients simpler. Finally, we leverage the static nature of the access points to make RobinHood more practical in networks where the mobility of clients inhibit the use of traditional interference alignment schemes. Results from our trace-driven simulations show that RobinHood obtains a throughput improvement of 6.08x and 24.2x over omniscient TDMA and IEEE 802.11, respectively.

A scalable algorithm for maintaining perpetual system connectivity in dynamic distributed systems

We investigate the problem of maintaining a topology with small degree as well as small diameter in a dynamic distributed system such that the system always stays connected and processes that wish to leave the system can do so quickly. Perpetual system connectivity is necessary to solve many important problems in dynamic distributed systems, including atomic broadcast and stable property detection, that need strict (deterministic) guarantees about system connectivity to be solvable. To our knowledge, in all existing topology maintenance algorithms for asynchronous distributed systems that provide perpetual system connectivity, either: (i) the topology has large worst-case degree and/or diameter (ii) a process may experience high worst-case delay when leaving the system, or (iii) processes cannot join and/or leave concurrently. In this paper, we present a spanning tree maintenance algorithm that satisfies the following desirable properties. First, the spanning tree has small maximum degree of O(1) and small maximum diameter of O(log N), where N denotes the maximum size of the system. Second, any process can leave the system within O(log N) time even in the presence of concurrent arrivals and departures. Third, the system always stays connected. We show using a simple knowledge-based argument that, in any algorithm that maintains perpetual connectivity such that the topology has either worst-case diameter of Ω(log N) or worst-case degree of O(1), the departure of a process may be delayed by Ω(log log N) time in the worst-case.