Model Uncertainty Research Articles

Confidence-based design optimization using multivariate kernel density estimation under insufficient input data

The uncertainty quantification of the input statistical model in reliability-based design optimization (RBDO) has been widely investigated for accurate reliability analysis, and it could be estimated through its characteristics, cumulative experiences, and available data. However, uncertainty quantification of random variables in existing RBDO studies has exploited parametric distributions quantifying the uncertainty through the Bayes' theorem. In addition, a correlation between random variables is often underestimated due to a lack of knowledge and difficulty to describe the high-dimensional correlation. Hence, it has been a challenge to properly quantify input statistical model and its uncertainty. Therefore, a multivariate kernel density estimation (KDE) is employed to perform data-driven confidence-based design optimization (CBDO) for effective quantification of input model uncertainty. Any assumption on input distribution is not necessary since it is established only with the given input data. Moreover, the input model uncertainty due to insufficient data is quantified using bootstrapping and optimal adaptive bandwidth matrices through the Bayes’ theorem using cross-validation error. Consequently, the proposed CBDO with given input data is capable of finding a conservative optimum of RBDO accounting for both aleatory uncertainty of random variables and epistemic uncertainty induced by a limited number of input data through the multivariate KDE.

Accelerating material discovery with a threshold-driven hybrid acquisition policy-based Bayesian optimization

Advancements in materials play a crucial role in technological progress. However, the process of discovering and developing materials with desired properties is often impeded by substantial experimental costs, extensive resource utilization, and lengthy development periods. To address these challenges, modern approaches often employ machine learning (ML) techniques such as Bayesian Optimization (BO), which streamline the search for optimal materials by iteratively selecting experiments that are most likely to yield beneficial results. However, traditional BO methods, while beneficial, often struggle with balancing the trade-off between exploration and exploitation, leading to sub-optimal performance in accelerated material discovery processes. This paper introduces a novel Threshold-Driven UCB-EI Bayesian Optimization (TDUE-BO) method, which dynamically integrates the strengths of Upper Confidence Bound (UCB) and Expected Improvement (EI) acquisition functions to optimize the material discovery process. Unlike the classical BO, our method focuses on efficiently navigating the high-dimensional material design space (MDS). TDUE-BO begins with an exploration-focused UCB approach, ensuring a comprehensive initial sweep of the MDS. As the model gains confidence, indicated by reduced uncertainty, it transitions to the more exploitative EI method, focusing on promising areas identified earlier. The UCB-to-EI switching policy, which is steered by ongoing monitoring of model uncertainty at each stage of sequential sampling, enables more efficient navigation through the MDS while guaranteeing quicker convergence. The effectiveness of TDUE-BO is demonstrated through its application on three different material science datasets, showing significantly better approximation and optimization performance over traditional EI and UCB-based BO methods in terms of the RMSE scores and convergence efficiency, respectively.

Addressing Model Uncertainties in Finite Element Simulation of Electrically Stimulated Implants for Critical-Size Mandibular Defects.

Electrical stimulation is known to enhance bone healing. Novel electrostimulating devices are currently being developed for the treatment of critical-size bone defects in the mandible. Previous numerical models of these devices did not account for possible uncertainties in the input data. We present the numerical model of an electrically stimulated minipig mandible, including optimization and uncertainty quantification (UQ) methods that allow us to determine the most influential parameters. Uncertainties in the optimized finite element model are quantified using the polynomial chaos method that is implemented in the open-source Python toolbox Uncertainpy. The volumes of understimulated, beneficially stimulated, and overstimulated tissue are considered quantities of interest because they may significantly impact the expected healing success. Further, the current is a substantial quantity, limiting the lifetime of a battery-driven stimulation unit. With sensitivity analyses, the most critical parameters in the numerical model can be identified. Thus, we can learn which parameters are particularly relevant, for example, when conceptualizing the stimulation unit or planning the manufacturing process. The results of this study show that the parameters of the electrode-tissue interface (ETI), as well as the conductivity within the defect volume, have the most significant impact on the model results. The UQ results suggest that careful characterization of the ETI and the dielectric tissue properties is crucial to reduce these uncertainties. The numerical model regarding uncertainties yields important implications for reliable implant design and clinical translation.

Can input reconstruction be used to directly estimate uncertainty of a dose prediction U-Net model?

