Non-uniform Input Research Articles

SparseDC: Depth completion from sparse and non-uniform inputs

We propose SparseDC, a model for Depth Completion from Sparse and non-uniform inputs. Unlike previous methods focusing on completing fixed distributions on benchmark datasets (e.g., NYU with 500 points, KITTI with 64 lines), SparseDC is specifically designed to handle depth maps with poor quality in real usage. Our SparseDC makes two major contributions. First, we design a simple strategy, called SFFM, to improve the robustness under sparse inputs by explicitly filling the unstable depth features with stable image features. Second, we suggest utilizing a two-branch feature embedder in order to forecast the exact local geometry of regions with available depth values and accurate structures in regions with no depth. The key of the embedder is an uncertainty-based feature fusion module called UFFM to balance the local and long-term information extracted by CNNs and ViTs. Numerous experiments conducted both indoors and outdoors show how robust and effective our framework is when facing sparse and non-uniform input depths. And SparseDC outperforms the current state-of-the-art (SOTA) method, CFormer, with varying inputs. We use less parameters (38.2M < 45M) and achieve a +18.6% increase in the REL metric and a +9.9% increase in the RMSE metric on NYU Depth dataset. The pre-trained model and the complete code repository can be accessed at https://github.com/WHU-USI3DV/SparseDC.

Implications of variable synaptic weights for rate and temporal coding of cerebellar outputs.

Purkinje cell (PC) synapses onto cerebellar nuclei (CbN) neurons allow signals from the cerebellar cortex to influence the rest of the brain. PCs are inhibitory neurons that spontaneously fire at high rates, and many PC inputs are thought to converge onto each CbN neuron to suppress its firing. It has been proposed that PCs convey information using a rate code, a synchrony and timing code, or both. The influence of PCs on CbN neuron firing was primarily examined for the combined effects of many PC inputs with comparable strengths, and the influence of individual PC inputs has not been extensively studied. Here, we find that single PC to CbN synapses are highly variable in size, and using dynamic clamp and modeling we reveal that this has important implications for PC-CbN transmission. Individual PC inputs regulate both the rate and timing of CbN firing. Large PC inputs strongly influence CbN firing rates and transiently eliminate CbN firing for several milliseconds. Remarkably, the refractory period of PCs leads to a brief elevation of CbN firing prior to suppression. Thus, individual PC-CbN synapses are suited to concurrently convey rate codes and generate precisely timed responses in CbN neurons. Either synchronous firing or synchronous pauses of PCs promote CbN neuron firing on rapid time scales for nonuniform inputs, but less effectively than for uniform inputs. This is a secondary consequence of variable input sizes elevating the baseline firing rates of CbN neurons by increasing the variability of the inhibitory conductance. These findings may generalize to other brain regions with highly variable inhibitory synapse sizes.

Output feedback consensus for input-delayed nonlinear multi-agent systems via event-triggered communication

This paper designs two periodic event-triggered control protocols to achieve the output feedback leaderless consensus for nonlinear multi-agent systems with uniform input delays, and leader-follower consensus with nonuniform input delays. The involved system permits uncertainties in input delays and inherent nonlinearities, which brings the challenge on the design of observers. To overcome this challenge, a compensator is constructed independent of nonlinearities and input delays for every agent. To guarantee the event-triggered communication among different agents, two new periodic triggering mechanisms are designed based on the discrete output information. Moreover, these triggering mechanisms naturally avoid constantly monitoring and the Zeno behavior since triggering conditions are only checked on sampling points. Specifically, the designed event-triggered output feedback control protocols achieve convergence not boundedness of consensus errors. Finally, two examples are listed to demonstrate the validity of the two event-triggered strategies.

Novel perturbation mechanism underlying the network fragility evolution

Studies have shown that fragility is an effective marker for seizures and seizure onset zone (SOZ). Through analysis and simulation of a probabilistic neural network under different inputs, the regularization mechanism of external input perturbations on the fragility is explored. It is theoretically found that the fragility of a perturbed node within seizure network is inversely associated with the received perturbation input, while the fragility of the other unperturbed nodes always oppositely changes with this perturbed node. By terming the node with high fragility as the fragile node (FN), it is interestingly shown that the FN would evolve to the node with the smallest input. Then, the network fragility is further investigated. Results show that the non-uniform perturbation inputs can more easily impact the network fragility. In addition, noise-induced variations of network connection can degrade the network fragility to some extent. Finally, the real data from a patient with epilepsy have verified the universality of the above obtained findings. These results may provide possible insights into stimulation strategies for seizure control in clinic.

