Research

From CORES
Jump to: navigation, search

Massive MIMO in Interference Management in Cognitive Radio Networks

MassiveMIMO.png

In coming years, wireless networks have to serve a large number of data hungry devices due to proliferation of new devices and applications every year. Increasing number of wireless devices (users) are competing for physical layer resources, namely time, space and frequency, to download the data required for different applications. In order to serve a large number of users, the wireless network needs to either create new resources or utilize existing resources more efficiently.

We propose to create spatial resources in the channel by using massive MIMO technique in a cognitive cell. The cognitive cell coexists with the primary cell in underlay paradigm. In underlay CR network, the cognitive base station (CBS) can serve CRs if the interference to the primary users (PUs) is kept below a threshold. By using a large antenna array at CBS, cognitive cell simultaneously serves a large number of CRs, while restricting the interference to PUs below a specified threshold. In this project, we investigate how the network parameters, namely number of antennas at CBS (Mb), number of antennas at CRs (Mu), and number of PUs affect the ability of CBS to serve a large number of CRs. We consider three constraints in this study: 1) maximum interference at PU (I0), 2) minimum rate required by each SU (R0) , and 3) maximum transmit power (P0). The imperfect knowledge of the channel between CBS and PU is assumed in our approach.

Staff

Selected publications

  • S. Chaudhari and D. Cabric, "Downlink Transceiver Beamforming and Admission Control for Massive MIMO Cognitive Radio Networks", Asilomar Conference, May 2015.
  • S. Chaudhari and D. Cabric, "Feasibility of Serving K Secondary Users in Underlay Cognitive Radio Networks using Massive MIMO", in proc. of 11th international ITG conference on Systems, Communications, and Coding (SCC), Feb. 2017
  • S. Chaudhari and D. Cabric, "QoS Aware Power Allocation and User Selection in Massive MIMO Underlay Cognitive Radio Networks", submitted to IEEE Transactions on Cognitive Communications and Networking, May 2017.

Load Balancing and Interference Management for Heterogeneous Cellular Networks

Hetnet.png

With the proliferation of wireless devices and the advent of bandwidth hungry applications such as video streaming, cloud-based technologies, etc., incremental improvements to existing networks and architectures such as 4G are insufficient to meet the projected data demands in the next five years. This has called for paradigm shifts in next generation networks, i.e., 5G, and particularly, the ongoing evolution towards very dense and unplanned deployment of low-power small cell base stations (BSs) of various types, commonly known as HetNets.

Small cells can significantly enhance network’s throughput, but there are several key challenges associated with the deployment of HetNets such as load-balancing and the high interference. For instance, low-power BSs can be overshadowed by the high transmit power of the macro BS, rendering existing user association techniques such as the max-power user association impractical due to the load-imbalance it causes in the network. Similarly, co-channel deployment of these cells prohibitively increases the interference, degrading the quality-of-service especially for users located at cell edges.

We propose load-aware user association policies that significantly improve the user’s throughput by balancing the load for different network settings. In addition, we study interference not only via the allocation of time-frequency resources but also via the allocation of spatial resources based on the premise that massive MIMO will be an integral technology in 5G.

Staff

Selected publications

  • G. Hattab and D. Cabric, "Joint Resource Allocation and User Association in Multi-Antenna Heterogeneous Networks", in IEEE Globecom, Dec. 2016.
  • G. Hattab and D. Cabric, "Inter-tier Interference Mitigation in Multi-Antenna HetNets: A Resource Blanking Approach", in IEEE Globecom, Dec. 2016.
  • G. Hattab and D. Cabric, "Load-aware User Association for Downlink 5G Heterogeneous Networks", submitted to IEEE Trans. on Wireless Commun., Jan. 2017.

Cooperatively Learning Locations and Footprints of Multiple Coexisting Primary Users

Network.png
Topology.jpg

This project focuses on developing cooperative algorithms for the identification of spectrum holes and classification of primary network activity by geographically large-scale cognitive radio networks. The project consists of three parts, namely localization, footprint learning and network topology learning.

