Chamith Wijenayake

Multidimensional (MD) Circuits and Systems for Emerging Applications Including Cognitive Radio, Radio Astronomy, Robot Vision and Imaging

Advances in the performance of VLsi circuits are leading to a number of emerging applications of multidimensional (md) filters. Early progress was focused on the numerical design of two dimensional (2-d) transfer functions and the challenging stability issues associated with low-complexity infinite impulse response (iir) implementations. However, over the last decade or so, important practical advances have occurred in the design of 3-d and 4-d iir filters, leading to some important emerging applications. in this tutorial article, some of these applications are described, with emphasis on 2-d spatio-temporal beamforming and 4-d light field processing. in particular, advances in spatio-temporal beamforming for cognitive radio systems and for synthetic aperture radio telescopes are considered. in the 4-d case, we describe a class of 4-d light field filters for image processing, 4-d hyper-fan filters for low-light imaging, depth filtering, denoising and the attenuation of distracting objects, with applications in computational photography and habitat monitoring. Both analog and digital systolic VLsi circuit implementations are described with emphasis on recent progress using field programmable gate array (fPgA)-based and digital VLsi circuits that can potentially operate at radio frequencies in the multi-gHz range. these new innovations open up exciting possibilities for real-time md filters having frames rates in the multi-gHz for emerging radio frequency (rf) antenna signal processing and imaging systems.

RF Analog Beamforming Fan Filters Using CMOS All-Pass Time Delay Approximations

A continuous-time (CT) radio frequency (RF) antenna array beamformer and analog circuit based on a discrete-space-continuous-time (DSCT) 2-D fan-filter having transfer function H_F,A(z_x,s_ct) is derived. The proposed transfer function is based on a 2-D FIR discrete domain fan filter. The discrete domain prototype is converted to the proposed mixed-domain DSCT analog filter by replacing unit sampled delays with CT analog first-order all-pass networks corresponding to the bilinear transform. First-order all-pass network Φ(s_t) is a poor approximation to a CT delay exp(-<i>sT</i>) . To address this, a novel broadband pre-warping method is proposed to exactly compensate for such “bilinear warping”. A 65 nm CMOS VLSI circuit for Φ(<i>s</i>_t) is proposed and an example fan filter with axis oriented at θ₀=35^°, half-fan-angle ε = 5^° and maximum frequency <i>F</i>_u = 2.6 GHz is simulated employing closed-form expressions, an ABCD parameter based model and 65 nm CMOS simulations in Cadence. A stop-band interference rejection of 38 dB is verified by BSIM4 based simulations. The proposed circuit for Φ(<i>s</i>_t) operates at 3.7 mA from a 1.2 V supply. The beamfomer is shown to operate correctly in the presence of PVT variations of Φ(<i>s</i>_t).

All-Pass Filter-Based 2-D IIR Filter-Enhanced Beamformers for AESA Receivers

An active electronically scanned array (AESA) beamforming method that provides enhanced selectivity (interference rejection) for the same number of antennas compared to conventional delay-and-sum (DAS) beamforming is proposed. Conventional DAS 2-D transfer function is modified by introducing complex pole-manifolds based on recently proposed 2-D infinite impulse response (IIR) beam filters, at guaranteed stability. A continuous-time domain signal flow graph is proposed based on first order all-pass filters that eliminate the need of transmission line-based delays used in conventional DAS beamformers. Improved interference rejection is verified using closed-form signal processing models. For an array of 64 antennas, with desired signal direction of arrival (DOA) 10 ° and interference DOA -60° from array broadside, the proposed scheme shows an improvement in the signal-to-interference ratio (SIR) around 7 dB for the same number of antennas, compared to DAS beamforming. The improvement in interference rejection is observed for both uniform and non-uniform aperture weights in terms of side lobe performance. A feasibility study is presented on potential CMOS circuit implementation of the proposed AESA for a linear array of eight antennas and maximum operational frequency of 1 GHz.

Continuous-Time Analog Two-Dimensional IIR Beam Filters

Analog filter circuits that are realized in continuous-time discrete-space domain are proposed for implementing 2-D infinite impulse response (IIR) spatiotemporal transfer functions having beam-shaped passbands. These filters may be used for analog beamforming applications including seismic, audio, sonar, and ultrasonic signal processing. Unlike digital implementations, the proposed 2-D analog beam filters avoid the need for A/D conversions and digital circuitry. Importantly, the beam distortion (beam squint) due to bilinear transformation that is unavoidable in digital 2-D IIR beam filters is significantly reduced. A demonstrative example using operational amplifiers for first-order analog filters is provided.

2-D-IIR Time-Delay-Sum Linear Aperture Arrays

A uniform linear array (ULA) digital beamformer with enhanced selectivity (interference rejection) over conventional true-time delay-sum (TTDS) beamforming is proposed for scanned aperture applications. Conventional TTDS transfer function, which contains only zero-manifolds, is modified by introducing complex pole-manifolds based on 2-D infinite impulse response (IIR) digital beam filters, thereby achieving enhanced selectivity for the same number of antennas in the ULA. For a ULA of 64 antennas, desired direction of arrival (DOA) 30° and interference DOA -60° from array broadside, a relative improvement in signal-to-interference ratio (SIR) of 7 dB is verified. For 128-element ULA, interference DOA -10°, a relative improvement in SIR of 5-10 dB is obtained for the desired DOA in the range 10°-90°. The projected improvements are independent of the weights of the conventional phased-array, TTDS, or filter-sum beamformer used because the method does not change the zero-manifolds of the original array pattern.