A random non-uniform wave input method based on complex incident site waves

To improve the deficiency of the wave input method (WIM) in expressing complex incident site waves, in this paper, a random non-uniform wave input method (RNUWIM) is derived from the use of the random incident waves and the visco-elastic boundary condition. The random incident waves are generated and adjusted by the comparison of the horizontal ground motion’s response spectrum and the target spectrum, so that they are representative of the actual wavefield in project sites. Furthermore, the RNUWIM is implemented and verified by the finite element method. The result shows that the RNUWIM inherits the advantages of the WIM in simulating the wave propagation effect and the radiation damping effect. In the new method, the spatial variation of ground motions such as time delays, peak variation, and time-history shape differences can be simulated synthetically. The errors of RNUWIM in ground acceleration, velocity, and displacement at the reference point are respectively 3.23%, 0.66%, and 0.07% in one case, which proves the efficiency of the RNUWIM in seismic response analyses.

Statistical Study of Bisection Method for Cubic Equations with Random Coefficients

The bisection method is an iterative approach used in numerical analysis to find solutions to nonlinear equations. The main purpose of this paper is to study how the parameters of a probability distribution characterizing the coefficients of a cubic polynomial can influence the convergence of the bisection method. The study covers discrete and continuous distributions, including discrete uniform, continuous uniform, and normal distributions. It was found that for both types of uniform distribution inputs, a second-degree polynomial equation can predict the average iteration for a given parameter r, where r indicates the distribution interval [−r, r]. Interestingly, the coefficients of the second-degree polynomial are nearly identical for discrete and continuous uniform distributions. For normal distribution input, the average iteration does not depend upon the standard deviation when the mean is fixed and the standard deviation is varying. But when the standard deviation is fixed and the mean is varying, the second-degree polynomial is still the best fit. This means the average iteration depends upon the mean of the normal distribution. Overall, our paper concludes that: I. For uniform distribution input, the average iteration does not depend on whether the distribution is discrete or continuous but rather depends on the range of the distribution which is its parameter. II. For non-uniform distribution input, the average iteration depends on the mean of the distribution (location parameter) but not on the standard deviation (scale parameter). Finally, a curtain is raised in the future direction of research in which we propose to combine the bisection method with the regula falsi and Newton-Raphson methods to increase the rate of convergence.

Seismic response analysis of submerged floating tunnel considering non-uniform excitation and flexible-support boundary effect

In order to ensure the safety of submerged floating tunnel(SFT) during operation, it is necessary to carry out SFT response analysis under environmental loads. Given that SFTs are often constructed in valley areas situated in seismic zones, the aseismic performance of SFT needs to be investigated. However, there are important factors of non-uniform excitation and flexible-support boundary which have not been explored thoroughly. Non-uniform excitation affects the load input characteristics while flexible-support boundary affects the dynamic characteristics of structure itself. In this paper, we built a multi-point seismic wave synthesis model which fits well with the target power spectrum and a numerical response analysis model which reflects the dynamic characteristics of SFT structure well. The results based on the proposed model show that vertical non-uniform excitation can significantly increase the peak value of SFT tether tension and tube displacement with flexible-support boundary under whether artificial or natural seismic wave series input. While flexible-support boundary can availably produce vibration control effect on peak value of tether tension and tube response of whether displacement or stress indices under non-uniform wave series input. This work can provide reference for dynamic safety assessment and vibration control design of SFT to improve the development of ocean space.

