Inverse Modulation Research Articles

Image Reconstruction with Multiscale Interest Points Based on a Conditional Generative Adversarial Network

A new image reconstruction (IR) algorithm from multiscale interest points in the discrete wavelet transform (DWT) domain was proposed based on a modified conditional generative adversarial network (CGAN). The proposed IR-DWT-CGAN model generally integrated a DWT module, an interest point extraction module, an inverse DWT module, and a CGAN. First, the image was transformed using the DWT to provide multi-resolution wavelet analysis. Then, the multiscale maxima points were treated as interest points and extracted in the DWT domain. The generator was a U-net structure to reconstruct the original image from a very coarse version of the image obtained from the inverse DWT of the interest points. The discriminator network was a fully convolutional network, which was used to distinguish the restored image from the real one. The experimental results on three public datasets showed that the proposed IR-DWT-CGAN model had an average increase of 2.9% in the mean structural similarity, an average decrease of 39.6% in the relative dimensionless global error in synthesis, and an average decrease of 48% in the root-mean-square error compared with several other state-of-the-art methods. Therefore, the proposed IR-DWT-CGAN model is feasible and effective for image reconstruction with multiscale interest points.

Geotechnical RBDO: Coupling the inverse reliability algorithm with multi-objective reliability-based design optimization of geotechnical systems

The analysis and design of geotechnical engineering inevitably deal with various sources of uncertainties. To further expand the application of reliability methods in geotechnical engineering design works, the inverse reliability algorithm is modified to incorporate unknown geotechnical design parameters in this paper. Based on the established HLRF algorithm for first order reliability analysis, the proposed approach can easily realize traditional single-objective inverse reliability-based design (RBD). Further, the multi-objective inverse reliability design is achieved by combining several single-objective inverse reliability analysis modules through a coupling matrix without seeking for any other surrogate functions. The Nataf transformation is integrated into the method to deal with the correlation of random variables. In addition, the algorithm can be combined with any nonlinear optimization method when the number of design parameters is greater than the number of performance functions, at which point the RBD is extended to reliability-based design optimization (RBDO) with an objective function related to the optimum design purpose. Verifications of the proposed approach are illustrated through several examples, which show that the proposed reliability design optimization framework is effective and can be used as a basic tool for RBDO problems in geotechnical engineering.

Dual-microphone active noise cancellation paved with Doppler assimilation for TADS

An improved dual-microphone adaptive Active Noise Cancellation (ANC) method called IM-RM-LMS is proposed to solve the problem of the bearing sound signal being interfered with by wheel-track contact noise in the TADS. Two microphones, the main microphone and the reference microphone, are used to collect the bearing acoustic signal and the wheel-track contact noise signal, respectively. Due to the different relative spatial positions of the two microphones and the noise source, the noise signal components collected by the two microphones obey different Doppler distortion laws, which makes the LMS-based ANC method ineffective. Therefore, the main research content of this paper is the ANC of two moving sound source signals that obey different Doppler distortion laws. An Inverse Modulation and Re-Modulation (IM-RM) processing is innovatively performed to assimilate the Doppler distortion law of the noise signal components collected by the two microphones. The effectiveness of IM-RM-LMS is verified by simulation and experiment. Compared with the traditional LMS algorithm without considering Doppler assimilation, it can achieve a significantly better denoising effect. This method provides a new theoretical basis for noise cancellation of two moving sound sources obeying different Doppler distortion laws and it is expected to be applied to the noise cancellation of TADS.