The reliable and efficient estimation of uncertainty in artificial intelligence (AI) models poses an ongoing challenge in many fields such as radiation therapy. AI models are intended to automate manual steps involved in the treatment planning workflow. We focus in this study on dose prediction models that predict an optimal dose trade-off for each new patient for a specific treatment modality. They can guide physicians in the optimization, be part of automatic treatment plan generation or support decision in treatmentindication. Most common uncertainty estimation methods are based on Bayesian approximations, like Monte Carlo dropout (MCDO) or Deep ensembling (DE). These two techniques, however, have a high inference time (i.e., require multiple inference passes) and might not work for detecting out-of-distribution (OOD) data (i.e., overlapping uncertainty estimate for in-distribution (ID) and OOD). In this study, we present a direct uncertainty estimation method and apply it for a dose prediction U-Net architecture. It can be used to flag OOD data and give information on the quality of the doseprediction. Our method consists in the addition of a branch decoding from the bottleneck which reconstructs the CT scan given as input. The input reconstruction error can be used as a surrogate of the model uncertainty. For the proof-of-concept, our method is applied to proton therapy dose prediction in head and neck cancer patients. A dataset of 60 oropharyngeal patients was used to train the network using a nested cross-validation approach with 11 folds (training: 50 patients, validation: 5 patients, test: 5 patients). For the OOD experiment, we used 10 extra patients with a different head and neck sub-location. Accuracy, time-gain, and OOD detection are analyzed for our method in this particular application and compared with the popular MCDO andDE. The additional branch did not reduce the accuracy of the dose prediction model. The median absolute error is close to zero for the target volumes and less than 1% of the dose prescription for organs at risk. Our input reconstruction method showed a higher Pearson correlation coefficient with the prediction error (0.620) than DE (0.447) and MCDO (between 0.599 and 0.612). Moreover, our method allows an easier identification of OOD (no overlap for ID and OOD data and a Z-score of 34.05). The uncertainty is estimated simultaneously to the regression task, therefore requires less time and computationalresources. This study shows that the error in the CT scan reconstruction can be used as a surrogate of the uncertainty of the model. The Pearson correlation coefficient with the dose prediction error is slightly higher than state-of-the-art techniques. OOD data can be more easily detected and the uncertainty metric is computed simultaneously to the regression task, therefore faster than MCDO or DE. The code and pretrained model are available on the gitlab repository: https://gitlab.com/ai4miro/ct-reconstruction-for-uncertainty-quatification-of-hdunet.

Hierarchical deep compartment modeling: A workflow to leverage machine learning and Bayesian inference for hierarchical pharmacometric modeling.

Population pharmacokinetic (PK) modeling serves as the cornerstone for understanding drug behavior within a specific population. It integrates subject covariates to elucidate the variability in PK parameters, thus enhancing predictive accuracy. However, covariate modeling within this framework can be intricate and time-consuming due to the often obscure structural relationship between covariates and PK parameters. Previous attempts, such as deep compartment modeling (DCM), aimed to streamline this process using machine learning techniques. Nonetheless, DCM fell short in assessing residual errors and interindividual variability (IIV), potentially leading to model misspecification and overfitting. Furthermore, DCM lacked the ability to quantify model uncertainty. To address these limitations, we introduce hierarchical deep compartment modeling (HDCM) as an advancement of DCM. HDCM harnesses machine learning to discern the interplay between covariates and PK parameters while simultaneously evaluating diverse levels of random effects and quantifying uncertainty through Bayesian inference. This tutorial provides a comprehensive application of the HDCM workflow using open-source Julia tools.

Can the young water fraction reduce predictive uncertainty in water transit time estimations?

Transit time distributions (TTDs) of streamflow are informative descriptors of catchment hydrological functioning and solute transport mechanisms. Conventional methods for estimating TTDs generally require model calibration against extensive tracer data time series, which are often limited to well-studied experimental catchments. We challenge this limitation and propose an alternative approach that uses the young water fraction (Fywobs), an increasingly used water age metric which represents the proportion of streamflow with a transit time younger than 2-3 months, and that can be robustly estimated with sparsely measured tracer data. To this end, we conducted a proof of concept study by modelling TTDs using StorAge Selection (SAS) functions with oxygen isotopes (δ18O) measurements for 23 diverse catchments in Germany. In a Monte-Carlo approach, we computed the (averaged) marginal TTDs of a prior parameter distribution and derived a model-based Fyw (Fywsim). We compared Fywsim with Fywobs, obtained from δ18O measurements, and constrained the prior SAS parameters distribution. Subsequently, we derived a posterior distribution of parameters and resulting model simulations. Our findings showed that using Fywobs to constrain the model effectively reduced parameter equifinality and simulation uncertainty. However, the value of Fywobs on reducing model uncertainty varied across sites, with larger values (Fywobs≥0.10) leading to simulations with a narrower uncertainty band and higher model efficiency, whilst smaller values (Fywobs≤0.05) had limited influence on reducing model output uncertainty. We discussed the potential and limitations of combining SAS functions with Fywobs, and considered broader implications of this approach for enhancing our understanding of catchment functioning and water quality status.