Optical and thermal analysis of solar parabolic dish cavity receiver system for hydrogen production using deep learning

The solar thermochemical cycle, responsible for converting solar energy to chemical fuel, is an emerging clean hydrogen production technique. The distribution of the concentrated heat flux and temperature on the receiver is crucial in deciding solar-to-fuel conversion efficiency. Thus, the optimum configuration of a solar parabolic dish collector (PDC) is necessary for providing efficient thermal energy. The dish diameter, rim angle, and distance of the receiver aperture from the focal point of the dish characterize the optimization of the PDC. The non-uniform heat flux captured is coupled with a heat transfer model of a solar thermochemical reactor for a thermal dissociation of ZnO involved in a ZnO/Zn thermochemical cycle for hydrogen production. The directly irradiated rotating reactor comprises a cylindrical cavity packed with insulation layers and a quartz glass window. The rotating cavity is loaded with ZnO granular particles held by a centrifugal force. A 3D CFD numerical model is analyzed by coupling the Eulerian multiphase model for ZnO particle dynamics, fluid flow of gaseous mixture, RNG k-ԑ turbulence model, Energy and Discrete Ordinates (DO) radiation model, and Arrhenius reaction rate chemical reaction of ZnO dissociation. A numerical model with non-uniform heat flux input is validated for a cavity temperature, mass fraction of zinc and oxygen, and oxygen molar flow rate at the reactor outlet. An artificial neural network model is adopted to predict the heat flux and uniformity factor values of PDC. Correlation Coefficient (R2), Root Mean Square Error (RMSE), and Mean Absolute Percentage Error (MAPE) error metrices were used to evaluate the model's performance having the receiver diameter (d), the dish diameter (D), the rim angle (φ), and the distance between the focus and receiver (h) as input variables.

Wind Field Numerical Simulation of a Cable-stayed Bridge in a Mountainous Area Using Improved Inlet Boundary by CIRFG Method

Bridges in mountainous areas are indispensable nodes in transportation networks, and wind resistance capabilities have become a controlling factor of long-span bridges built in mountain areas. Therefore, it is necessary to study the characteristics of wind fields under complex terrain. An improved inlet boundary by fitting the boundary curve was proposed in this study. The inlet fluctuating wind field was generated by the Correlation Improved Random Flow Generation method (CIRFG). The results of the numerical simulations show that the fluctuating wind input generated by CIRFG tallies with the target wind field, which proves the reliability of the proposed method. The method of fitting boundary curves to give inlet wind speed profiles can achieve non-uniform wind profile inputs. The results show the wind direction of the gorge varies significantly by height. The wind speed at the summit will accelerate affected by the terrain. Also influenced by the terrain, the turbulence intensity profiles in the simulated area show an S-shape. The transverse wind and angle of attack are uneven along the main girder, especially near slopes. The conclusions obtained in the study can provide references for the wind resistance of bridges built in mountainous areas.

Distributed observer-based consensus tracking for nonlinear MASs with nonuniform input delays and disturbances

In this paper, we investigate the consensus tracking problem of nonlinear MASs with nonuniform time-varying input delays and external disturbances. For each follower, the composited disturbance observer and the state observer are employed to estimate bounded composited disturbances and unmeasured states, and a distributed observer based on output-feedback is proposed to approximate the leader’s states approachably. Sequentially, the consensus tracking control is converted into a stability control problem for the nonlinear MASs with nonuniform time-varying input delays. Subsequently, a distributed controller based on the truncated prediction approach is presented, which only depends on the boundary value of time-varying input delays. The distributed controller can render each follower synchronically stable via the Lyapunov stability theory. Finally, a group of single-link manipulators is used as an example to verify the effectiveness of the theoretical results.

Construction and Calculation Analysis of Mechanical Model of Crushed Rock Pile Treatment of Foundation Earthquake

