Search

Deep neural networks (DNNs) are penetrating into a broad spectrum of applications and replacing manual algorithmic implementations, including the radio frequency communications domain with classical signal processing algorithms. However, the high throughput (gigasamples per second) and low latency requirements of this application domain pose a significant hurdle for adopting computationally demanding DNNs. In this article, we explore highly specialized DNN inference accelerator approaches on field-programmable gate arrays (FPGAs) for RadioML modulation classification. Using an automated end-to-end flow for the generation of the FPGA solution, we can easily explore a spectrum of solutions that optimize for different design targets, including accuracy, power efficiency, resources, throughput, and latency. By leveraging reduced precision arithmetic and customized streaming dataflow, we demonstrate a solution that meets the application requirements and outperforms alternative FPGA efforts by 3.5× in terms of throughput. Against modern embedded graphics processing units (GPUs), we measure $>\mathsf{10}\times$>10× higher throughput and $>\mathsf{100}\times$>100× lower latency under comparable accuracy and power envelopes. [ABSTRACT FROM AUTHOR]

The holding force acting on a levitated object during near-field acoustic levitation has not been statically and directly measured so far. In this study, it was considered to realize such a measurement when a levitated object has a large displacement from the vibration source. In previous studies, under restricted conditions, the holding force has been calculated indirectly by image processing or measured from the balance with gravity by tilting the apparatus. In this article, the force was measured based on the magnetic force (MF) compensation principle. This is the first attempt to measure the force directly and statically under arbitrary conditions. One side of a rectangular rotor, which was placed above the acoustic radiation surface, received the holding force. The other side of the rotor received the MF generated by a solenoid. The holding force was estimated from the electric current flowing in the solenoid when the holding force and the MF were balanced. The holding force acting on the rotor surfaces was affected by the deviation between the rotor and the radiation surface. The holding force increased with increasing vibration amplitude or decreasing air gap between the rotor and the radiation surface. When the vibration sources were opposed, the holding force was affected by their vibration phase difference. The holding force was maximum at the opposite vibration phase and decreased with decreasing vibration phase difference. When the vibration sources were arranged in the transportation direction, the transportation force occurred and increased with increasing vibration amplitude of the destination vibration source. These measurement results agree well with the analysis results. [ABSTRACT FROM AUTHOR]

Modern advanced minimally invasive surgery has been implemented for most of the significant gastrointestinal diseases. However, patients with coagulopathy or unresectable tumors cannot be cured by current treatment methods. Moreover, other existing medical devices for targeted drug release are too large to be applied in gastric endoscope because the diameter of the biopsy channel is smaller than 3 mm. To address it, in this work, we developed a piezoelectric single crystal ultrasonic transducer (the diameter was only 2.2 mm and the mass was 0.076 g) to produce acoustic waves, which could promote the drug release in the designed position of the digestive tract through an endoscope. It exhibited the electromechanical coupling coefficient of 0.36 and the center frequency of 6.9 MHz with the −6-dB bandwidth of 23%. In in vitro sonophoresis experiment, the gastric mucosa permeability to Bovine Serum Albumin increased about 5.6 times when the ultrasonic transducer was activated at 40 Vpp and 60% duty ratio, proving that employment of this transducer could facilitate drug penetration in the gastric mucosa. Meanwhile, the permeability could be adjusted by tuning the duty ratio of the ultrasonic transducer. The corresponding sonophoresis mechanism was related to the acoustic streaming and the thermal effect produced by the transducer. In addition, the measured maximum power density was 128 mW/cm2 and the mechanical index of the ultrasonic transducer was 0.02. The results held a great implication for applications of the transducer for targeted drug release in the gastrointestinal tract. [ABSTRACT FROM AUTHOR]

