Adaptive Inversion Research Articles

Online Learning Human Behavior for a Class of Human-in-the-Loop Systems via Adaptive Inverse Optimal Control

To enhance the machines’ intelligence, it is important for them to learn how humans perform tasks. In this article, the issue of online adaptive learning human behavior is addressed for a class of human-in-the-loop (HiTL) systems using the state measurement only. The hypothesis underlying our study is that human behavior can be described by a linear quadratic optimal control model with an unknown weighting matrix for the quadratic cost function. In this model, the weighting matrix depicts the human tradeoff of various objectives. Our aim is thus to only use the system state measurement for learning the weighting matrix under the condition that human feedback gain matrix is unknown. A novel adaptive inverse optimal control approach to online learning human behavior is proposed for the HiTL system, which integrates adaptive estimation and linear matrix inequality (LMI) optimization techniques. Our approach consists of two steps: First, an adaptive law is developed to learn the human feedback gain matrix online using the system state measurement only, and second, the weighting matrix of human cost function is retrieved by solving an LMI optimization problem with the learned feedback gain matrix. Finally, simulation and experiment results on a steering assist system of intelligent vehicles are presented to illustrate the effectiveness of the proposed method.

Improved Neural Adaptive Control for Nonlinear Oscillatory Dynamic of Flapping Wings

Modeling uncertainties and oscillatory dynamics are control challenges for flapping vehicles. Flapping-wing aerial vehicles are nonlinear time-varying oscillatory systems with flexible multibody dynamics. As such, accurate modeling of these complex systems considering unsteady aerodynamics is a formidable task. Though adaptive controllers can work well with uncertain approximate models, inherent oscillation of flapping systems can degrade their control performance and create undesired control actuations. This paper addresses improvements of neural adaptive dynamic inversion controller toward efficient performance for flapping flight via effective utility of actuator capacity. In this respect, first an assumptive model of a bird-mimetic flapping-wing is considered for simulation and analysis. It is shown that the usage of oscillatory feedback data produces additional error in trajectory tracking and unnecessary oscillatory control actuations. Subsequently, the control loop is modified by adding an adaptive notch filter for real-time estimation, tracking, and removal of dominant oscillatory modes from the feedback data and consequently from the controller output. As a result, the required control effort is effectively reduced and the controller performance is significantly improved. Simulations are provided to demonstrate the efficiency of the proposed scheme in presence of noise, variation of system frequency, disturbances, as well as delay in the control loop.

On convergence rates of adaptive ensemble Kalman inversion for linear ill-posed problems

In this paper we discuss a deterministic form of ensemble Kalman inversion as a regularization method for linear inverse problems. By interpreting ensemble Kalman inversion as a low-rank approximation of Tikhonov regularization, we are able to introduce a new sampling scheme based on the Nyström method that improves practical performance. Furthermore, we formulate an adaptive version of ensemble Kalman inversion where the sample size is coupled with the regularization parameter. We prove that the proposed scheme yields an order optimal regularization method under standard assumptions if the discrepancy principle is used as a stopping criterion. The paper concludes with a numerical comparison of the discussed methods for an inverse problem of the Radon transform.

L1 Adaptive Structure-Based Nonlinear Dynamic Inversion Control for Aircraft with Center of Gravity Variations

L1 Adaptive Structure-Based Nonlinear Dynamic Inversion Control for Aircraft with Center of Gravity Variations

Robust Adaptive Path-Tracking Control of Autonomous Ground Vehicles With Considerations of Steering System Backlash

Backlash phenomena have haunted control engineers for ages. Concerning the ground vehicle path-tracking control system design, the steering system's backlash attribute is often overlooked. If not properly compensated, the backlash nonlinearity tends to provoke unfavorable consequences including erroneous positioning of the road-wheel steering angle, response chattering, limit-cycle-like oscillations, and delays. This paper pioneers to employ an adaptive inverse controller to offset the dynamics of the steering system's backlash. Overall, the path-following objective is accomplished via a novel trio-loop control architecture, which decomposes the tracking objectives into kinematic, vehicular lateral dynamics, and steering backlash compensation sub-levels. All adaptive control laws are robustified by means of sigma modification. Hardware-in-the-loop experimental results are presented to demonstrate the superiorities of the proposed solution.