The knowledge of PU location is required in a cognitive radio (CR) network for advanced spectrum sharing, e.g., location-based interference management. Our proposed Cyclic Weighted Centroid Localization (WCL) algorithm uses the distinct cyclostationary properties of the PUs in order to estimate the location of each. The proposed Cyclic WCL reduces the localization error by 5x as compared to traditional WCL when the interferer's transmit power is 40 dB higher than the target.

When multiple CRs and PUs coexist in the same spectrum, different sets of CRs may receive signals from each PU. Learning these footprints of each PU is necessary for disambiguating the PU signals which, in turn, is necessary for computing individual PUs’ radio environment map, location, and for signal classification. To increase versatility of our proposed algorithms, we use only the received energy measurements at the CRs as input. Furthermore, they do not rely on knowledge of the channel propagation models or the location of any radios, PUs or CRs. Their ability to distinguish the signals from each PUs can be used to extend existing algorithms, such as WCL, designed for single PU systems.

The learned footprints and the detected spectrum occupancy will be used to analyze the PU activity over time. The PU activity will be used to identify sets of PUs that are part of the same network. Learning the network topology will improve the spatial resolution of the spectrum holes by identifying the receiver of each PU transmission.

Staff

Selected publications

  • S. Chaudhari and D. Cabric, "Cyclic Weighted Centroid Algorithm for Transmitter Localization in the Presence of Interference", accepted to IEEE Transactions on Cognitive Communications and Networking.
  • S. Chaudhari and D. Cabric, "Cyclic Weighted Centroid Localization for Spectrally Overlapped Sources in Cognitive Radio Networks", Global Communications Conference (GLOBECOM), 2014 IEEE , vol., no., pp.935-940, 8-12 Dec. 2014.
  • M. Laghate and D. Cabric, "Identifying the Presence and Footprints of Multiple Incumbent Transmitters", in 2015 Asilomar Conference on Signals, Systems, and Computers, Nov. 2015, pp. 146-150.
  • M. Laghate and D. Cabric, "Cooperatively learning Footprints of Multiple Incumbent Transmitters by Using Cognitive Radio Networks", Submitted to IEEE Transactions on Cognitive Communications and Networking, Sept. 2016

Robust Wideband Spectrum Sensing under Interferers

Impairments that affect the spectrum sensing performance
Proposed architecture

Wideband spectrum sensing significantly reduces the sensing time of detecting unoccupied frequency bands. As a result, the throughput of secondary user network increases. However, the appearance of interferers (detrimentally strong signals) severely degrades wideband spectrum sensing performance of current cognitive radio (CR) prototypes. Reasons of such degradation includes intermodulation (IMD) interference and power leakage in power spectrum density (PSD) estimation.

In this project, we aim to design a CR receiver architecture and algorithm for wideband spectrum sensing with interferers. The following two techniques are under investigation for robust sensing: (1) Interferers identification that quickly identifies dominant interferers in a wideband spectrum and feedforwards interferers’ frequencies to other system modules, i.e., reconfigurable RF notch filters and DSP, in order to suppress interferers’ impacts. The technique requires low ADC sampling rate, and it features low computational complexity. The saving comes from utilizing a-priori knowledge of interferers’ bands and the sparsity of dominant interferers. (2) Adaptive IMD compensation: once the dominant interferers have been identified, our system decodes these signals and digitally reconstructs IMD interference. An algorithm is developed to adaptively compute the optimal compensation weight and mitigate IMD from distorted signal. Simulation results demonstrate a significant improvement in spectrum sensing performance with IMD compensation.