Penetration of nanoscale therapeutic agents into the extracellular matrix (ECM) of a tumor is a limiting factor for the sufficient delivery of drugs in tumors. Ultrasound (US) in combination with microbubbles causing cavitation is reported to improve delivery of nanoparticles (NPs) and drugs to tumors. Acoustic radiation force (ARF) could also enhance the penetration of NPs in tumor ECM. In this work, a collagen gel was used as a model for tumor ECM to study the effects of ARF on the penetration of NPs as well as the deformation of collagen gels applying different US parameters (varying pressure and duty cycle). The collagen gel was characterized, and the diffusion of water and NPs was measured. The penetration of NPs into the gel was measured by confocal laser scanning microscopy and numerical simulations were performed to determine the ARF and to estimate the penetration distance and extent of deformation. ARF had no effect on the penetration of NPs into the collagen gels for the US parameters and gel used, whereas a substantial deformation was observed. The width of the deformation on the collagen gel surface corresponded to the US beam. Comparing ARF caused by attenuation within the gel and Langevin pressure caused by reflection at the gel-water surface, ARF was the prevalent mechanism for the gel deformation. The experimental and theoretical results were consistent both with respect to the NP penetration and the gel deformation. [ABSTRACT FROM AUTHOR]

Patients with unresectable or nonablatable tumors are difficult to cure, but nanotherapy combining targeted nanoparticles has many severe side effects due to the toxicities of anticancer drugs. We found that acoustic energy can produce a local region with high concentration from a low concentration suspended liquid of nano-SiO2 particles at 2.5 MHz. Our calculated results show that the main reason for aggregation is the synthesized effect of the potential well of acoustic energy and streaming to trap them. In addition, the aggregated region can be manipulated to a targeted position in the vessel phantom by moving the ultrasound transducer external to the body. This noninvasive manipulation of suspended nanoparticles can rapidly increase the local drug concentration, but reduce the total dosage of anticancer drugs, which has the potential to be used for patients with advanced tumors by improving the physiological effects and reducing the side effects. [ABSTRACT FROM AUTHOR]

The ability to drive efficient micromixing on a microfluidic platform is crucial for a wide range of lab-on-a-chip applications. Here, we investigate the ability of acoustic waves generated on different geometric surfaces (concave and convex) to enhance the micromixing efficiency in droplet acoustomicrofluidic systems, and, concomitantly, to reduce the power consumption in these devices for a given performance requirement. Quite counterintuitively, we observe that although the acoustic streaming velocity, which scaled inversely with the droplet size, tended to be generally lower (by approximately 45%) when the flow is generated by transducers with convex surfaces compared to those with concave surfaces, the mixing efficiency is disproportionately higher: compared to pure diffusional mixing in the absence of the acoustic forcing, the mixing efficiency due to the acoustically driven convection increased by up to 25% and 43% on these respective surfaces. As such, the mixing enhancement cannot simply be attributed to an increase in the convective flow arising from the acoustic forcing. Rather, we observe the mixing enhancement to be due to the stronger chaotic advection arising in the transducer with the convex surface due to its diverging acoustic field into the droplet. [ABSTRACT FROM AUTHOR]

A combined top-down/bottom-up microfluidic assembly system is demonstrated in which microspheres undergo directed assembly within microchannels. The approach is adapted from macroscale templated assembly by selective removal (TASR), in which high-frequency acoustic excitations drive selective assembly into shape- and size-matched wells patterned in a planar surface. A PDMS layer patterned with microchannels is bonded to a PDMS layer molded to define assembly wells. A suspension of 10 $\mu \text{m}$ diameter polystyrene microspheres is flowed into the microchannels, and an external, high-frequency ultrasonic transducer circulates the microspheres within the channels and removes them from everywhere except the size-matched wells. Microspheres are successfully assembled in wells inside the microchannels with yields of up to 90%. Smaller channel dimensions reduce assembly yield in part because smaller internal channel volumes reduce the availability of microspheres relative to assembly wells. In addition, reducing channel width below about $500~\mu \text{m}$ is shown to be an independent predictor of low assembly yields; this is attributed to the physics of acoustic streaming flow under the acoustic excitation. Assembly yield is largely independent of the location of the assembly wells within the microchannels, and selectivity is confirmed by low assembly yield of $6~\mu \text{m}$ diameter microspheres into the larger wells. [2020-0072] [ABSTRACT FROM AUTHOR]