An Intelligent Nonlinear Control Method for the Multistage Electromechanical Servo System

In order to meet the requirements of servo systems, including sensitive and rapid adjustment, high control and motion accuracy, and strong working adaptability, in special application fields, such as high thrust and long travel, an adaptive inversion control method is proposed for the lateral force and other nonlinear factors of multistage electromechanical servo system (MEMSS). The position tracking controller of permanent magnet synchronous motor (PMSM), based on an improved adaptive inversion method, was designed and its stability was analyzed, and the Luenberger load torque observer model of PMSM was established. The EMESS simulation model of an adaptive inversion controller was built using the Simulink platform, and the motor multibody dynamics model was established in ADAMS software. Through the joint simulation of Simulink and ADAMS software, the results of EMESS under adaptive inversion controller and traditional PID controller were compared, and the feasibility and reliability of the designed adaptive inversion controller were verified. Finally, the designed controller was tested based on the experimental platform. The experimental results show that the adaptive inversion controller designed in this paper has better robustness and stability than the traditional PID controller.

Time-lapse adaptive inversion on resistivity monitoring data with Lp norm regularization and cross-time weight

Time-lapse electrical resistivity tomography (ERT) can describe the dynamic processes occurring underground and has been widely employed in the fields of hydrology, environmental management, and engineering. The time-lapse inversion method is based on the space- and time-dimension regularization technique as it considers complementary information about more than one time interval and reduces the potential causes of the geophysical data inversion of the artifact. However, the L2 norm inversions smear out the models and cannot distinguish between the resistivity interfaces. Moreover, the determination of the two optimal regularization parameters for time-lapse inversion remains challenging, and the fixed-time regularization parameter smooths out the low-sensitivity area and inhibits the real variation on the target body of the deep region. To overcome this problem, we propose a time-lapse ERT inversion method of adaptively imaging the time-lapse model and focusing on the resistivity variation on the time dimension. This method is based on the Lp norm regularization provided by the M-estimator and uses the adaptive regularization parameters based on the model roughness and data misfit. In addition, a cross-time matrix improving the sensitivity in the deep region of model space is added to the time regularization term. We present two synthetic examples and a field dataset to validate the method. Numerical simulation and field dataset inversion results revealed that the proposed time-lapse inversion strategy can avoid the excessive smoothness caused by the low sensitivity in the model deep region and can thus enhance the response from the true ground variation to improve the accuracy of the inversion results. Owing to the adaptive regularization parameters that can preserve the dynamic balance between the objective function components, the inversion iteration progress can remain stable and can maintain the trend toward convergence.

Constrained Gravity Inversion With Adaptive Inversion Grid Refinement in Spherical Coordinates and Its Application to Mantle Structure Beneath Tibetan Plateau

AbstractWe develop a novel gravity inversion algorithm in spherical coordinates based on adaptive inversion mesh refinement and multiphysical parameter constraints. The inversion mesh is discretized into tesseroids (spherical prisms) to take the curvature of the Earth into account. To reduce the number of unknowns and computational cost, the inversion mesh is adaptively refined according to the spatial variation of parameters at each iteration. Wavelet compression is used to further reduce the computational requirement. Besides, to alleviate the nonuniqueness of inversion, the algorithm is capable of incorporating a priori models from other geophysical methods via cross‐gradient coupling and/or direct parameter coupling. For cross‐gradient coupling, the vector product of spatial gradients of two model parameters is included in the objective function. For direct parameter coupling, an a priori model is converted to a density model using the relation between two physical properties, which is then used as the initial and reference model in the inversion. A synthetic example is presented to demonstrate the effectiveness of our inversion method. Finally, we invert the gravity data over the Tibetan Plateau to obtain the density variations in the upper mantle, with a global S wave tomographic model used as a constraint. The inversion model shows that the mantle lithosphere beneath the Himalayan collision zone varies from west to east. Low‐density anomalies are observed beneath the southern and the northern Tibetan Plateau, which are consistent with published low‐velocity anomalies and magmatism distributions, possibly indicating two asthenospheric upwellings.