Staff

Selected publications

  • H. Yan and D. Cabric, "Digitally Enhanced Inter-modulation Distortion Compensation in Wideband Spectrum Sensing", accepted to Asilomar Conference on Signals, Systems, and Computers, May. 2016

Energy Efficient Blind Automatic Modulation Classification

Spectral correlation function of overlapped BPSK, QPSK, and MSK
Higher order modulation classifier architecture

Modulation classification is becoming an increasingly important subject in military applications and more recently in the context of cognitive radio to detect the type of the licensed primary user. Furthermore, future public safety systems require modulation classification to distinguish between different types of early responders and coordinate communication in a better way. With that in mind, we are exploring blind cyclostationary-based energy efficient algorithms for modulation classification that would help unlicensed users in the detection of the primary user. We are investigating the classification of overlapped signals using a single receive antenna for a low hardware complexity system (see Spectral Correlation Function for overlapped BPSK, QPSK, and MSK signals). Given that this method doesn't differentiate among different levels of the same modulation class, we propose a hierarchical classifier which first estimates the modulation class, and then finds the modulation level.

In the same spirit, we are also investigating highly energy efficient algorithms for accurate modulation classification that distinguish among higher-order modulations. The novel techniques utilize different statistical distribution distance tests such as the Kuiper test and the Kolmogorov-Smirnov test with the goal of improving the performance of current techniques based on cumulants.

Staff

Selected publications

Self-Healing Mixed-Signal Baseband Processor for Cognitive Radios

Healing.png

Cognitive radios (CRs) are expected to emerge as a critical component in military and commercial applications. These radios are addressing the fact that spectrum is actually poorly utilized in many bands, in spite of the increasing demand for wireless connectivity. On a conceptual level, cognitive radio networks sense the spectral environment and adapt transmission parameters to dynamically reuse available spectrum. This is not just a hypothetical concept, there is actual evidence that the FCC supports this technology and is currently working on the rules for cognitive radio operation in licensed TV bands. The FCC’s interest also extends to higher frequencies, where the spectrum utilization is even lower. The realization of this vision could open up to 100 GHz of spectrum and a new frontier of opportunities for radio designers and wireless application developers. However, the novelty of this approach requires new mechanisms for using radio frequencies through sharing rather than fixed allocations.

A challenging task in cognitive radios is spectrum sensing, i.e., determining which channels are not occupied. In the presence of shadowing, the system must identify signals that are as much as 20 dB below noise. Moreover, this detection has to take place in a very short amount of time, in order to vacate the frequency band before generating any significant amount of interference. Our project focuses on the implementation of a baseband spectrum sensing processor using energy detection and pilot detection.

Staff

Selected publications

  • T.-H. Yu, S. Rodriguez-Parera, D. Marković, D. Čabrić, Cognitive Radio Wideband Spectrum Sensing Using Multitap Windowing and Power Detection with Threshold Adaptation, in Proc. IEEE International Conference on Communications (IEEE ICC), 23-27 May 2010, Cape Town, South Africa

DAPRA logo.jpg

Joint Spectrum Sensing and Medium Access Control

File.png

Cognitive Radio (CR) is highlighted as a future technology for wireless communications, expected to solve spectrum scarcity problem by allowing secondary use of spectrum which is detected as idle. For this brand-new technology, we face features that are not considered in traditional design approach: the spectrum sensing (SS) and the opportunistic spectrum access (OSA). The SS is a feature to detect if a channel is idle or busy reliably, and the OSA is another feature for effective utilization of the bandwidth that may be dynamically changing depending on the activity of the Primary User (PU) who has primary right to access channels. In addition, because the effective utilization is based on reliable detection of the idle bandwidth, we also have to deal with joint effect between the SS and OSA. To design the CR with solving these design problems, we need a cohesive design approach which can capture the interaction between the SS and OSA. Moreover, we want to apply any specific option for the SS or the OSA to this comprehensive design approach, so that we can evaluate and design any option proposed.

Therefore, we set up a framework to design the CR by jointly optimizing the SS and the OSA. Also, by deriving parameters in the framework based on specific options of the SS and the OSA, we can design the CR. As a first step, we apply this framework to small ad hoc network. Assuming one-hop network, we apply the collaborative sensing as a specific scheme for the SS, which is a well-known sensing strategy to make a decision of the channel availability based on sensing results of multiple users. Moreover, we assume that one channel is assigned to each user and the bandwidth of connection is not changed during connection. With this assumption, we consider two OSA strategies to deal with sudden PU appearance: buffering and switching. If a PU appears on a channel that is being utilized for data transmission, a CR user may terminate connection (non buffering with non switching), buffer data until the PU disappears (buffering with non switching), switch to other channel that is detected idle (non buffering with switching), or switch to idle channel if possible and otherwise buffer data (buffering with switching).