Evaluating extensions to LCDM: an application of Bayesian model averaging and selection

We present a powerful and innovative statistical framework to address key cosmological questions about the universe's fundamental properties, performing Bayesian model averaging (BMA) and model selection. Utilizing this framework, we systematically explore extensions beyond the standard ΛCDM model, considering a varying curvature density parameter ΩK, a spectral index ns = 1 and a varying n run, a constant dark energy equation of state (EOS) w 0CDM and a time-dependent one w 0 w aCDM. We also assess cosmological data against a varying effective number of neutrino species N eff. Our analysis incorporates data from various combinations of cosmic microwave background (CMB) data from the latest Planck PR4 analysis, CMB lensing from Planck 2018, baryonic acoustic oscillations (BAO), and the Bicep-KECK 2018 results.We reinforce the standard ΛCDM model statistical preference when combining CMB data withCMB lensing, BAO, and Bicep-KECK 2018 data against the K-ΛCDM model anddns /d ln k-ΛCDM with a probability > 80%. Whenevaluating the dark energy EOS, we find that this dataset does not exhibit a strong preferencebetween the standard ΛCDM model and the constant dark energy EOS model w 0CDM,with a model posterior probability distribution of approximately ≈ 40%:60% in favour of w 0CDM, while the time-varying dark energy EOS model only holds below 1%probability. We find a similar result also when considering the N eff-ΛCDM model, with asplit probability almost 50%-50% from both our datasets.Overall, our application of BMA reveals that including model uncertainty in these cases does notsignificantly impact the Hubble tension, showcasing BMA's robustness and utility in cosmologicalmodel evaluation.

Long-Term Health Outcomes of Huntington Disease and the Impact of Future Disease-Modifying Treatments: A Decision-Modeling Analysis.

Disease-modifying treatments (DMTs) such as gene therapy are currently under investigation as a potential treatment for Huntington disease (HD). Our objective was to estimate the long-term natural history of HD progression and explore the potential efficacy impacts and value of a hypothetical DMT using a decision-analytic modeling framework. We developed a health state transition model that separately analyzed 40-year-old individuals with prefunctional decline (PFD, HD Integrated Staging System [HD-ISS] stage <3, total functional score [TFC] 13), active functional decline Shoulson and Fahn category 1 (SF1, HD-ISS stage 3, TFC 13-11), and SF2 (HD-ISS stage 3, TFC 10-7). Three-year outcomes from the TRACK-HD longitudinal study were linearly extrapolated to estimate the long-term health outcomes and costs of each population. For PFD individuals, we used the HD-ISS to predict the onset of functional decline. HD costs and quality-adjusted life years (QALYs) were estimated over a lifetime horizon by applying health state-specific costs and utilities derived from a related HD burden-of-illness study. We then estimated the long-term health impacts of hypothetical DMTs that slowed or delayed onset of functional decline. We conducted sensitivity analyses to assess model uncertainties. The expected life years for 40-year-old PFD, SF1, and SF2 populations were 20.46 (95% credible range [CR]: 19.05-22.30), 13.93 (10.82-19.08), and 10.99 (8.28-22.07), respectively. The expected QALYs for PFD, SF1, and SF2 populations were 15.93 (14.91-17.44), 8.29 (6.36-11.79), and 5.79 (4.14-12.91), respectively. The lifetime costs of HD were $508,200 ($310,300 to $803,700) for the PFD population, $1.15 million ($684,500 to $1.89 million) for SF1 individuals, and $1.07 million ($571,700 to $2.26 million) for SF2 individuals. Although hypothetical DMTs led to cost savings in the PFD population by delaying the cost burdens of functional decline, they increased costs in SF1 and SF2 populations by prolonging time spent in expensive progressive HD states. Our novel HD-modeling framework estimates HD progression over a lifetime and the associated costs and QALYs. Our approach can be used for future cost-effectiveness models as positive DMT clinical trial evidence becomes available.