Neural-Network-State-Observation-Based Adaptive Inversion Control Method of Maglev Train

Maglev train is a new vehicle of rail transit. The time-varying load caused by large capacity is an important test of suspension system. If the state variable matrix and parameter matrix caused by time-varying mass cannot be estimated effectively, it is very likely to lead to the decline of suspension performance and instability. Firstly, an NNs state observer based on neural network is proposed to effectively estimate the system state and parameter matrix. Secondly, based on the NNs observer model, an inverse control algorithm with stronger regulation ability is designed by integrating the output limited characteristics and Lyapunov function, which can ensure the stability of the system when disturbed. The simulation results show that the control algorithm proposed in this paper can have faster response speed and stability in the flotation process of 16mm-10mm, and has better anti-disturbance ability in the face of time-varying mass disturbance. Finally, based on the single point suspension experimental platform, the effectiveness and practicability of the proposed control algorithm are verified, and the correctness of the proposed NNs observer model is also verified.

Determination of the vocal tract model order in iterative adaptive inverse filtering

This paper is part of a special issue on Music Technology. Accurate estimation of glottal wave is an important method for singing voice analysis via voicing patterns. In this paper, an optimising method is proposed for automatic determination of vocal tract linear prediction analysis order that follows the specific situation of different voicing scenes based on Iterative Adaptive Inverse Filtering (IAIF) aiming at getting a more accurate glottal wave from speech or singing voice signal in a non-invasive way. In comparison with existing methods that use fixed experience order, the proposed method can achieve up to 8.41% improvement in correlation coefficient with the real glottal wave.

4D inversion of resistivity monitoring data with adaptive time domain regularization

The time-lapse or 4D inversion aims to image significant underground changes and efficiently suppress inversion artefacts. However, the inverted model changes are frequently over-smoothed along the time-axis. Particularly in cases where models have poor resolution, even significant model changes are severely blurred or blunted in the reconstructed image. In this study, we developed a new 4D inversion algorithm that automatically determines time domain (TD) regularization based on model changes and the resolution to achieve a more accurate interpretation of resistivity monitoring data. By assigning a smaller TD constraint to model parameters with poor resolution, model changes at depth can be imaged if the models are significantly changed over time. Moreover, TD regularization was adaptively adjusted according to the TD error roughness, which can effectively control model smoothness. We conducted inversion experiments on synthetic and field data to test the performance of the proposed inversion algorithm. The inversion results for synthetic data demonstrated that the 4D inversion with adaptive TD regularization yields the model change that is most similar to the true change, among various options of TD regularization. The adaptive 4D inversion confirmed its capability to clearly image widespread and locally confined model changes. Model changes with poor resolution could likewise be imaged. Finally, the applicability of the proposed 4D inversion algorithm was demonstrated with real resistivity monitoring data collected for leakage detection at an embankment dam in Korea.

Adaptive weight-bit inversion for state error reduction for robust and efficient deep neural network inference using MLC NAND Flash

When Flash memory is used to store the weights of a deep neural network (DNN), the inference accuracy can degrade owing to the state errors of the Flash memory. To protect the weights from state errors, the existing methods rely on an error correction code (ECC) or parity, which can incur power/storage overhead. We propose a weight-bit inversion method that minimizes accuracy loss caused by state errors without using ECC or parity. First, the method applies weight-bit inversion for state elimination (WISE), which removes the most error-prone state from MLC NAND, thereby improving the error robustness and the most significant bit (MSB) page read speed. If the initial accuracy loss caused by the WISE is unacceptable, we apply weight-bit inversion for state error reduction (WISER), which reduces weight mapping to error-prone states with minimum changes in weight value. To further improve the read speed with minimum accuracy loss, we propose an adaptive weight-bit inversion scheme that selectively applies WISE or WISER to the unit of a weight group. The simulation results imply that after 16K program-erase cycles in NAND Flash, WISER reduces the CIFAR-100 accuracy loss by 1.33X for LeNet-5, 2.92X for VGG-16, and 2.74X for Resnet-20 compared with the existing methods. In addition, the adaptive inversion technique improves the read speed by 48.6% without accuracy loss, compared with the WISER-only scheme.

Prediction of glutenite reservoirs using an adaptive Bayesian seismic inversion in the slope zone of Mahu Depression in Junggar Basin, NW China

The slope zone of Mahu Depression in Junggar Basin contains the largest glutenite oilfield discovered over the world so far. However, because of the mechanism of near-source sedimentation and rapid facies transition, the glutenite reservoirs are strongly heterogeneous which makes it difficult to be discriminated directly from seismic data. To characterize the lateral variation and improve the vertical resolution of glutenite reservoirs, we develop an adaptive impedance inversion method based on the Bayesian theory. In this method, an automatically adjusted damping factor is derived to obtain the best balance between the vertical resolution and the inversion stability according to the noise level. And the trace-by-trace recursive inversion strategy, which uses the inversion result of the previous adjacent trace as the initial model for the next, is adopted to ensure the lateral variation. Synthetic data tests on 1-D and 2-D models verify the adaptive ability and the high resolution advantage of the introduced method. Real data application results in Mahu Oilfield shows that the prediction results of glutenite reservoir have distinct lateral variation and high vertical resolution, with good agreement with the actual drilling and sedimentary trend, which indicates a good application prospect in the heterogeneous glutenite reservoir exploration.