Staff

Selected publications

Reliable Wideband Spectrum Sensing under Strong Interference Environments in Cognitive Radios

Spectrum hole.png

Issues of low signal-to-noise ratio (SNR) and hidden terminal problems in a cognitive radio (CR) system are quite common and have been well addressed before. Most previous literature pursues more spatial diversity from cooperative networks to raise the detection probability and obtain more access opportunity. In the situation, the interference model for the CR system could be as similar as in the figure where the hidden PU is surrounded by strong interference (could be higher than 60 dB) at near frequency subbands. This could severely damage the detection performance (especially miss detection) even with a great sensing technique. Besides, to provide a more flexible and faster sensing, efficient wideband sensing methods instead of single band detection become necessary and have drawn much attention in these years.

In this project, the aim is to develop a reliable and efficient wideband spectrum sensing system that is applicable to the strong interference environments. (1) Wideband Sensing: The critical challenge for a wideband sensing system is to reliably identify available spectral holes with the required spectral resolution in an efficient way, both in system algorithm and in hardware architecture design. To solve these problems, an adaptive multitaper spectral detector has been introduced with the advantages of robust multitaper gains, highly adjustable spectra-resolution capability, and delicate weighting mechanism. (2) Hidden PUs detection: Hidden PUs detection under wideband sensing faces crucial sensing problems under "ultra low SNR" conditions, which is caused by colored interference, channel fading, and additive white Gaussian noise (AWGN). Rather accurate interference power estimation results (usually they are not achievable due to the uncertainty of noise or knowledge of interference) are required by conventional power spectrum- or energy-based detection methods. We are investigating fundamental statistical and spectral signal processing approaches to relieve the requirement.

Staff

Selected publications

  • T. W. Chiang, J. M. Lin, and H. P. Ma, Optimal detector for multitaper spectrum estimator in cognitive radios, in Proc. IEEE Global Telecommunications Conference (IEEE GLOBECOM), Nov. 30-Dec. 4. 2009

Secure Spectrum Sensing in Cognitive Radio Networks

Secure.png

Reliable and swift spectrum sensing is a crucial technical challenge of cognitive radio (CR) that must be overcome before the widespread deployment of CR networks. In spectrum sensing, the incoming signal is received and processed, and then a decision on whether the sensed band can be accessed is made. If cooperative sensing is considered, the sensing results (either raw observation or hard decision) from each CR are reported and a global decision is made. Therefore sensing functionality exposes two kinds of secure vulnerabilities at the physical layer, which are categorized as sensing link attack and sensing cooperation attack. (1) Sensing link attack: In this attack, malicious attacker launches electromagnetic signals in the spectrum bands which CR is observing to affect the system sensing performance, e.g. primary user emulation attack (2) Sensing cooperation attack: When cooperation in spectrum sensing is involved, the malicious attacker could control or emulate as a CR and send false information and mislead the spectrum sensing results to cause collision or inefficient spectrum usage. The goal of this research is to investigate how to improve the security of spectrum sensing in physical layer.

We analyze the impact caused by different types of misbehaviors on sensing performance and develop a misbehaved CR detection algorithm by introducing a reputation-based mechanism.

Staff

Selected publications

  • K. Zeng, P. Pawełczak, D. Čabrić, Reputation-based Cooperative Spectrum Sensing with Trusted Node Assistance, accepted to IEEE Communications Letters, 7 Dec. 2009

VANET Design and Simulation for ITS

VANET proj.png

Safer, more efficient, and more comfortable automotive transport are the primary goals of Intelligent Transportation Systems (ITS). Dedicated Short Range Communications (DSRC) wireless links between vehicles, transit authorities, and consumer service providers will provide the communications infrastructure to realize these goals. The IEEE 802.11p WAVE, Wireless Access for the Vehicular Environment, standard is currently being developed to provide high rate, mid-range, mobile wireless connectivity.

