A New Class of Biased Linear Estimators in Deficient-rank Linear Models

In this paper, we define a new class of biased linear estimators of the vector of unknown parameters in the deficient_rank linear model based on the spectral decomposition expression of the best linear minimun bias estimator. Some important properties are discussed. By appropriate choices of bias parameters, we construct many interested and useful biased linear estimators, which are the extension of ordinary biased linear estimators in the full_rank linear model to the deficient_rank linear model. At last, we give a numerical example in geodetic adjustment.

Extremely thin but very robust:Surprising cryptogam trait combinations at the end of the leaf economics spectrum

Leaf economics spectrum(LES)describes the fundamental trade-offs between leaf structural,chemical,and physiological investments.Generally,structurally robust thick leaves with high leaf dry mass per unit area(LMA)exhibit lower photosynthetic capacity per dry mass(Amass).Paradoxically,“soft and thinleaved”mosses and spikemosses have very low Amass,but due to minute-size foliage elements,their LMA and its components,leaf thickness(LT)and density(LD),have not been systematically estimated.Here,we characterized LES and associated traits in cryptogams in unprecedented details,covering five evolutionarily different lineages.We found that mosses and spikemosses had the lowest LMA and LT values ever measured for terrestrial plants.Across a broad range of species from different lineages,Amass and LD were negatively correlated.In contrast,Amass was only related to LMA when LMA was greater than 14 g cm^(-2).In fact,low Amass reflected high LD and cell wall thickness in the studied cryptogams.We conclude that evolutionarily old plant lineages attained poorly differentiated,ultrathin mesophyll by increasing LD.Across plant lineages,LD,not LMA,is the trait that represents the trade-off between leaf robustness and physiology in the LES.

Robustifying Biased Estimation in Linear Model

The parameter estimation problem in linear model is considered when multicollinearity and outliers exist simultaneously.A class of new estimators,robust general shrunken estimators,are proposed by grafting the robust estimation techniques philosophy into the biased estimator,and their statistical properties are discussed.By appropriate choices of the shrinking parameter matrix,we obtain many useful and important estimators.A numerical example is used to illustrate that these new estimators can not only effectively overcome difficulty caused by multicollinearity but also resist the influence of outliers.

A Class of Biased Estimators Based on SVD in Linear Model

In this paper, a class of new biased estimators for linear model is proposed by modifying the singular values of the design matrix so as to directly overcome the difficulties caused by ill_conditioning in the design matrix. Some important properties of these new estimators are obtained. By appropriate choices of the biased parameters, we construct many useful and important estimators. An application of these new estimators in three_dimensional position adjustment by distance in a spatial coordiate surveys is given. The results show that the proposed biased estimators can effectively overcome ill_conditioning and their numerical stabilities are preferable to ordinary least square estimation.

A Polynomial Prediction Filter Method for Estimating Multisensor Dynamically Varying Biases

The estimation of the sensor measurement biases in a multisensor system is vital for the sensor data fusion. A solution is provided for the estimation of dynamically varying multiple sensor biases without any knowledge of the dynamic bias model parameters. It is shown that the sensor bias pseudomeasurement can be dynamically obtained via a parity vector. This is accomplished by multiplying the sensor uncalibrated measurement equations by a projection matrix so that the measured variable is eliminated from the equations. Once the state equations of the dynamically varying sensor biases are modeled by a polynomial prediction filter, the dynamically varying multisensor biases can be obtained by Kalman filter. Simulation results validate that the proposed method can estimate the constant biases and dynamic biases of multisensors and outperforms the methods reported in literature.