Multi-granularity scenarios understanding network for trajectory prediction

Understanding agents’ motion behaviors under complex scenes is crucial for intelligent autonomous moving systems (like delivery robots and self-driving cars). It is challenging duo to the inherent uncertain of future trajectories and the large variation in the scene layout. However, most recent approaches ignored or underutilized the scenario information. In this work, a Multi-Granularity Scenarios Understanding framework, MGSU, is proposed to explore the scene layout from different granularity. MGSU can be divided into three modules: (1) A coarse-grained fusion module uses the cross-attention to fuse the observed trajectory with the semantic information of the scene. (2) The inverse reinforcement learning module generates optimal path strategy through grid-based policy sampling and outputs multiple scene paths. (3) The fine-grained fusion module integrates the observed trajectory with the scene paths to generate multiple future trajectories. To fully explore the scene information and improve the efficiency, we present a novel scene-fusion Transformer, whose encoder is used to extract scene features and the decoder is used to fuse scene and trajectory features to generate future trajectories. Compared with the current state-of-the-art methods, our method decreases the ADE errors by 4.3% and 3.3% by gradually integrating different granularity of scene information on SDD and NuScenes, respectively. The visualized trajectories demonstrate that our method can accurately predict future trajectories after fusing scene information.

Signal Inversion and Amplification of Circularly Polarized Luminescence in a Poly(phenylacetylene)-Based Composite System Assisted by Achiral PMMA.

Multichannel regulable circularly polarized luminescence (CPL) is fascinating because of its fundamental and application interest. There are few reports on helical sense (P-/M-helix) modulation and chiral signal amplification of polyacetylenes with the assistance of achiral polymers and further applications in precisely and conveniently regulating CPL handedness and magnitude. Herein, a helical poly(phenylacetylene)-based CPL-active system was constructed, in which CPL inversion occurred by adding achiral PMMA into a helical poly(phenylacetylene)-CHCl3 solution. Significantly, there is almost 10 times magnification of luminescence dissymmetry factor values (|glum|) during this process. The above phenomena could be ascribed to the PMMA-assisted polyene backbone elongation and the formation of a more ordered helical structure for the poly(phenylacetylene)s. More interestingly, the CPL signal can be facilely inverted and switched by simply changing the thickness of the PPhAD/PMMA layer. The temperature-driven dynamic CPL handedness inversion and magnitude modulation can also be achieved. Based on the multiple regulations for CPL, logic operations were developed, and the practical application is further facilitated by designing various CPL patterns. This study establishes effective and convenient strategies to switch the handedness, magnitude, and wavelength of CPL, which may generate a breakthrough in the manufacturing of CPL-active smart materials and devices with promising application potential.

Time-reversal symmetries in two-dimensional reversible partitioned cellular automata and their applications

Time-reversal symmetry (T-symmetry) in a reversible cellular automaton (CA) is the property in which forward and backward evolutions of configurations are governed by the same local transition function. We show that the framework of partitioned cellular automata (PCAs) is useful to study T-symmetries of reversible CAs. Here, we investigate reversible elementary square PCAs (ESPCAs) and reversible elementary triangular PCAs (ETPCAs), and prove that a large number of reversible ESPCAs and all reversible ETPCAs are T-symmetric under some kinds of simple transformations on configurations. As applications, these results are used to find and analyse backward evolution processes in reversible PCAs. For example, for a given functional module implemented in a reversible PCA, such as a reversible logic element, we can obtain its inverse functional module very easily using its T-symmetry.

90 nm CMOS Implementation of Multiplicative Inverse of the S-Box for AES Algorithm Using Six Transistor XOR Gate

The Substitution-BOX is the most difficult architecture and is at the heart of any Advanced Encryption Standard algorithm implementation. It is the most complicated non-linear architecture using multiplicative inverse. It consumes maximum amount of the energy and power budget of the algorithm. This paper introduces a full-custom CMOS implementation of Multiplicative Inverse module for Substitution/ Inverse Substitution transformation in composite field arithmetic using Galois field GF (28). The multiplicative inversion module utilizes large number of XOR gates in its implementation. This paper introduces implementation of a novel XOR gate using six transistors. Using this six transistor XOR gate, the multiplicative inverse module is implemented in 90 nm CMOS technology. Simulation of the proposed design is achieved using Tanner EDA v.16 software. The area of the multiplicative inverse circuit is 39.92 µm2 requiring 776 transistors. With 0.6 Volts supply voltage, the design shows a power dissipation of 2.386 µWatts making it ideal for applications such as smart cards and RFID tags.

