In this article, we design a memristive competitive neural network circuit based on the winner-take-all (WTA) mechanism and the online Hebbian learning rule. Each synapse of the network contains two memristors whose terminals of signal inputs are opposite. However, only one memristor participates in the calculation each time, and that one is determined by the original input signal. The competitive neural network circuit includes two parts: forward calculation and weight update. In this article, the forward calculation part of the circuit is designed based on the WTA mechanism. The combination of the leaky-integrate-and-fire (LIF) model and pMOS realizes the lateral inhibition of neurons. The design of the weight updating part is based on Hebbian learning rules. In each cycle, only synaptic memristors connected to the winner output neuron in forward calculation can be adjusted. The voltage used for synaptic memristor adjustment comes from the membrane voltage of the winner output neuron. The whole neural network circuit does not need the participation of a central processing unit (CPU) or a field-programmable gate array (FPGA) and really realizes parallel calculation, the saving of area, power consumption, and a certain extent computing-in-memory. Based on the circuit designed in PSPICE, we simulated the classification of $5\times3$ pixel pictures. The changing trend of weights in the training phase and the high recognition accuracy in the recognition phase prove that the network can learn and recognize different patterns. The competitive neural network can be applied to the neuromorphic system of visual pattern recognition.
Two-dimensional materials have potential applications for flexible thermoelectric materials because of their excellent mechanical and unique electronic transport properties. Here we present a functionalization method by a Lewis acid-base reaction to modulate atomic structure and electronic properties at surface of the MoS2nanosheets. By AlCl3solution doping, the lone pair electronics from S atoms would enter into the empty orbitals of Al3+ions, which made the Fermi level of the 1T phase MoS2move towards valence band, achieving a 1.8-fold enhancement of the thermoelectric power factor. Meanwhile, benefiting from the chemical welding effect of Al3+ions, the mechanical flexibility of the nanosheets restacking has been improved. We fabricate a wearable thermoelectric wristband based on this improved MoS2nanosheets and achieved 5 mV voltage output when contacting with human body. We think this method makes most of the transition metal chalcogenides have great potential to harvest human body heat for supplying wearable electronic devices due to their similar molecular structure.
We investigate the propagation of a partially coherent beam in optical lattice. The condition for formation of partially coherent solitons is obtained using mutually coherent function, and the influence of partial coherence on the evolution of light beam is identified.
In this paper, we study a midinfrared femtosecond optical parametric amplifier (OPA) that is severely affected by group velocity dispersion (GVD). Both theoretical and experimental results show that GVDs in nonlinear crystals will significantly degrade the performance of a femtosecond OPA. By introducing a prechirp to the pump pulse, the effect of GVD can be effectively compensated. A lithium-niobate-crystal-based femtosecond OPA demonstrates that the conversion efficiency with optimally prechirped pumping is nearly twice that of the nonchirp case, and the output pulses can be further compressed to nearly their Fourier-transform limit by prism pairs.
More durable (with high impact force), lighter, and more compact flexible azo dye micropolarizers are attractive candidates for low-cost, simple polarization imaging systems. The liquid crystal polymer (LCP), as an emerging material developed by photo-alignment technology, is a potential material for organizing the long-range ordered structure of azo dyes. However, little research has been done on LCP aligned azo dyes. This paper points out and solves a key problem that restricts the fabrication of high-precision arrays in guest (azo dye)-host (LCP) systems: the doping of dyes leads to disorder of the LCP during curing. After solving the problem, the relationship between the thickness of the LCP and the extinction ratio of the polarizing film was investigated, which effectively improved the extinction ratio. Alignment of azo dye molecules in the range of 2 µm (0°–180°) and arrays of micropolarizers (0°, 45°, 90°, -45°) with 8 µm × 8 µm pixel pitch was achieved by laser direct writing technology. The bending cycle test demonstrates the mechanical stability of the ultrathin flexible polarizer. The flexible patterned polarizer with robust chemical and mechanical stabilities provides a flexible way to capture the polarization of the light and highly integrated advanced flexible optoelectronic devices.
We have theoretically and experimentally investigated polarization insensitive all-optical wavelength conversion for polarization multiplexing signal based on orthogonal pump four-wave mixing in nonlinear optical fiber. After wavelength conversion based on 1 km high-nonlinear optical fiber with polarization insensitivity, the power penalties of 2.5 Gbit/s optical OOK intensity and 10 Gbit/s differential phase-shift keying orthogonal signals are less than 0.5 and 0.8 dB, respectively.
Carrier frequency offsets and mobility induced Doppler shifts introduce time-selective in wireless links. To mitigate the resulting time-selective, in this paper, optimal training sequence was designed for linear minimum mean square error (LMMSE) estimation of multi-input multi-output time-selective channels by maximizing a lower bound on the average capacity while at the same time minimizing the mean-square channel estimation error. Optimal number and placement of pilot symbols are also analyzed. Numerical analysis and simulation corroborate our theoretical designs.
At first, the grid and grid resource are introduced simply;Then, analyzes three parts of grid resource management; Finally, a tactic to grid resource distribute are given in detail.
Abstract The production of broadband, terawatt terahertz (THz) pulses has been demonstrated by irradiating relativistic lasers on solid targets. However, the generation of extremely powerful, narrow-band and frequency-tunable THz pulses remains a challenge. Here, we present a novel approach for such THz pulses, in which a plasma wiggler is elaborated by a table-top laser and a near-critical density plasma. In such a wiggler, the laser-accelerated electrons emit THz radiations with a period closely related to the plasma thickness. The theoretical model and numerical simulations predict that a THz pulse with a laser–THz energy conversion of over 2.0%, an ultra-strong field exceeding 80 GV/m, a divergence angle of approximately 20° and a center frequency tunable from 4.4 to 1.5 THz can be generated from a laser of 430 mJ. Furthermore, we demonstrate that this method can work across a wide range of laser and plasma parameters, offering potential for future applications with extremely powerful THz pulses.
The deficiencies of the existed centralized network supervisory system are first point out, then software agent technology to distribute intelligence and control of the network supervisory system inside the entire network in order to enable the entire network management system with features of adjustability and autonomy are applied. Compared with traditional network supervisory system, it not only enhances the monitoring efficiency but also adapts the distributional dynamic network environment. System with this structure is easier to expand, and enhanced the reuse of the management service.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)