Integrating small data and shape prior knowledge with gradient-enhanced Kriging through adaptive knowledge sampling

Integrating prior knowledge into surrogate models constitutes an advanced approach to tackling the challenges posed by small data. The existing gradient-enhanced Kriging (GEK) method can enhance accuracy by incorporating gradient points of shape prior knowledge. Nonetheless, our findings indicate that the strategy employed for setting gradient points significantly impacts accuracy, even when the number of gradient points remains constant. To enhance the efficient utilization of gradient information, we propose a novel Adaptive Knowledge Sampling (AKS) strategy with an innovative acquisition function that balances local exploitation and global exploration, effectively extracting gradient points from shape prior knowledge. The acquisition function consists of three parts: (1) The Local Information Filling Criterion (LIFC), designed to mitigate model error; (2) The Global Information Filling Criterion (GIFC), aimed at reducing model uncertainty in data-sparse areas; and (3) The Minimum Distance Constraint (MDC), implemented to prevent excessive clustering of gradient points. To strike a balance between model accuracy and computational cost, an adaptive stopping condition is introduced to regulate the number of gradient points sampled. The impact of hyperparameters in the AKS strategy was analyzed in detail, and the method’s performance was compared to existing approaches using six benchmark functions and a missile aerodynamic prediction case. The results demonstrate that AKS-GEK offers higher prediction accuracy and robustness when integrating small data with shape prior knowledge.

Stochastic analysis of combined primary-secondary structures subjected to imprecise seismic excitation

The imprecise stochastic process model is able to incorporate both random and epistemic uncertainties, leading to a more accurate description of environmental excitations, such as earthquakes or wind loads, when limited data are available. Epistemic uncertainties in the loading model may significantly affect the performance of structural systems.The present paper addresses the stochastic analysis of combined primary-secondary structures subjected to imprecise seismic excitation modeled as a zero-mean stationary Gaussian random process, characterized by an interval-valued Power Spectral Density (PSD) function. The power and energy content of the imprecise ground motion acceleration may vary, even for the same soil category, affecting the complex dynamic behavior of combined structures and the vibration control capacity of secondary substructures under different frequency tuning conditions. The main purpose of this study is to develop a framework for investigating both the influence of epistemic uncertainties in the loading model and the interaction effects between the subsystems on the seismic performance of combined primary-secondary structures. To this aim, the stochastic analysis of combined structures under imprecise ground motion acceleration is addressed in both the frequency and time domains by an efficient approach capable of decoupling the propagation of interval and random uncertainties. Seismic safety assessment is performed in the framework of the first-passage theory by interval extension.

Incremental Value of Radiomics Features of Epicardial Adipose Tissue for Detecting the Severity of COVID-19 Infection.

Epicardial adipose tissue (EAT) is known for its pro-inflammatory properties and association with Coronavirus Disease 2019 (COVID-19) severity. However, existing detection methods for COVID-19 severity assessment often lack consideration of organs and tissues other than the lungs, which limits the accuracy and reliability of these predictive models. The retrospective study included data from 515 COVID-19 patients (Cohort 1, n=415; Cohort 2, n=100) from two centers (Shanghai Public Health Center and Brazil Niteroi Hospital) between January 2020 and July 2020. Firstly, a three-stage EAT segmentation method was proposed by combining object detection and segmentation networks. Lung and EAT radiomics features were then extracted, and feature selection was performed. Finally, a hybrid model, based on seven machine learning models, was built for detecting COVID-19 severity. The hybrid model's performance and uncertainty were evaluated in both internal and external validation cohorts. For EAT extraction, the Dice similarity coefficients (DSC) of the two centers were 0.972 (±0.011) and 0.968 (±0.005), respectively. For severity detection, the area under the receiver operating characteristic curve (AUC), net reclassification improvement (NRI), and integrated discrimination improvement (IDI) of the hybrid model increased by 0.09 (p&lt;0.001), 19.3 % (p&lt;0.05), and 18.0 % (p&lt;0.05) in the internal validation cohort, and by 0.06 (p&lt;0.001), 18.0 % (p&lt;0.05) and 18.0 % (p&lt;0.05) in the external validation cohort, respectively. Uncertainty and radiomics features analysis confirmed the interpretability of increased certainty in case prediction after inclusion of EAT features. This study proposed a novel three-stage EAT extraction method. We demonstrated that adding EAT radiomics features to a COVID-19 severity detection model results in increased accuracy and reduced uncertainty. The value of these features was also confirmed through feature importance ranking and visualization.