We propose a system that controls the spatial distribution of odors in an environment by generating electronically steerable ultrasound-driven narrow air flows. The proposed system is designed not only to remotely present a preset fragrance to a user, but also to provide applications that would be conventionally inconceivable, such as: 1) fetching the odor of a generic object placed at a location remote from the user and guiding it to his or her nostrils, or 2) nullifying the odor of an object near a user by carrying it away before it reaches his or her nostrils (Fig. 1). These are all accomplished with an ultrasound-driven air stream serving as an airborne carrier of fragrant substances. The flow originates from a point in midair located away from the ultrasound source and travels while accelerating and maintaining its narrow cross-sectional area. These properties differentiate the flow from conventional jet- or fan-driven flows and contribute to achieving a midair flow. In our system, we employed a phased array of ultrasound transducers so that the traveling direction of the flow could be electronically and instantaneously controlled. In this paper, we describe the physical principle of odor control, the system construction, and experiments conducted to evaluate remote fragrance presentation and fragrance tracking. [ABSTRACT FROM PUBLISHER]

This paper presents the results of preliminary modeling of the boron particles behavior in a mixing chamber downstream, the inductively coupled plasma or radio frequency (RF) torch, for boron nitride nanotubes (BNNTs) production. The 3-D computational fluid dynamics calculations of heating of boron particles by nitrogen plasma with plume power 15 and 30 kW in the mixing chamber with one and four feeding ports have been performed. The obtained results can be used for further design and optimization of the RF system for BNNTs production and new materials synthesis. [ABSTRACT FROM AUTHOR]

The acoustic manipulation has the merits such as no selectivity to manipulated objects, less heat damage to biological samples (for some principles), simple and compact structure, and light weight. We demonstrate a strategy to ultrasonically rotate a single nanowire in a water film on a substrate surface. Controlled mobile acoustic streaming field generated by the ultrasound around a vibrating fiberglass is used to rotate the nanowire at any position on the substrate surface. The nanowire center or end moves to the location directly under the vibrating fiberglass tip due to the acoustic streaming field, and serves as the rotation center. The driving torque has an order of magnitude of 10^-20 N·m. Stable angular speed of a rotating nanowire can be up to several radians per second, and the temperature rise at the manipulation spot is less than 0.1 °C. The method has potential applications in the assembling of micro/nanostructures, orientation of biological samples, and measurement of dynamics properties of micro/nanoentities. [ABSTRACT FROM PUBLISHER]

In this paper, we have put forward and demonstrated a noncontact manipulation strategy that uses an ultrasonic microbeak to suck, align, trap and transfer a single nanowire in a film of water on the surface of a glass substrate. The sucking and trapping force is generated by symmetric acoustic streaming eddies flowing into the microbeak tip in ultrasonic vibration, from the front of the microbeak and along the direction perpendicular to the microbeak vibration. A nanowire in front of the microbeak can be sucked to and trapped under the microbeak, and aligned perpendicularly to the vibration direction approximately. Sucking range of the microbeak becomes larger as its vibration increases, and the transfer speed of a sucked nanowire depends on its relative position to the microbeak. During the sucking process, a nanowire rotates to the direction perpendicular to the microbeak vibration, and the angular speed depends on the orientation angle of the nanowire. A trapped nanowire, which is not in contact with the microbeak, can be dragged on the substrate surface by moving the microbeak, along an arbitrary 2-D path in the water film. The driving force on a transferred nanowire is estimated to have an order of magnitude of 10 ^-14 N. [ABSTRACT FROM PUBLISHER]

This paper presents a novel framework to uncertain one-class learning and concept summarization learning on uncertain data streams. Our proposed framework consists of two parts. First, we put forward uncertain one-class learning to cope with data of uncertainty. We first propose a local kernel-density-based method to generate a bound score for each instance, which refines the location of the corresponding instance, and then construct an uncertain one-class classifier (UOCC) by incorporating the generated bound score into a one-class SVM-based learning phase. Second, we propose a support vectors (SVs)-based clustering technique to summarize the concept of the user from the history chunks by representing the chunk data using support vectors of the uncertain one-class classifier developed on each chunk, and then extend k-mean clustering method to cluster history chunks into clusters so that we can summarize concept from the history chunks. Our proposed framework explicitly addresses the problem of one-class learning and concept summarization learning on uncertain one-class data streams. Extensive experiments on uncertain data streams demonstrate that our proposed uncertain one-class learning method performs better than others, and our concept summarization method can summarize the evolving interests of the user from the history chunks. [ABSTRACT FROM AUTHOR]