An LMIs‐based back‐stepping sliding mode control framework for robust velocity tracking of the automatic driving vehicle on sloped roads

AbstractAutomatic driving has received a broad of attention from academia and industry since it is effective in greatly reducing the severity of potential traffic accidents and achieving the ultimate automobile safety and comfort. This article presents a back‐stepping sliding mode controller (BSMC) through linear matrix inequalities (LMIs) for the highly automatic driving vehicle on sloped roads. It includes three key modules, namely, an extended Kalman filter (EKF)‐based road slope estimation module, a robust BSMC‐LMIs velocity‐tracking controller based on the input–output feedback linearization, as well as a longitudinal inverse vehicle dynamics module. The nonlinear combined slip tire model with the transient behavior is introduced to calculate the tire forces properly, which would be further proven to offer more accurate road slope estimations even in a fierce acceleration or deceleration situation. The proposed BSMC‐LMIs controller for velocity tracking can handle the lumped uncertainties which include the modeling error, the parameter perturbation, external disturbances, and noises, and guarantee the reachability of the sliding surface, meanwhile, alleviating the chattering phenomenon inherited from the sliding mode structure. Besides, a sufficient condition for the existence of the proposed BSMC is derived by using the LMIs, which ensures the asymptotical stability on the sliding surface. Finally, the robustness, feasibility, and effectiveness of the proposed BSMC‐LMIs controller for velocity‐tracking are verified by simulation tests in various working scenarios, which shows satisfying results when dealing with the lumped uncertainties on sloped roads. Moreover, the comparative study also shows that the proposed BSMC‐LMIs controller has the best tracking performance when compared to the model predictive control, conventional sliding mode control, and the cubic proportional‐integral controller.

Method for increasing the degree of protection of message encryption based on an algorithm for a constant component in time

Currently, asymmetric cryptosystems are used everywhere, in document management for cryptocurrencies, providing a high level of protection to end users, relying on the mathematical complexity of calculating a discrete algorithm. But, it is possible to make a cryptocurrency attack on the so-called ephemeral key, which is an auxiliary key when creating a signature. Recent works have shown examples of cryptocurrencies on the random number generator, processor cache, timing attacks. However, these attacks do not work when the numerical value of the bits is unknown. Also, recent work shows the main vulnerability in the case signature, namely the inverse module calculation algorithm that is vulnerable to timing attacks. The article considers the damage of cryptosystems such as DSA and ECDSA before the attack based on the analysis of the variable time of signing the message. A mathematical model has been developed to test this type of lesion, based on lattice attacks. It is shown that if there are enough signatures with the same signing time, it is possible to identify the presence of common bits of ephemeral keys, which will restore the sender's private key. It is proved that the cause of the lesion is the lack of execution of the operation of calculating the inverse module of the time variable, which provides ephemeral key data to the attacker. To solve this problem, an extended Euclidean algorithm for calculating the inverse module for a fixed time is proposed. In this paper, the advanced Euclidean algorithm for calculating the inverse module is improved, namely, its constant time execution is achieved, which prevents timed attacks.

Image Compression Based on Hybrid Domain Attention and Postprocessing Enhancement.

Deep learning-based image compression methods have made significant achievements recently, of which the two key components are the entropy model for latent representations and the encoder-decoder network. Both the inaccurate estimation of the entropy estimation model and the existence of information redundancy in latent representations lead to a reduction in the compression efficiency. To address these issues, the study suggests an image compression method based on a hybrid domain attention mechanism and postprocessing improvement. This study embeds hybrid domain attention modules as nonlinear transformers in both the main encoder-decoder network and the hyperprior network, aiming at constructing more compact latent features and hyperpriors and then model the latent features as parametric Gaussian-scale mixture models to obtain more precise entropy estimation. In addition, we propose a solution to the errors introduced by quantization in image compression by adding an inverse quantization module. On the decoding side, we also provide a postprocessing enhancement module to further increase image compression performance. The experimental results show that the peak signal-to-noise rate (PSNR) and multiscale structural similarity (MS-SSIM) of the proposed method are higher than those of traditional compression methods and advanced neural network-based methods.