Determination of soil environmental criteria for high-risk trace metals in urban park soils using improved CLEA model

The trace metals (TMs) accumulated in urban park soils can pose potential threats to human health, making the management of soil quality based on health risks critically important. Based on the human health risk assessment (HHRA) model coupled with Monte Carlo Simulation, this study improved the contaminated land exposure assessment (CLEA) model. Combined with local parameters, the Soil Environmental Criteria (SEC) for high-risk trace metals (TMs) in urban park soils were calculated. Results indicated that all the mean TCR (Total carcinogenic risk) values of seven TMs exceeded the risk threshold of 1E-06, suggesting a higher likelihood of carcinogenic risks for all populations. As and Cr presented the highest potential carcinogenic risks, and were identified as high-risk TMs in the study area. The traditional CLEA model was enhanced by incorporating region-specific data, optimizing exposure parameter calculations, and addressing parameter sensitivity and uncertainty. Using the improved CLEA model, the SEC values for high-risk TMs were calculated, revealing that the SEC values gradually increased from ages 1 to 18, while significantly decreased for individuals over 80 years old. This study effectively addresses issues of parameter uncertainty and sensitivity in the CLEA model, offering new insights for the development of soil environmental quality standards.

Cost-effectiveness of sodium-glucose cotransporter-2 inhibitors in the treatment of diabetic nephropathy in Japan.

To assess the cost-effectiveness of diabetic nephropathy treatment with sodium-glucose cotransporter-2 (SGLT2) inhibitors in Japanese clinical practice, considering diabetes-related complications. A population-based Monte Carlo simulation was used to estimate the cost-effectiveness for people with diabetic nephropathy who initiated pharmacotherapy with an SGLT2 inhibitor plus conventional treatment or conventional treatment alone, based on quality-adjusted life-years (QALYs) and healthcare costs. The Canagliflozin and Renal Events in Diabetes with Established Nephropathy Clinical Evaluation study (CREDENCE) and the Japanese Society for Dialysis Research statistical survey were the primary sources of probability and mortality, while Japanese Health Insurance Claims Data were the cost source. The state transition model included diabetic nephropathy, hospitalization due to cardiovascular disease, dialysis, and death. One-way and probabilistic sensitivity analyses were used to explore model uncertainty. Using the threshold of JPY 5 000 000 per QALY, SGLT2 inhibitor plus conventional treatment was more cost-effective than conventional treatment alone, with an incremental cost-effectiveness ratio of JPY 654 309 per QALY. Treating 100 000 people, SGLT2 inhibitor plus conventional treatment prevented 2234 deaths and reduced 5793 fewer heart failure cases, 3967 fewer myocardial infarctions and stroke events. Sensitivity analysis affirmed the robustness of these results for patients aged under 70 years. The SGLT2 inhibitor treatment appeared to be cost-effective for the overall population of our study and particularly for younger patients (<70 years old). For older patients (≥70 years old), the cost-effectiveness was less clear and may require further evaluation. Decision-makers should consider this age-based heterogeneity when making recommendations about SGLT2 inhibitor treatment.

Bayesian Lower and Upper Estimates for Ether Option Prices with Conditional Heteroscedasticity and Model Uncertainty

This paper aims to leverage Bayesian nonlinear expectations to construct Bayesian lower and upper estimates for prices of Ether options, that is, options written on Ethereum, with conditional heteroscedasticity and model uncertainty. Specifically, a discrete-time generalized conditional autoregressive heteroscedastic (GARCH) model is used to incorporate conditional heteroscedasticity in the logarithmic returns of Ethereum, and Bayesian nonlinear expectations are adopted to introduce model uncertainty, or ambiguity, about the conditional mean and volatility of the logarithmic returns of Ethereum. Extended Girsanov’s principle is employed to change probability measures for introducing a family of alternative GARCH models and their risk-neutral counterparts. The Bayesian credible intervals for “uncertain” drift and volatility parameters obtained from conjugate priors and residuals obtained from the estimated GARCH model are used to construct Bayesian superlinear and sublinear expectations giving the Bayesian lower and upper estimates for the price of an Ether option, respectively. Empirical and simulation studies are provided using real data on Ethereum in AUD. Comparisons with a model incorporating conditional heteroscedasticity only and a model capturing ambiguity only are presented.