This paper presents a numerical model of high-pressure sodium (HPS) lamps, including acoustic streaming. The model is implemented in COMSOL and is based on the finite-element method using a 2-D axisymmetric geometry. Moreover, a revision concerning experimental acoustic resonance (AR) characterizations of HPS lamps will be also provided. The results presented in this paper will contribute to the understanding of AR phenomena and will help in the electronic ballast design methods for AR avoidance. [ABSTRACT FROM PUBLISHER]

In the megasonic cleaning of wafers, size and motion of cavitating bubbles and fluid flow due to acoustic streaming play a very important role. In this paper, chronoamperometric technique has been used to seek information on acoustic streaming and bubble activity in a 1 MHz sound field. Specifically, current transients during reduction of potassium ferricyanide were recorded. Data collected at 1–6 MHz sampling rate using a 25 \mum platinum electrode show current “peaks” indicative of the approach of oscillating bubbles to the electrode and current “valleys” due to blocking of the electrode by bubbles. Acoustic streaming velocity (\sim1.5~cm/s) and bubble size (maximum radius of \sim1~\mum for oscillating bubbles) have been estimated from local current transients caused by bubble activity near the electrode. [ABSTRACT FROM AUTHOR]

An ultrasonic air pump that uses a traveling wave along a small air gap between a bending vibrator and a reflector is discussed. The authors investigate ultrasonic air pumps that make use of bending vibrators and reflectors and confirm that air can be induced to flow by generating an asymmetric acoustic standing wave along an air gap. In this paper, we proposed a novel ultrasonic air pump in which a traveling wave along an air gap induces acoustic streaming and achieves one-way airflow. Two new reflector configurations, stepped and tapered, were designed and used to generate traveling waves. To predict airflow generation, sound pressure distribution in the air gap was calculated by means of finite element analysis (FEA). As a preliminary step, 2 FEA models were compared: one piezoelectric-structure-acoustic model and one piezoelectric-structure-fluid model, which included the viscosity effect of the fluid. The sound pressure distribution in the air gap, including fluid viscosity, was calculated by the FEA because it is expected to be dominant and thus have a strong effect on the sound pressure field in such a thin fluid layer. Based on the FEA results of the stepped and the tapered reflectors, it was determined that acoustic traveling waves could propagate along the gaps. Experiments were carried out with the designed bending vibrator and the reflectors. The acoustic fields in the air gap were measured via a fiber optic probe, and it was determined that the sound pressure and the phase distribution tendencies corresponded well with the results computed by FEA. Through our experiments, one-way airflow generation, in the same direction of the traveling wave and with the maximum flow velocity of 5.6 cm/s, was achieved. [ABSTRACT FROM AUTHOR]

A fast algorithm for the evaluation of acoustic fields produced by given source distributious is developed with the aim of accelerating iterative boundary element method (BEM) solvers. The algorithm is based on field smoothing by phase and amplitude compensation, which allows for sampling of the fields radiated by finite-size source distributions over coarse nonuniform (spherical) grids (NGs). Subsequently, the fields at the desired target points can be obtained by an interpolation and phase and amplitude restoration. Combining this approach with the divide-and-conquer strategy, the total field is computed via a hierarchical decomposition of the source domain. In this computational scheme, the phase and amplitude compensated fields produced by neighboring subdomains are gradually aggregated through a multilevel process involving interpolation between increasingly dense NGs and the scatterer surface. This multilevel NC algorithm is used to reduce the computational cost of applying the field evaluation operator and its adjoint, as required in each iteration of the conjugate gradient solver based on the BEM-discretized integral representation of scattering problems. Accuracy and computational efficiency of the NC algorithm are demonstrated on representative examples of elongated, quasi-planar, and full 3-D scatterers. [ABSTRACT FROM AUTHOR]

In the context of next-generation wireless systems, it is expected that services targeted at mass content distribution will become widely popular. Multicast services, such as audio/video streaming and mobile television, have the characteristic that the same information has to be transmitted to a group of recipients, which may include user terminals as well as relay stations, in the case of relaying networks. This paper deals with the problem of multigroup multicast beamforming for multiantenna wireless cellular networks, which assumes that multiple multicast groups can share the same resource. The intergroup interference, which appears due to the resource sharing, needs to be suppressed by the beamforming algorithms. In this paper, new linear transmit beamforming techniques are proposed, which aim to provide a reasonable tradeoff between performance and computational complexity. These techniques correspond to extensions of known algorithms, such as zero forcing (ZF), minimum mean square error (MMSE), and signal-to-interference-plus-noise ratio (SINR) balancing, to the multicast case. Furthermore, new "multicast- aware" techniques that take into account the peculiarities of multicast transmission are proposed to improve performance. To provide the necessary mathematical framework for the analysis, a novel general multigroup multicast system model is proposed as well. [ABSTRACT FROM AUTHOR]