An enhanced formalism for the inverse reactor kinetics problem

Abstract The inverse kinetics problem in reactor physics is a standard formalism to unfold reactivity on the basis of registered power (flux) profile. The classical inverse point kinetics framework has been retrofitted herein to comprise thermal reactivity feedback effects. The instantaneous fuel and coolant temperatures are thus computed by way of the exponential time-differencing scheme and the corresponding thermal reactivity feedback is plugged into the inverse kinetics module. The core external reactivity is therefore unfolded employing only two consecutive time-steps of the power (flux) profile. A history independent yet straightforward numerical routine is accrued enjoying noticeable robustness with regards to the time-step resolution.

IRLSOT: Inverse reinforcement learning for scene‐oriented trajectory prediction

Forecasting pedestrians' future trajectory in unknown complex environments is essential to autonomous navigation in real-world applications, for example, for self-driving cars and collision warnings. However, modern observed trajectory-based prediction methods may easily over-fit to complex or rare scenes because they do not entirely understand the correlations between scenes and trajectories. To address the over-fitting problem, an Inverse Reinforcement Learning for Scene-oriented Trajectory Prediction (IRLSOT) is proposed in this work. The authors' method can be divided into three modules. First, the inverse reinforcement learning module generates the optimal policy by extracting features from scenes and pedestrians' observed trajectories. A lightweight ENet is used to extract features from scenes. Afterwards, the path sampling module introduces a Gumbel Softmax Trick (GST) to improve the accuracy of optimal policy sampling. Different paths are generated on the basis of the optimal policies. Finally, the information fusion module uses the proposed Scene Based Attention (SBA) to fuse the path and trajectory information, then outputs the predicted trajectories. Comparison results show that IRLSOT improves performance on Stanford Drone Database(SDD) by 5.9 % $\%$ . Furthermore, the authors' test IRLSOT on multi-agent scenarios and the authors' own data sets, and results demonstrate that IRLSOT can enhance the generalization of trajectory prediction to rare or new scenes.

Real-time multimode dynamics of terahertz quantum cascade lasers via intracavity self-detection: observation of self mode-locked population pulsations.

Mode-locking operation and multimode instabilities in Terahertz (THz) quantum cascade lasers (QCLs) have been intensively investigated during the last decade. These studies have unveiled a rich phenomenology, owing to the unique properties of these lasers, in particular their ultrafast gain medium. Thanks to this, in QCLs a modulation of the intracavity field intensity gives rise to a strong modulation of the population inversion, directly affecting the laser current. In this work we show that this property can be used to study in real-time the dynamics of multimode THz QCLs, using a self-detection technique combined with a 60GHz real-time oscilloscope. To demonstrate the potential of this technique we investigate a 4.2THz QCL operating in free-running, and observe a self-starting periodic modulation of the laser current, producing trains of regularly spaced, ∼100ps-long pulses. Depending on the drive current we find two distinct regimes of oscillation with dramatically different properties: a first regime at the fundamental cavity repetition rate, characterised by large amplitude and phase noise, with coherence times of a few tens of periods; a much more regular second-harmonic-comb regime, with typical coherence times of ∼105 oscillation periods. We interpret these measurements using a set of effective semiconductor Maxwell-Bloch equations that qualitatively reproduce the fundamental features of the laser dynamics, indicating that the observed carrier-density and optical pulses are in antiphase, and appear as a rather shallow modulation on top of a continuous wave background. Thanks to its simple implementation and versatility, the demonstrated broadband self-detection technique is a powerful tool for the study of ultrafast dynamics in THz QCLs.