In recent years, earthquakes have occurred frequently at home and abroad, causing huge losses to the local society. Therefore, it is important to study the mechanical response of foundations under earthquakes. In this paper, firstly, gravel piles are selected as soft ground improvement materials, and A single bulk material pile section and a peripile soil section within the influence of a single pile are considered as a study unit, and the pore water discharge from the consolidation process is equal to the volume reduction of the unit, considering the consolidation deformation of the pile body, the composite foundation consolidation equation is derived, and the consolidation equation is solved by the separation variable method through the isostrain assumption and initial boundary conditions, and the average superporous water pressure of the soil between the body and the pile is obtained. Then, a more realistic non-uniform ground motion input method is adopted, and the seismic fluctuations are converted into equivalent loads on artificial boundary nodes through the Matlab program, and the equivalent loads are applied to the established viscoelastic artificial boundary by OpenSees software to realize ground motion input. Finally, a mechanical analysis model is established by hypothesis, and with the help of mode orthogonal theory and Hilbert yellow transform, the calculation method of ground seismic mechanical response based on non-uniform ground motion input is obtained. The results show that the effect of gravel pile encryption can reduce the accumulation rate and peak of super-static pore pressure during vibration, and the surface settlement of the site is reduced by 40–50% and the surface lateral shift is reduced by 30–50% after encryption; the drainage effect has a significant effect on reducing the accumulation rate of super-static pore pressure and the post-earthquake dissipation time of the soil during vibration, and the surface settlement is reduced by about 10–20% and the lateral shift is reduced by about 30–40% after increasing the drainage effect.

Conditional Entropy-Based Construction of Multilevel Polar-Coded Modulation With Probabilistic Shaping

Code construction is a critical issue for polar-coded systems to ensure competitive performance. In this letter, we propose a conditional entropy-based method to construct multilevel polar-coded modulation (MLC-PCM) with probabilistic shaping. More specifically, we allocate the bit positions in accordance with their conditional entropies induced by channel and source polarization. We show that a surrogate channel can be used to determine the frozen set in each component polar code efficiently. The remaining positions are assigned to the information and shaping bits by the source Bhattacharyya parameter, where the shaping bits are used to achieve non-uniform channel inputs. Approximate bounds are also derived to evaluate the decoding error probability of the probabilistically shaped MLC-PCM.

Design of Generalized Enhanced Static Segment Multiplier with Minimum Mean Square Error for Uniform and Nonuniform Input Distributions

In this paper, we analyze the performances of an Enhanced Static Segment Multiplier (ESSM) when the inputs have both uniform and non-uniform distribution. The enhanced segmentation divides the multiplicands into a lower, a middle, and an upper segment. While the middle segment is placed at the center of the inputs in other implementations, we seek the optimal position able to minimize the approximation error. To this aim, two design parameters are exploited: m, defining the size and the accuracy of the multiplier, and q, defining the position of the middle segment for further accuracy tuning. A hardware implementation is proposed for our generalized ESSM (gESSM), and an analytical model is described, able to find m and q which minimize the mean square approximation error. With uniform inputs, the error slightly improves by increasing q, whereas a large error decrease is observed by properly choosing q when the inputs are half-normal (with a NoEB up to 18.5 bits for a 16-bit multiplier). Implementation results in 28 nm CMOS technology are also satisfactory, with area and power reductions up to 71% and 83%. We report image and audio processing applications, showing that gESSM is a suitable candidate in applications with non-uniform inputs.

Formation Techniques Used in Shape-Forming Microrobotic Systems with Multiple Microrobots: A Review.

Multiple robots are used in robotic applications to achieve tasks that are impossible to perform as individual robotic modules. At the microscale/nanoscale, controlling multiple robots is difficult due to the limitations of fabrication technologies and the availability of on-board controllers. This highlights the requirement of different approaches compared to macro systems for a group of microrobotic systems. Current microrobotic systems have the capability to form different configurations, either as a collectively actuated swarm or a selectively actuated group of agents. Magnetic, acoustic, electric, optical, and hybrid methods are reviewed under collective formation methods, and surface anchoring, heterogeneous design, and non-uniform control input are significant in the selective formation of microrobotic systems. In addition, actuation principles play an important role in designing microrobotic systems with multiple microrobots, and the various control systems are also reviewed because they affect the development of such systems at the microscale. Reconfigurability, self-adaptable motion, and enhanced imaging due to the aggregation of modules have shown potential applications specifically in the biomedical sector. This review presents the current state of shape formation using microrobots with regard to forming techniques, actuation principles, and control systems. Finally, the future developments of these systems are presented.