This work presents an approach for determining the streaming patterns that are generated by Rayleigh surface aconstic waves in arbitrary 3-D geometries by finite element method (FEM) simulations. An efficient raytracing algorithm is applied on the acoustic subproblem to avoid the unbearable memory demands and computational time of a conventional FEM acoustics simulation in 3-D. The acoustic streaming interaction is modeled by a body force term in the Stokes equation. In comparisons between experiments and simulated flow patterns, we demonstrate the quality of the proposed technique. [ABSTRACT FROM AUTHOR]

A slider for a self-running standing wave-type, ultrasonically levitated, thin linear stage is discussed. The slider can be levitated and moved using acoustic radiation force and acoustic streaming. The slider has a simple configuration and consists of an aluminum vibrating plate and a piezoelectric zirconate titanate (PZT) element. The large asymmetric vibration distribution for the high thrust and levitation performance was obtained by adjusting the configuration determined by finite elemental analysis (FEA). As a preliminary step, the computed results of the sound pressure distribution in the 1-mm air gap by FEA was compared with experimental results obtained using a fiber optic probe. The direction of the total driving force for the acoustic streaming in the small air gap was estimated by the sound pressure distribution calculated by FEA, and it was found that the direction of the acoustic streaming could be altered by controlling the vibration mode of the slider. The flexural standing wave could be generated along the vibrating plate near the frequencies predicted based on the FEA results. The slider could be levitated by the acoustic radiation force radiated from its own vibrating plate at several frequencies. The slider could be moved in the negative and positive directions at 68 kHz and 69 kHz, which correspond to the results computed by FEA, with the asymmetric vibration distribution of the slider's vibrating plate. Larger thrust could be obtained with the smaller levitation distance, and the maximum thrust was 19 mN. [ABSTRACT FROM AUTHOR]

This paper describes an efficient model to describe an autoregressive (AR) signal with slowly-varying amplitude in additive white Gaussian noise (WGN). Even a simple low-order AR model becomes complicated by varying amplitude and additive white noise. However, by approximating the signal amplitude as piecewise-constant, an efficient filtering approach can be applied in order to compute the maximum likelihood (ML) estimate for the entire data record. The model is efficient both in terms of having a compact set of parameters and in the computational sense. Simulation results are provided. The algorithm has applications in signal modeling for underwater acoustic signals, particularly active wideband signals such as explosive sources. [ABSTRACT FROM AUTHOR]

Nonspecific binding of proteins is an ongoing problem that dramatically reduces the sensitivity and selectivity of biosensors. We demonstrate that ultrasonic waves generated by surface acoustic wave (SAW) devices remove nonspecifically bound proteins from the sensing and nonsensing regions of the microarrays. We demonstrate our approach for controllably and nondestructively cleaning the microarray interface. In this work, SAWs were generated using 128 YX lithium niobate, chosen for its high coupling coefficient and efficient power transfer to mechanical motion. These waves propagating along the surface were coupled into specifically bound and nonspecifically bound proteins on a patterned surface of 40 mum feature size. Fluorescence intensity was used to quantify cleaning efficacy of the microarrays. Our results have shown that excess protein layers and aggregates are removed leaving highly uniform films as evidenced by fluorescence intensity profiles. Selected antigen-receptor interactions remained bound during the acoustic cleaning process when subjected to 11.25 mW of power and retained their efficacy for subsequent antigen capture. Results demonstrate near-complete fluorescence signal recovery for both the sensing and nonsensing regions of the microarrays. Of significance is that our approach can be integrated into existing array technologies where sensing and nonsensing regions are extensively fouled. We believe that this technology will be pivotal in the development and advancement of microsensors and their biological applications. [ABSTRACT FROM PUBLISHER]