A versatile inversion approach for space/temperature/time-related thermal conductivity via deep learning

Identifying the thermophysical properties of unknown material through the measurement of temperature is of great significance in computational heat transfer. Existing numerical algorithms are generally computationally cumbersome and resource demanding. Rapid advances in deep learning (DL) offer an alternative pathway to speed up the inversion process by fully utilizing the parallel computing ability of Graphics Processing Units (GPUs). In this paper, a DL framework is proposed to reconstruct the thermal conductivity related to space, temperature or time. The whole framework consists of a forward data generation module, a denoising module and an inversion module. It is noteworthy that the physics informed neural network (PINN) is employed in the process of generating training data, which avoids the use of commercial software based on traditional methods. In order to simulate the measurement error in practical scenarios, a certain intensity of Gaussian noise is added to the generated data. After denoising by the U-net, the measured temperature is fed to the nonlinear mapping module (NMM) for the inversion of the unknown thermal conductivity. As a result, a well-trained framework can realize high precision real-time inversion even with intensive environmental noise, offering great potential for applications pertaining to the reconstruction of thermophysical properties.

2.5-D Deep Learning Inversion of LWD and Deep-Sensing EM Measurements Across Formations With Dipping Faults

Deep learning (DL) inversion of induction logging measurements is used in well geosteering for real-time imaging of the distribution of subsurface electrical conductivity. We develop a DL inversion workflow to solve 2.5-D inverse problems arising in well geosteering. The inversion workflow employs three DL modules: a “look-around” fault detection module and two inversion modules for reconstructing anisotropic resistivity models in the presence or absence of fault planes, respectively. Our DL approach is capable of detecting and quantifying arbitrary dipping fault planes in real time. We compare inversion performance considering only short logging-while-drilling (LWD) measurements versus using both short LWD and deep-sensing measurements. The latter measurements provide enhanced depth-of-investigation while minimizing uncertainty. We also obtain improved results when using multidimensional inversion, especially nearby fault planes. This study verifies the applicability of real-time 2.5-D DL inversion across arbitrary faulted formations for well geosteering.

3D interacting hand pose and shape estimation from a single RGB image

Estimating 3D interacting hand poses and shapes from a single RGB image is challenging as it is difficult to distinguish the left and right-hands in interacting hand pose analysis. This paper proposes a network called GroupPoseNet using a grouping strategy to address this problem. GroupPoseNet extracts the left- and right-hand features respectively and thus avoids the mutual affection between the interacting hands. Empowered by a novel up-sampling block called MF-Block predicting 2D heat-maps in a progressive way by fusing image features, hand pose features, and multi-scale features, GroupPoseNet is effective and robust to severe occlusions. To achieve an effective 3D hand reconstruction, we design a transformer mechanism based inverse kinematics module(termed TikNet) to map 3D joint locations to hand shape and pose parameters of MANO hand model. Comprehensive experiments on the InterHand2.6M dataset show GroupPoseNet outperforms existing methods by a significant margin. Additional experiments also demonstrate it has a good generalization ability in the problems including left-hand, right-hand and interacting hand pose estimation from a single RGB image. We also show the efficiency of TikNet by the quantitative and qualitative results.

Increasing speech intelligibility and naturalness in noise based on concepts of modulation spectrum and modulation transfer function

This study focuses on identifying effective features for controlling speech to increase speech intelligibility under adverse conditions. Previous approaches either cancel noise throughout speech presentation or preprocess speech by controlling its intensity and/or spectra. Among them, a method based on modulation transfer function theory, inverting the environmental effects to anticipate attenuation of speech modulation spectrum, shows excellent potential due to its systematic and explicit derivation of intelligibility enhancement against environmental smears. However, strictly following the inverse modulation transfer function is dangerous and inefficient as important speech features can be damaged, and it costs lots of energy to boost all smeared regions. This study takes a different approach: analyzing the relations of smeared modulation spectra by the environments for intelligibility to extract effective modifying features. First, we conduct listening tests for intelligibility in noise with different types of enhanced speech. Next, we extract acoustic and modulation frequency components in the smeared modulation spectra by noise showing high correlation with intelligibility scores. Finally, we examine the intelligibility benefits of modifying these components by performing listening tests. The results show that these components effectively increase intelligibility by at most 18%, which demonstrates that our concept is valid.