All state constrained decentralized adaptive implicit inversion control for a class of large scale nonlinear hysteretic systems with time-delays

This paper proposes an all state constrained decentralized adaptive implicit inversion control scheme for a class of large scale nonlinear systems with unknown time delays and asymmetric saturated hysteresis. First, to address the states constrained problem, the asymmetric barrier Lyapunov function is introduced to keep the error surface within an appropriate range from the view of engineering practice to ensure the performance and safety, such as for attitude tracking of rigid spacecraft, for spacecraft approach and intersection. Second, the transmission delays between different subsystems are considered and approximated through the incorporation of the neural-network approximators and the finite coverage lemma. Third, a new hysteresis implicit inverse algorithm is designed to effectively mitigate asymmetric and saturated hysteresis nonlinearities. It should be noted that the implicit inverse implies that the analytical inverse of the asymmetric and saturated hysteresis is not required. Instead, the decoupling algorithms are designed to extract the actual control signal from the temporarily hysteretic control signal , which reduces the preliminary work of the control algorithm. Finally, all of the signals in the closed-loop system are proved to be semi-globally ultimately uniformly bounded and the tracking errors converge to an arbitrarily small residual set. The experimental results on two-machine excitation power systems in the hardware-in-loop system are presented to illustrate the effectiveness of the proposed scheme.

Feature Extraction of Music Signal Based on Adaptive Wave Equation Inversion

The digitization, analysis, and processing technology of music signals are the core of digital music technology. There is generally a preprocessing process before the music signal processing. The preprocessing process usually includes antialiasing filtering, digitization, preemphasis, windowing, and framing. Songs in the popular wav format and MP3 format on the Internet are all songs that have been processed by digital technology and do not need to be digitalized. Preprocessing can affect the effectiveness and reliability of the feature parameter extraction of music signals. Since the music signal is a kind of voice signal, the processing of the voice is also applicable to the music signal. In the study of adaptive wave equation inversion, the traditional full-wave equation inversion uses the minimum mean square error between real data and simulated data as the objective function. The gradient direction is determined by the cross-correlation of the back propagation residual wave field and the forward simulation wave field with respect to the second derivative of time. When there is a big gap between the initial model and the formal model, the phenomenon of cycle jumping will inevitably appear. In this paper, adaptive wave equation inversion is used. This method adopts the idea of penalty function and introduces the Wiener filter to establish a dual objective function for the phase difference that appears in the inversion. This article discusses the calculation formulas of the accompanying source, gradient, and iteration step length and uses the conjugate gradient method to iteratively reduce the phase difference. In the test function group and the recorded music signal library, a large number of simulation experiments and comparative analysis of the music signal recognition experiment were performed on the extracted features, which verified the time-frequency analysis performance of the wave equation inversion and the improvement of the decomposition algorithm. The features extracted by the wave equation inversion have a higher recognition rate than the features extracted based on the standard decomposition algorithm, which verifies that the wave equation inversion has a better decomposition ability.

Adaptive neural inverse optimal tracking control for uncertain multi-agent systems

An adaptive neural inverse optimal consensus control method is presented for a class of multi-agent systems (MASs) with uncertain dynamics and time-varying disturbance. The auxiliary system is constructed for every agent, and then, based on backstepping technique, neural networks are utilized to approximate the unknown function and develop the distributed controllers. Additionally, there is only one unknown parameter need to be learned for each agent. It is proved that the proposed control scheme can ensure that all states of MASs are bounded and each agent is Input-to-state stabilizable (ISS). In the meantime, the inverse optimal controller also minimizes a meaningful cost function given in advance, which contains system input and disturbance. That is, the control scheme also reduces the cost of input. Simulation experiments illustrate the feasibility of the proposed method.