Space-Time Spectral White Spaces in Cognitive Radio: Theory, Algorithms, and Circuits

Space-time spectral white spaces in a cognitive radio environment are defined based on multidimensional spatio-temporal spectral properties of radio waves received by a planar array of antennas. Spectral occupancy of a given carrier frequency pertaining to a particular direction in space is expressed by the volume of a semi-cone shaped geometrical region in the 3-D spatio-temporal frequency space ω. A combined approach employing low complexity array processing and conventional time-frequency spectrum sensing is proposed towards the detection of space-time white spaces in ω. The detection scheme employs four subsystems; antenna array, front-end processing, 3-D spatio-temporal array processing, and 1-D spectrum sensing. Key components in the antenna array and front-end processing subsystems are described including an example of a broadband Vivaldi antenna simulated in the frequency range 1.25-2 GHz. The array processing subsystem employs 3-D infinite impulse response digital beam filters, as a low complexity alternative to conventional phased arrays. One potential realization of the 1-D spectrum sensing subsystem is described by using a tunable bandpass filter followed by an energy detector. Simulation examples are provided by considering different directions of arrival, effect of multi-path replicas, signal to noise ratio changes and both narrow band and wideband signals in the normalized temporal frequency range (0,π).

Analog 2D fan filters from discrete domain transfer functions

A continuous-time discrete-space 2D IIR fan filter is proposed for ultra-wideband beamforming for uniform linear arrays of antennas. The proposed signal-flow-graph replaces sampled unit-delays in a 2D FIR digital prototype fan filter using time-delays of the same duration using an ideal delay line. Thereafter, ideal delays are approximated using an RC-active VLSI circuit consisting of high-order all-pass filters aimed at the 90 nm CMOS technology. The paper covers transferfunction design, 2D input spectra in continuous-time, delay approximation, CMOS all-pass circuits and simulation. Finally, a prototype with fan axis directed along θ = 30° (from temporal frequency axis) and half fan angle ε = 6° is simulated and the frequency response from DC to 3.6 GHz is obtained using data from a BSIM4 model of a 10th- order all-pass filter. The proposed novel analog 2D fan filters are free of aliasing, quantization noise, switching power consumption, and do not require an array of high-speed analog-to-digital converters.

Characterizing and classifying IoT traffic in smart cities and campuses

Campuses and cities of the near future will be equipped with vast numbers of IoT devices. Operators of such environments may not even be fully aware of their IoT assets, let alone whether each IoT device is functioning properly safe from cyber-attacks. This paper proposes the use of network traffic analytics to characterize IoT devices, including their typical behaviour mode. We first collect and synthesize traffic traces from a smart-campus environment instrumented with a diversity of IoT devices including cameras, lights, appliances, and health-monitors; our traces, collected over a period of 3 weeks, are released as open data to the public. We then analyze the traffic traces to characterize statistical attributes such as data rates and burstiness, activity cycles, and signalling patterns, for over 20 IoT devices deployed in our environment. Finally, using these attributes, we develop a classification method that can not only distinguish IoT from non-IoT traffic, but also identify specific IoT devices with over 95% accuracy. Our study empowers operators of smart cities and campuses to discover and monitor their IoT assets based on their network behaviour.

A Steerable DC-1 GHz all-pass filter-Sum RF space-time 2-D beam filter in 65 nm CMOS

An electronically steerable broadband radio frequency (RF) filter-sum beamforming filter using 1st-order all-pass filters is proposed. The beamforming filter has the 2-D transfer function Ha (zx, sct), which uses M-section cascaded 1st-order all-pass filters and analog combiners as building blocks. Beam steering is achieved by tuning the group delay of the all-pass filters via a control voltage. For beam directions ψ ≤ 20o from array broadside, M = 1 provides close to ideal broadband response, and 1 <; M ≤ 4 is shown to provide nearly ideal broadband response for beam directions ψ > 20°. The array factor of the beamformer is evaluated using 65 nm CMOS BSIM 4 simulations of the all-pass filters and a 4-channel RF combiner for a linear antenna array of 4 antennas and is shown to provide steerable beams at 1 GHz. The CMOS simulations verify a broadband response from DC to 1 GHz.

Directional spectrum sensing using tunable multi-D space-time discrete filters

The potential use of multi-dimensional (MD) signal processing concepts and 1-D tunable bandpass filtering (BPF) for 2-D space-time (ST) spectrum sensing in a cognitive radio environment is proposed. The 2-D ST spectrum sensing allows the secondary users to search and detect white spaces along different spatial directions at a given time instance. This 2-D ST interrogation of the radio environment enables efficient sharing of primary frequency bands over multiple spatial directions in different instances of time. The use of frequency agile antennas is highlighted to mitigate interferences at known frequencies. Broadband/intermediate-frequency (IF) and analog/digital realizations of 2-D ST filters are proposed to perform the directional enhancement of signals. A low complexity highly-selective tunable 1-D digital BPF is proposed to be used with energy detection to perform spectral sensing along the required direction in space. The tunable BPF eliminates the need of computationally intensive high resolution FFT for frequency scanning. A prototype FPGA implementation of the BPF operating at 66 MHz is used to study the effect of finite precision arithmetic with a word length of 8 bits.