Ultrasensitive ethanol sensor based on nano-Ag&ZIF-8 co-modified SiNWs with enhanced moisture resistance

Silicon nanowires (SiNWs) is considered as one of the most promising candidates for gas sensing application due to its capability of operating at room temperature (20 °C) and compatibility with modern semiconductor process. However, there still remains a challenge that SiNWs is poorly sensitive to volatile organic compounds (VOCs). In this study, we reveal that the co-decoration of zeolite imidazole frameworks-8 (ZIF-8) and Ag nanoparticles can improve the gas-sensing response to ethanol and enhance the humidity resistance at 20 °C. Due to the inversion modulation introduced by nano-Ag modificaiton and the enhanced gas adsorption provided by ZIF-8 modification, the resultant co-decorated sensor of ZIF-8&[email protected] shows obviously enhanced response upon exposure to ethanol with wide concentration range. It meanwhile exhibits a practical capability to sense rarefied ethanol vapor low to 125 ppb at 20 °C and relative humidity of 25 %. Especially, the sensor exhibits a low limit of detection of 5.45 ppb of ethanol vapor. When working at high humidity of 85 %, the developed ZIF-8&[email protected] sensor still remain a significant response. The superior moisture resistance performance is clarified by first-principles caculations to be highly related to the hydrophobicity characteristic of the modified ZIF-8 and the hygroscopic center formed by Ag nanoparticles. A resistance model was proposed to demonstrate the enhanced sensing response of the ZIF-8&[email protected] Further, a two-level moisture resistance mechanism is demonstrated based on the first-principles caculations and the unique microstructure of ZIF-8&[email protected] The beneficial effect of ZIF-8&nano-Ag co-modification demonstrated in present work suggest an effective strategy for developing superior gas sensors capable of sensitive detection of VOCs with strong stability in humditiy ambient.

Modulation instability in one-dimensional bi-periodic optical lattice made of negative index metamaterials waveguide arrays

In this paper, the dispersion relation and modulation instability (MI) have been studied in one-dimensional bi-periodic optical lattice of negative index metamaterials waveguide arrays. It is found that the plateaus will be appeared on the dispersion curve when the coupling coefficient is around 0.25. The plateau will become wider with the increase of the nonlinear and longitudinal modulation depth. The phase shift has little effect on MI the gain spectrum when the values of coupling coefficient, nonlinear coefficient and longitudinal modulation depth coefficient are in the range that is conducive to the generation of plateaus. However, if the values of these coefficients are in the range that is not conducive to the generation of plateaus, the effect of phase shift on the MI gain spectrum will become significant. In addition, it is found that the inverse periodic modulation between waveguides causes the light intensity to change periodically during transmission.

A split-label design for simultaneous measurements of perfusion in distant slices by pulsed arterial spin labeling.

PurposeMultislice arterial spin labeling (ASL) MRI acquisitions are currently challenging in skeletal muscle because of long transit times, translating into low‐perfusion SNR in distal slices when large spatial coverage is required. However, fiber type and oxidative capacity vary along the length of healthy muscles, calling for multislice acquisitions in clinical studies. We propose a new variant of flow alternating inversion recovery (FAIR) that generates sufficient ASL signal to monitor exercise‐induced perfusion changes in muscle in two distant slices.MethodsLabel around and between two 7‐cm distant slices was created by applying the presaturation/postsaturation and selective inversion modules selectively to each slice (split‐label multislice FAIR). Images were acquired using simultaneous multislice EPI. We validated our approach in the brain to take advantage of the high resting‐state perfusion, and applied it in the lower leg muscle during and after exercise, interleaved with a single‐slice FAIR as a reference.ResultsWe show that standard multislice FAIR leads to an underestimation of perfusion, while the proposed split‐label multislice approach shows good agreement with separate single‐slice FAIR acquisitions in brain, as well as in muscle following exercise.ConclusionSplit‐label FAIR allows measuring muscle perfusion in two distant slices simultaneously without losing sensitivity in the distal slice.