Cooperative Constrained Control of Autonomous Vehicles With Nonuniform Input Quantization

Responsive and highly controlled autonomous vehicles (AVs) on the road easily form platoons with small inter-vehicle gaps to increase traffic throughput and reduce energy consumption. Unlike human drivers, AVs typically employ discrete quantization signals and only generate finite discrete acceleration. Finite discrete acceleration easily causes unstable inter-vehicle gaps and inter-vehicle collisions, especially when there are inaccurate quantization parameters due to quantizer aging. Based on the backstepping technique, this paper proposes the cooperative constrained control algorithm under unknown quantization parameters for AVs with nonuniform input quantization and motion uncertainties. The control laws of each AV in the platoon only depend on the state information of its consecutive AVs. The position trajectories of consecutive AVs are constructed as inter-vehicle collision-free constraints to avoid emergency braking and inter-vehicle collisions. In addition, the consistent characteristic of two types of nonuniform quantizers (i.e., logarithmic quantizers and hysteresis quantizers) is proposed, and unknown disturbances in vehicle motion are compensated. The Lyapunov based local stability and string stability are proved. Numerical simulation shows that the control algorithm is applicable to AVs with either logarithmic quantizers or hysteresis quantizers. Compared with the previous control algorithms without considering input quantization and inter-vehicle collision avoidance, the control algorithm proposed reduces inter-vehicle spacing errors, avoids high acceleration, and avoids inter-vehicle collisions.

Application of an integrated nonuniform seismic random excitation method in tunnel engineering in soft soil areas

Nonuniform seismic excitation cannot be neglected in the seismic design of long-span tunnels in soft soil areas, as it might induce tunnel damage. This paper proposes a method that combines an integrated nonuniform seismic random excitation method with the finite element method to solve the problem of nonuniform seismic input for geotechnical tunnel seismic design in soft soil areas. The integrated nonuniform seismic random excitation method of producing multipoint random seismic waves is investigated considering the effects of passage, incoherence, and energy attenuation. Design codes are developed to achieve random seismic ground waves, baseline adjustment and filtering, and inversion of bedrock. The correctness of an integrated nonuniform seismic random excitation is verified by analysis of the differences (5%–10%) between the simulated and theoretical power spectrum functions. Based on this modelling, the dynamic responses of tunnels subjected to longitudinal and lateral seismic excitations are investigated. Nonuniform seismic excitation will cause greater tunnel deformation and internal forces than uniform seismic excitation due to the tunnel consuming more seismic energy under nonuniform seismic excitation. Longitudinal nonuniform seismic excitation induces compressive and tensile states in the tunnel, and the maximum tension is close to the tensile strength of the concrete. Lateral nonuniform seismic excitation can enlarge the bending moment and shear force, and bending moment and shear force distribution ranges are recommended for this case study.

Influence of seismic wave type and incident direction on the dynamic response of tall concrete-faced rockfill dams

Owing to the stochastic behavior of earthquakes and complex crustal structure, wave type and incident direction are uncertain when seismic waves arrive at a structure. In addition, because of the different types of the structures and terrains, the traveling wave effects have different influences on the dynamic response of the structures. For the tall concrete-faced rockfill dam (CFRD), it is not only built in the complex terrain such as river valley, but also its height has reached 300 m level, which puts forward higher requirements for the seismic safety of the anti-seepage system mainly comprising concrete face slabs, especially the accurate location of the weak area in seism. Considering the limitations of the traditional uniform vibration analysis method, we implemented an efficient dynamic interaction analysis between a tall CFRD and its foundation using a non-uniform wave input method with a viscous-spring artificial boundary and equivalent nodal loads. This method was then applied to investigate the dynamic stress distribution on the concrete face slabs for different seismic wave types and incident directions. The results indicate that dam-foundation interactions behave differently at different wave incident angles, and that the traveling wave effect becomes more evident in valley topography. Seismic wave type and incident direction dramatically influenced stress in the face slab, and the extreme stress values and distribution law will vary under oblique wave incidence. The influence of the incident direction on slab stress was particularly apparent when SH-waves arrived from the left bank. Specifically, the extreme stress values in the face slab increased with an increasing incident angle. Interestingly, the locations of the extreme stress values changed mainly along the axis of the dam, and did not exhibit large changes in height. The seismic safety of CFRDs is therefore lower at higher incident angles from an anti-seepage perspective. Therefore, it is necessary to consider both the seismic wave type and incident direction during seismic capacity evaluations of tall CFRDs.