Research of these emerging vehicular ad-hoc networks (VANET) has become increasingly pertinent. Safety applications are of particularly high priority for ITS. Consequently, the accuracy and applicability of VANET performance analysis to real-world scenarios is of utmost importance. Unfortunately, the shear complexity of VANETs often precludes direct analysis and encumbers efficient simulation.

Application, network, medium access, and physical layers interact with each other and channel and mobility models to realize a VANET simulation. Our research investigates the cross-layer interactions between various layers and the environment. We identify parameters critical to performance and develop simulation models to account for these parameters. Algorithms are designed cognizant of cross-layer affects and top layer performance constraints.

Staff

Selected publications

Caltrans

Estimation and Characterization of Frequency Hopping Interferers Using Spectral Sensing Techniques

Freq blockdiag.jpg

Identifying and distinguishing among the individual frequency-hopping spread spectrum (FHSS) emitters in a band of interest is one of the most important aspects of spectrum sensing in cognitive radio. Characterization of FHSS interferers through spectrum sensing enables intelligent interference management and more secure communication in a cognitive radio network. One important attribute that can be measured to enable this characterization of the transmission environment is carrier frequency offset, inherent in all transmitters, such as Bluetooth. FHSS signals are known to be difficult to intercept, so the prior work was focused on methods for FHSS signal detection and interception. However, the methods for interception do not address two important cognitive radio spectrum sensing goals: determining the exact number of FHSS transmitters and their unique frequency offset fingerprints.

In practice, obtaining sufficiently accurate carrier frequency offset measurements can be challenging given constraints on sampling rate and DFT size. Individual emitters will often exhibit offsets separated by only a few KHz, which can make it difficult to measure carrier frequency offset sufficiently accurately to perform emitter differentiation based on a single transmitted pulse from each emitter. The frequency resolution can be increased if data from multiple pulses from the same emitter are combined. This is in some sense a classic estimation problem, though one which in the spectrum sensing context is made substantially more complex due to time and frequency variations of the transmissions made by the multiple emitters.

In our work, we utilize an approach in which temporal information is used to resolve this potential ambiguity, and thus to enhance the resolution of the frequency analysis, thereby increasing the ability to accurately characterize frequency offset. The method leverages the facts that 1) the start times of transmissions are uniformly distributed, and 2) the frequency hopping interval for standards such as Bluetooth is several orders of magnitude larger than the time resolution available in low cost detectors. Thus, observations of pulse start times can be utilized to form a list of likely different emitters, and the averaging over multiple pulses can then be performed. This in turn enables accurate frequency offset information which can then be used to identify future transmissions of devices, even when such transmissions occur at different time offsets.

Staff

Selected publications

  • A. Gök, S. Joshi, J. Villasenor, D. Čabrić, Determining and Characterizing The Number of Frequency Hopping Interferers Using Time and Frequency Offset Estimation, in Proc. IEEE Military Communications Conference (IEEE MILCOM), 18-21 Oct. 2009

Cisco

Weighted Centroid Localization Algorithm: Theoretical Analysis and Distributed Implementation

Localization.jpg

Information about primary transmitter location is crucial in enabling several key capabilities in dynamic spectrum access networks, including improved spatio-temporal sensing, intelligent location-aware routing, as well as aiding spectrum policy enforcement. Compared to other proposed non-interactive localization algorithms, the weighted centroid localization (WCL) scheme uses only the received signal strength information, which makes it simple to implement and robust to variations in the propagation environment. In this paper we present the first theoretical framework for WCL performance analysis in terms of its localization error distribution parameterized by node density, node placement, shadowing variance and correlation distance. Using this analysis, we quantify the robustness of WCL to various physical conditions and provide design guidelines, such as node placement, for the practical deployment of WCL. We also propose a practical method for employing WCL through a distributed cluster-based implementation. This approach achieves comparable accuracy with its centralized counterpart, and greatly reduces total transmit power.

Staff

Selected publications

Retrieved from "http://cores.ee.ucla.edu/index.php?title=Research&oldid=2886"