Acoustic streaming induced by longitudinal vibration at 30 kHz is visualized with particle imaging velocimetry (Ply). To gauge an increase in the velocity of air flow due to acoustic streaming, the velocity of air flow in a gap between the heat source and ultrasonic vibrator is measured using PIV. The ultrasonic wave propagating into air in the gap creates steady-state secondary vortex called acoustic streaming which enhances heat transfer from the heat source to neighboring air. Heat transfer through air in the gap is represented by experimental correlation involving Peclet number and Nusselt number. Theoretical analysis reveals that gaps for maximum heat transfer are the multiple of half wavelength. The acoustic streaming velocity of air in the gap is at maximum when the gap agrees with the multiples of half wavelength of the ultrasonic wave, which are specifically 6 mm and 1.2 mm. Considering the propagation loss of the ultrasonic wave and acoustic streaming velocity of air in the gap, gaps at which maximum heat transfer occurs are experimentally found to be half wavelength and one wavelength. [ABSTRACT FROM AUTHOR]

Audio streaming applications have become very popular in recent years, owing to their low cost and convenience. However, during network congestions, data packets are often delayed or discarded, creating an annoying gap in the streamed media. This letter presents a new approach to audio packet loss concealment designed for MPEG-Audio streaming applications. In a previous work, we introduced a receiver-based concealment algorithm based on applying the gapped-data amplitude and phase estimation (GAPES) interpolation algorithm in the discrete short-time Fourier transform (DSTFT) complex domain and obtained better results compared to past methods. The current approach applies the same algorithm on a different complex domain, formed from combining the modified discrete cosine transform (MDCT) domain as its real part and the modified discrete sine transform (MDST) domain as the imaginary part. The new approach significantly reduces the complexity demands while maintaining similar high-quality results. [ABSTRACT FROM AUTHOR]

This paper describes a novel phenomenon "liquid needle" produced by the self-focusing acoustic transducer (SFAT). The SFAT is composed of a metal/piezo-material/metal plate (with and without a support layer) which has one side air backed while the other side is liquid backed. The SEAT generates and focuses acoustic waves (through wave interference) at the top surface of the liquid (with acoustic intensity up to 15W/cm²). The focused acoustic beam deforms the free liquid surface through the acoustic streaming effect, and produces a needle-shaped liquid column on the free liquid surface. With various SFATs and different types of liquids, we have obtained liquid needles that are 400-12 000 µm tall. The liquid needle diameter at its tip varies from less than 15 µm to more than 200 µm. [ABSTRACT FROM AUTHOR]

In this paper, a model is developed for the stability analysis of an acoustically levitated disk on the basis of analyzing eddy acoustic streaming and acoustic viscous stress. In the model, the effect of the acoustic streaming outside the boundary layer that is on the surface of the levitated disk is properly taken into account. Also, the calculation of sound field and acoustic viscous stress is limited to the range that has a dominant effect on the stability. By this method, we obtain a quite accurate solution of the stability coefficient. For the small horizontal shift of a large levitated disk, the model is verified by the good agreement between the experimental and theoretical results. By means of this model and relevant experiments, some factors that affect the stability of the levitated disk are investigated, and useful guidelines for design and application are obtained. It is found that the range from the edge to the outermost nodal circle of the disk-shaped vibrator has a large effect on the stability of the levitated disk. To stabilize the levitated disk by acoustic viscous force, the distance between the edge and the outermost nodal circle of the vibrator must be larger than a critical value, which is determined by the driving frequency and the sound velocity of the fluid between the levitated disk and the vibrator. When this condition is satisfied, increasing the distance between the edge and the outermost nodal circle leads to a decrease in the stability. It is also found that the property of the fluid between the levitated disk and the vibrator has a large effect on the stability. It is easier to stabilize the levitated disk in steam than in air, but more difficult to do so in carbon dioxide and hydrogen. In addition, theoretical results show that increasing the weight per unit area of the levitated object increases the stability for a given vibrator velocity. The distribution of the acoustic viscous stress and the dependence of the stability... [ABSTRACT FROM AUTHOR]

Focuses on a study which presented the numerical simulation and experimental validation of acoustic streaming micromachined flexural plate wave (FPW) devices. Description of the model; Influence of the backpressure; Thermal transfer caused by acoustic streaming.

Presents information on a study wherein a finite element model of breast cyst fluid behavior during streaming detection implementation was developed and validated. Simulation methods used; Details of the phantom experiments; Streaming generation time experiments; Results and discussion; Conclusion.