Comparisons of Supercritical Loop Flow and Heat Transfer Behavior Under Uniform and Nonuniform High-Flux Heat Inputs

The heat transfer characteristic of supercritical water is one of the crucial issues in SuperCritical Water-Cooled Reactors (SCWRs). The efficiency and safety of the SCWR system are largely dependent on the local heat transfer performance. This paper establishes the numerical model for supercritical water in a long vertical circular loop (inside diameter = 10 mm) and analyzes the flow and heat transfer mechanism during the transition process from subcritical to supercritical states under various heat fluxes (uniform and nonuniform). The results reveal that the difference in thermophysical properties between the boundary layer and the core region is the main reason for the heat transfer behavior, especially during the transition from subcritical to supercritical and liquidlike to gaslike. The flow structure on the buffer layer is a dominating factor for heat transfer deterioration. The cases under variable nonuniform heat fluxes have a higher heat transfer coefficient compared with uniform heat fluxes. But, this will cause large changes of the parameter locally. The dominating factors of heat transfer deterioration under these conditions are also identified.

Dynamic Response of Lining Structure in a Long Tunnel with Different Adverse Geological Structure Zone Subjected to Non-Uniform Seismic Load

The damage of a long tunnel is found in parts with an adverse geological structure zone under an earthquake. The phenomenon is normally the consequence of a non-uniform seismic load. Thus, to reveal the mechanism of the phenomenon, the dynamic response of the lining structure in a long tunnel passing through an adverse geological structure zone subjected to a non-uniform seismic load is mainly studied in this paper. Firstly, based on the random ground motion synthesis theory, the non-uniform ground motion acceleration–time history curves that reflect local site effects, such as traveling wave effects and attenuation effects, are generated. Secondly, the behavior of the tunnel with a different adverse geological structure zone (including different inclinations, thicknesses, and lithologies) under non-uniform seismic input is studied. Then, the impact of the different adverse geological structure zone on the internal force and safety factor of the tunnel lining is analyzed. Finally, the failure characteristics of the lining structure in the tunnel crossing through the adverse geological structure zone subjected to a non-uniform seismic load are revealed. The results show that the seismic dynamic responses significantly increase under non-uniform seismic input compared with the results under uniform seismic input, and the dynamic responses distribution along the tunnel axial is distinctly different under non-uniform seismic input. The inclination and thickness of the adverse geological structure zone have a significant influence on the internal force and safety factor of the tunnel lining, while the lithology mainly acts around the adverse geological structure zone. When the inclination angle of the adverse geological structure zone is 45°, a large number of compression-bending cracks appear in the entrance and exit sections of the tunnel, and the tunnel is in the most dangerous state.

Non-Uniform Input-Based Adaptive Growing Neural Gas for Unstructured Environment Map Construction

The research and development of special robots such as excavation robots is an important way to achieve safe and efficient production in coal mines. Affected by the unstructured environment such as complex working conditions and unsteady factor disturbances, the real-time construction of section environment maps that can accurately describe the environment and facilitate trajectory planning and decision making has become a key scientific problem to be solved as soon as possible. Therefore, non-uniform input based adaptive growing neural gas for unstructured environment map construction has been proposed. Considering complex load identification, real-time location identification, and the types of unsteady disturbance factors and working conditions, a set of environment identification models has been established based on a large amount of underground measured data training. These models can express whether the section environment has changed, as well as the type and magnitude of the change, to realize the overall knowledge extraction and parametric representation of the unstructured environment. Then, in order to solve the problems of inaccurate topology, excessive aging of connecting edges, and excessive deletion of nodes in non-uniform input environment, an adaptive growing neural gas algorithm based on non-uniform input environment (AGNG-NU) is proposed. Featured by a dynamic response deletion mechanism and adaptive adjustment mechanism of neuron parameters, the generated nodes and their topology can be dynamically adjusted according to the density of regional sample points. Several sets of non-uniform input environments are set to test the algorithm. The experimental results show that the topological maps established by AGNG-NU express clearer environmental information and, at the same time, the accuracy and distribution are improved by 8% and 15%, respectively, compared with the basic GNG algorithm. The accuracy and the distribution have also been significantly improved compared with other common SOM and GCS algorithms.