Landslide Susceptibility Assessment Based on Multisource Remote Sensing Considering Inventory Quality and Modeling

The completeness of landslide inventories and the selection of evaluation models significantly impact the accuracy of landslide susceptibility assessments. Conventional field geological survey methods and single remote-sensing technology struggle to reliably identify landslides under complex environmental conditions. Moreover, prevalent landslide susceptibility evaluation models are often plagued by issues such as subjectivity and overfitting. Therefore, we investigated the uncertainty in susceptibility modeling from the aspects of landslide inventory quality and model selection. The study focused on Luquan County in Yunnan Province, China. Leveraging multisource remote-sensing technologies, particularly emphasizing optical remote sensing and InSAR time-series deformation detection, the existing historical landslide inventory was refined and updated. This updated inventory was subsequently used to serve as samples. Nine evaluation indicators, encompassing factors such as distance to faults and tributaries, lithology, distance to roads, elevation, slope, terrain undulation, distance to the main streams, and average annual precipitation, were selected on the basis of the collation and organization of regional geological data. The information value and two coupled machine-learning models were formulated to evaluate landslide susceptibility. The evaluation results indicate that the two coupled models are more appropriate for susceptibility modeling than the single information value (IV) model, with the random forest model optimized by genetic algorithm in Group I2 exhibiting higher predictive accuracy (AUC = 0.796). Furthermore, comparative evaluation results reveal that, under equivalent model conditions, the incorporation of a remote-sensing landslide inventory significantly enhances the accuracy of landslide susceptibility assessment results. This study not only investigates the impact of landslide inventories and models on susceptibility outcomes but also validates the feasibility and scientific validity of employing multisource remote-sensing technologies in landslide susceptibility assessment.

Inversing NOx emissions based on an optimization model that combines a source-receptor relationship correction matrix and monitoring data: A case study in Linyi, China

Inversion based on observational data and the source-receptor relationship (SRR) simulated by an air quality model is an effective means to estimate NOx emissions. However, the SRR bias induced by the inherent uncertainty of the simulated model leads to potential errors in the inversed emissions, but this is seldom considered in NOx inversion. In this study, we constructed an inversion model based on the SRR correction (IMSC) by introducing a correction matrix, combined with joint regularization scheme and genetic algorithm. We innovated the dynamic acquisition method of center-restricted parameters for the correction matrix, combined this with other parts of the IMSC, attaining the multi-month and multi-region pollutant emission inversion. Hypothetical examples demonstrated that the IMSC effectively corrected the SRR and accurately estimated the NOx emissions. The IMSC was used to estimate Linyi county-level NOx emissions for January, April, July and October (typical months) during 2020 and 2021. An inversion model without SRR correction (IMWSC) was developed for comparison with the IMSC. Results showed that the IMSC more accurately, robustly, and reasonably estimated NOx emissions. Compared to the IMWSC, the IMSC improved the mean correlation between NOx emissions and NO2 observational concentrations by 25.0%, enhancing the correlation between NOx emissions and NO2 column concentrations by 111.3%. The similar NOx emission change ratios (σaverage = 5.1%) between the typical months in 2021 and 2020 among the different counties indicated a more robust performance of the IMSC than the IMWSC (σaverage = 55.2%). In addition, the NOx emissions inversed by the IMSC also showed better consistency with social activity levels (i.e., electricity consumption). The typical monthly Linyi's county-level NOx emission characterization was also studied. NOx emissions were lower in April, July, and October 2021 than the same period in 2020 due to COVID-19 and pollution controls. This study provides strategies for swiftly and accurately estimating pollutant emissions.

Semi-analytic approximate time-optimal asteroid landing with dimensionality reduction shooting

Pinpoint landing on asteroid is very challenging due to the uncertainties of the gravitational field, which highlights the urgent need for autonomous landing algorithms under such uncertainties. Some well-known guidance algorithms, such as APDG and real-time convex programming, either ignore thrust boundary constraints or are severely time consuming. In addition, these algorithms have the same difficulty in analyzing landing errors due to model uncertainty propagation, making the reliability of the algorithms in actual landings questionable. To address these challenges, we propose a landing framework that combines model identification, trajectory analytical solution and closed-loop corrections to achieve near-optimal real-time landing control. Firstly, we simplified the landing scenario, focusing on reducing the dimension of shooting variables to enable rapid trajectory calculation. Secondly, we derived error propagation equations and established criteria for trajectory replanning based on landing error prediction. Finally, we validated our approach with practical examples of the Shoemaker probe landing on Eros 433. Through real-time landing error estimation, our proposed framework enables spacecraft to achieve near time-optimal land within a given error threshold.

