Search

Placement has always been the most time-consuming part of the field programmable gate array (FPGA) compilation flow. Conventional simulated annealing has been unable to keep pace with ever increasing sizes of designs and FPGA chip resources. Without utilizing information of the circuit topology, it relies on large amounts of random swap operations, which are time-costly. This paper proposes an adaptive range-based algorithm to improve the behavior of swap operations and limit the swap distances by introducing the concept of range-limiting strategy for nets. It avoids unnecessary design space exploration, and thus can converge to near-optimal solutions much more quickly. The experimental results are based on the Titan benchmarks, which contain 4K to 30K blocks, including logic array blocks, inputs and outputs, digital signal processors, and random access memories. This approach achieves $2.82\boldsymbol \times $ speed up, 4.8% reduction on wire length, 4.1% improvement on critical path compared with the SA from VTR with wire length-driven optimization, and $1.78\boldsymbol \times $ speed up, 10% reduction on wire length, 2% reduction on critical path with path timing-driven optimization. It also manifests better scalability on larger benchmarks. [ABSTRACT FROM AUTHOR]

This paper describes bi-directional (linear and rotational) propelling and 2-D steering of acoustic bubble-powered microswimmers that are achieved in a centimeter-scale pool (beyond chip level scale). The core structure of a microswimmer is a microtube with one end open in which a gaseous bubble is trapped. The swimmer is propelled by microstreaming flows that are generated when the trapped bubble is oscillated by an external acoustic wave. The bubble oscillation and thus propelling force are highly dependent on the frequency of the acoustic wave and the bubble length. This dependence is experimentally studied by measuring the resonance behaviors of the testing pool and bubble using a laser Doppler vibrometer (LDV) and by evaluating the generated streaming flows. The key idea in the present 2-D steering is to utilize this dependence. Multiple bubbles with different lengths are mounted on a single microswimmer with a variety of arrangements. By controlling the frequency of the acoustic wave, only frequency-matched bubbles can strongly oscillate and generate strong propulsion. By arranging multiple bubbles of different lengths in parallel but with their openings opposite and switching the frequency of the acoustic wave, bi-directionally linear propelling motions are successfully achieved. The propelling forces are calculated by a CFD analysis using the Ansys Fluent® package. For bi-directional rotations, a similar method but with diagonal arrangement of bubbles on a rectangular swimmer is also applied. The rotation can be easily reversed when the frequency of the acoustic wave is switched. For 2-D steering, short bubbles are aligned perpendicular to long bubbles. It is successfully demonstrated that the microswimmer navigates through a T-junction channel under full control with and without carrying a payload. During the navigation, the frequency is the main control input to select and resonate targeted bubbles. All of these operations are achieved by a single piezoelectric actuator. [ABSTRACT FROM AUTHOR]

This paper describes an underwater micropropulsion principle where a gaseous bubble trapped in a suspended microchannel and oscillated by external acoustic excitation generates a propelling force. The propelling swimmer is designed and microfabricated from parylene on the microscale (the equivalent diameter of the cylindrical bubble is around 60 μm) using microphotolithography. The propulsion mechanism is studied and verified by computational fluid dynamics (CFD) simulations as well as experiments. The acoustically excited and thus periodically oscillating bubble generates alternating flows of intake and discharge through an opening of the microchannel. As the Reynolds number of oscillating flow increases, the difference between the intake and discharge flows becomes significant enough to generate a net flow (microstreaming flow) and a propulsion force against the channel. As the size of the device is reduced, however, the Reynolds number is also reduced. To maintain the Reynolds number in a certain range and thus generate a strong propulsion force in the fabricated device, the oscillation amplitude of the bubble is maximized (resonated) and the oscillation frequency is set high (over 10 kHz). Propelling motions by a single bubble as well as an array of bubbles are achieved on the microscale. In addition, the microswimmer demonstrates payload carrying. This propulsion mechanism may be applied to microswimmers that navigate microfluidic environments and possibly narrow passages in human bodies to perform biosensing, drug delivery, imaging, and microsurgery. [ABSTRACT FROM AUTHOR]

This paper reviews and discusses the propulsion of water-floating natural creatures and man-made devices at the air-water interface. Natural creatures with a large ratio of body weight to surface tension force normally propel themselves by repeatedly slapping and stroking their feet while natural creatures with a small ratio of body weight to surface tension force are floated dominantly by surface tension and are propelled by various propulsion mechanisms. Some animals developed unique walking and propelling methods such as leaping, sailing, meniscus climbing, and utilizing of Marangoni effect. Inspired by the natural creatures, a large number of water walking robots and propulsion methods have been developed. Many of them are based on the walking and propelling of water striders, although they require complicated fabrication and structures. Some of the recently developed devices are propelled by the gradient of surface tension (via addition of heat or surfactant) or by streaming flows (induced from oscillations of bubbles or interfaces). These methods need relatively simple mechanical structures and fabrication but are effective only in limited conditions. The bio-inspired water floating and propelling systems can be applied to environment monitoring systems, surveillance security systems, educational robots, and toys. In particular, this paper suggests applying those systems to 'smart dust' sensory robots and pollutant collecting devices. [ABSTRACT FROM AUTHOR]

Given the difficulties in measuring the effective wind speed of wind turbine during the maximum power point tracking (MPPT) process and the unknown nature of the system, this paper proposes a neural network super-twisted sliding mode control (NNST-SMC) method with echo state network (ESN) wind speed estimation. The rotor speed and electromagnetic power are taken as ESN inputs, and the effective wind speed is estimated through the inverse model of wind turbine dynamics. The super-twisted algorithm (STA) can effectively improve the chattering problem of the traditional sliding mode control (SMC) system. The RBF neural network is introduced to compensate for disturbance and uncertainty characteristics of the wind turbine. Results show that compared with the neural network first-order sliding mode control (NNSMC) and super-twisted sliding mode control (ST-SMC), the proposed method can improve the efficiency of wind energy utilization. [ABSTRACT FROM AUTHOR]

A robust approximation-based event-triggered control method is presented for single input single output (SISO) nonlinear continuous-time systems with unmeasurable states and external disturbance. In the whole system, just one neural network (NN) is designed to approximate the unknown part in the controller, and output errors are directly used to construct the system controller and event-triggered mechanism to relieve the burden of system communication. The controller with a simple structure is easier to be realized in practical engineering. The system control signals and adaptive parameters are updated only if the event trigger condition is met, and this way further reduces the waste of network resources caused by frequent system sampling. The application of stability theory of Lyapunov proves that the weight estimation of the NN and tracking errors are ultimate and uniform boundedness, and the efficacy of the proposed scheme is verified with numerical results on a robot. [ABSTRACT FROM AUTHOR]

Industrial defect image generation is a crucial technique for augmenting defect image data, catering to fulfill the exigencies of training data necessary for defect recognition methods. In recent years, generative adversarial networks have seen widespread use in generating defect images. However, constrained by the scarcity of training data and the intricate variability in various defect categories, the existing methods are susceptible to issues such as training instability, a dearth of diversity, and suboptimal image quality. To overcome these challenges, this paper proposes a defect image generation framework based on a progressive training diffusion model (PTDM). Firstly, this study adopts a denoising diffusion model, supplanting traditional generative adversarial networks, to mitigate issues related to training instability and the dearth of diversity observed in generated images. Secondly, a novel progressive training strategy based on the self-designed image quality evaluator is developed to efficiently generate numerous defect images while maintaining quality. Finally, extensive experiments on a steel surface defect image dataset were conducted to validate the performance of the proposed framework in defect image generation and recognition tasks. [ABSTRACT FROM AUTHOR]