Adaptive Inverse Control of a Vibrating Coupled Vessel-Riser System With Input Backlash

This article involves the adaptive inverse control of a coupled vessel-riser system with input backlash and system uncertainties. By introducing an adaptive inverse dynamics of backlash, the backlash control input is divided into a mismatch error and an expected control command, and then a novel adaptive inverse control strategy is established to eliminate vibration, tackle backlash, and compensate for system uncertainties. The bounded stability of the controlled system is analyzed and demonstrated by exploiting the Lyapunov's criterion. The simulation comparison experiments are finally presented to verify the feasibility and effectiveness of the control algorithm.

On cycle-skipping and misfit function modification for full-wave inversion: Comparison of five recent approaches

Full-waveform inversion, a high-resolution seismic imaging method, is known to require sufficiently accurate initial models to converge toward meaningful estimations of the subsurface mechanical properties. This limitation is due to the nonconvexity of the least-squares distance with respect to the kinematic mismatch. We propose a comparison of five misfit functions promoted recently to mitigate this issue: adaptive waveform inversion, instantaneous envelope, normalized integration, and two methods based on optimal transport. We explain which principles these methods are based on and illustrate how they are designed to better handle kinematic mismatch than a least-squares misfit function. By doing so, we can exhibit specific limitations of these methods in canonical cases. We further assess the interest of these five approaches for application to field data based on a synthetic Marmousi case study. We illustrate how adaptive waveform inversion and the two methods based on optimal transport possess interesting properties, making them appealing strategies applicable to field data. Another outcome is the definition of generic tools to compare misfit functions for full-waveform inversion.

An adaptive stratified joint PP and PS AVA inversion using accurate Jacobian matrix

Amplitude variation with incidence angle (AVA) analysis is an essential tool for discriminating lithology in hydrocarbon reservoirs. Compared to traditional AVA inversion using only compressional wave (P-wave) information, joint AVA inversion using PP and PS seismic data provides better estimation of rock properties (e.g., density, P- and shear wave [S-wave] velocities). Currently, the most used AVA inversions depend on the approximations of the Zoeppritz equations (e.g., the Shuey and Aki-Richards approximations), which are not suitable for formations with strong contrast interfaces and seismic data with large incidence angles. Based on the previous derivation of the accurate Jacobian matrix, we have found that the sign of each partial derivative of reflection coefficient with respect to the P-, S-wave velocities and density changes across the interface represents a good indicator for the reflection interfaces. Accordingly, we adopt an adaptive stratified joint PP and PS AVA inversion using the accurate Jacobian matrix that can automatically obtain the layer information and can be further used as a constraint in the inversion of in-layer rock properties (density and P- and S-wave velocities). Due to the use of the exact Zoeppritz equations and accurate Jacobian matrix, this proposed inversion method is more accurate than traditional AVA inversion methods, has higher computational efficiency, and can be applied to seismic wide-angle reflection data or seismic data acquired for formations with strong contrast interfaces. The model study shows that this proposed inversion method works better than the classic Shuey and Aki-Richards approximations at estimating reflection interfaces and in-layer rock properties. It also works well in handling a part of the complex Marmousi 2 model and real seismic data.

Adaptive individual weight-gain AVO inversion with smooth nonconvex regularization

Amplitude variation with offset (AVO) inversion is a widely used approach to obtain reliable estimates of elastic parameter in the fields of seismic exploration. However, the AVO inversion is an ill-posed problem because of the band-limited characteristic of seismic data. The regularization constraint plays an important role in improving inversion resolution. Total variation (TV) class regularization based on L $$_1$$ norm has been introduced in seismic inversion. But, these methods may underestimate the high-amplitude components and obtain low-resolution results. To tackle these issues, we propose to combine a smooth nonconvex regularization approach with adaptive individual weight-gain. Compared with the L $$_1$$ norm regularizers, the proposed smoothed nonconvex sparsity-inducing regularizers can lead to more accurate estimation for high-amplitude components. Different from previous regularization methods, the proposed approach also assigns different weight regularization parameters for different strata, which we call adaptive individual weight-gain strategy. To ensure sufficient minimization of the constructed objective function, a spectral Polak–Ribiere–Polyak conjugate gradient method with line search step size is used. Further, we prove that the proposed algorithm converges to a stationary point. The synthetic data tests illustrate that our approach has improved performance compared with the conventional TV class regularization methods. Field data example further verifies the higher resolution of the proposed approach.