Water Resources’ AI–ML Data Uncertainty Risk and Mitigation Using Data Assimilation

Artificial intelligence (AI), including machine learning (ML) and deep learning (DL), learns by training and is restricted by the amount and quality of training data. Training involves a tradeoff between prediction bias and variance controlled by model complexity. Increased model complexity decreases prediction bias, increases variance, and increases overfitting possibilities. Overfitting is a significantly smaller training prediction error relative to the trained model prediction error for an independent validation set. Uncertain data generate risks for AI–ML because they increase overfitting and limit generalization ability. Specious confidence in predictions from overfit models with limited generalization ability, leading to misguided water resource management, is the uncertainty-related negative consequence. Improved data is the way to improve AI–ML models. With uncertain water resource data sets, like stream discharge, there is no quick way to generate improved data. Data assimilation (DA) provides mitigation for uncertainty risks, describes data- and model-related uncertainty, and propagates uncertainty to results using observation error models. A DA-derived mitigation example is provided using a common-sense baseline, derived from an observation error model, for the confirmation of generalization ability and a threshold identifying overfitting. AI–ML models can also be incorporated into DA to provide additional observations for assimilation or as a forward model for prediction and inverse-style calibration or training. The mitigation of uncertain data risks using DA involves a modified bias–variance tradeoff that focuses on increasing solution variability at the expense of increased model bias. Increased variability portrays data and model uncertainty. Uncertainty propagation produces an ensemble of models and a range of predictions.

Hall thruster model improvement by multidisciplinary uncertainty quantification

We study the analysis and refinement of a predictive engineering model for enabling rapid prediction of Hall thruster system performance across a range of operating and environmental conditions and epistemic and aleatoric uncertainties. In particular, we describe an approach by which experimentally-observed facility effects are assimilated into the model, with a specific focus on facility background pressure. We propose a multifidelity, multidisciplinary approach for Bayesian calibration of an integrated system comprised of a set of component models. Furthermore, we perform uncertainty quantification over the calibrated model to assess the effects of epistemic and aleatoric uncertainty. This approach is realized on a coupled system of cathode, thruster, and plume models that predicts global quantities of interest (QoIs) such as thrust, efficiency, and discharge current as a function of operating conditions such as discharge voltage, mass flow rate, and background chamber pressure. As part of the calibration and prediction, we propose a number of metrics for assessing predictive model quality. Based on these metrics, we found that our proposed framework produces a calibrated model that is more accurate, sometimes by an order of magnitude, than engineering models using nominal parameters found in the literature. We also found for many QoIs that the remaining uncertainty was not sufficient to account for discrepancy with experimental data, and that existing models for facility effects do not sufficiently capture experimental trends. Finally, we confirmed through a global sensitivity analysis the prior intuition that anomalous transport dominates model uncertainty, and we conclude by suggesting several paths for future model improvement. We envision that the proposed metrics and procedures can guide the refinement of future model development activities.

Time-efficient model predictive control for autonomous tugs with adaptive input constraints

The marine autonomous surface ship (MASS) has gained significant attention because of its wide application in waterborne transportation in recent years. The trajectory control of the MASS is crucial in ensuring safety, particularly in narrow navigation areas and urgent encounter situations. Model predictive control (MPC) is well known as a sufficient method to solve the tracking problem considering the actuator saturation with model uncertainties and environmental disturbances. However, due to the constraints on computing resources and hardware limits, the controller may fail to find a feasible solution for MPC within a reasonable amount of time, especially when the system models are coupled and complex. To improve the solution efficiency, this paper introduces adaptive constraints to reduce search space for optimization, i.e., the current tracking point’s speed is transformed into input constraints in the MPC formulation to decrease the computation burden, as well as reduce mechanical wear on the thrusters with smaller amplitudes of control actions. Simulations are carried out to evaluate the effectiveness of the proposed method with different MPC forms.