A Homogeneous Linear Estimation Method for System Error in Data Assimilation

In this paper, a new bias estimation method is proposed and applied in a regional ensemble Kalman filter （EnKF） based on the Weather Research and Forecasting （WRF） Model. The method is based on a homogeneous linear bias model, and the model bias is estimated using statistics at each assimilation cycle, which is different from the state augmentation methods proposed in pre- vious literatures. The new method provides a good estimation for the model bias of some specific variables, such as sea level pres- sure （SLP）. A series of numerical experiments with EnKF are performed to examine the new method under a severe weather condi- tion. Results show the positive effect of the method on the forecasting of circulation pattern and meso-scale systems, and the reduc- tion of analysis errors. The background error covarianee structures of surface variables and the effects of model system bias on EnKF are also studied under the error covariance structures and a new concept ＇correlation scale＇ is introduced. However, the new method needs further evaluation with more cases of assimilation.

An approximate point-based alternative for the estimation of variance under big BAF sampling

Background:A new variance estimator is derived and tested for big BAF(Basal Area Factor)sampling which is a forest inventory system that utilizes Bitterlich sampling(point sampling)with two BAF sizes,a small BAF for tree counts and a larger BAF on which tree measurements are made usually including DBHs and heights needed for volume estimation.Methods:The new estimator is derived using the Delta method from an existing formulation of the big BAF estimator as consisting of three sample means.The new formula is compared to existing big BAF estimators including a popular estimator based on Bruce’s formula.Results:Several computer simulation studies were conducted comparing the new variance estimator to all known variance estimators for big BAF currently in the forest inventory literature.In simulations the new estimator performed well and comparably to existing variance formulas.Conclusions:A possible advantage of the new estimator is that it does not require the assumption of negligible correlation between basal area counts on the small BAF factor and volume-basal area ratios based on the large BAF factor selection trees,an assumption required by all previous big BAF variance estimation formulas.Although this correlation was negligible on the simulation stands used in this study,it is conceivable that the correlation could be significant in some forest types,such as those in which the DBH-height relationship can be affected substantially by density perhaps through competition.We derived a formula that can be used to estimate the covariance between estimates of mean basal area and the ratio of estimates of mean volume and mean basal area.We also mathematically derived expressions for bias in the big BAF estimator that can be used to show the bias approaches zero in large samples on the order of 1n where n is the number of sample points.

Controlling underestimation bias in reinforcement learning via minmax operation

Obtaining the accurate value estimation and reducing the estimation bias are the key issues in reinforcement learning.However,current methods that address the overestimation problem tend to introduce underestimation,which face a challenge of precise decision-making in many fields.To address this issue,we conduct a theoretical analysis of the underestimation bias and propose the minmax operation,which allow for flexible control of the estimation bias.Specifically,we select the maximum value of each action from multiple parallel state-action networks to create a new state-action value sequence.Then,a minimum value is selected to obtain more accurate value estimations.Moreover,based on the minmax operation,we propose two novel algorithms by combining Deep Q-Network(DQN)and Double DQN(DDQN),named minmax-DQN and minmax-DDQN.Meanwhile,we conduct theoretical analyses of the estimation bias and variance caused by our proposed minmax operation,which show that this operation significantly improves both underestimation and overestimation biases and leads to the unbiased estimation.Furthermore,the variance is also reduced,which is helpful to improve the network training stability.Finally,we conduct numerous comparative experiments in various environments,which empirically demonstrate the superiority of our method.

Adaptive Unified Biased Estimators of Parameters in Linear Model

To tackle multi collinearity or ill-conditioned design matrices in linear models, adaptive biased estimators such as the time-honored Stein estimator, the ridge and the principal component estimators have been studied intensively. To study when a biased estimator uniformly outperforms the least squares estimator, some sufficient conditions are proposed in the literature. In this paper, we propose a unified framework to formulate a class of adaptive biased estimators. This class includes all existing biased estimators and some new ones. A sufficient condition for outperforming the least squares estimator is proposed. In terms of selecting parameters in the condition, we can obtain all double-type conditions in the literature.

A closed-form solution for moving source localization using LBI changing rate of phase difference only

Due to the deficiencies in the conventional multiple-receiver localization syste,.ns based on direction of arrival （DOA） such as system complexity of interferometer or array and ampli- tude/phase unbalance between multiple receiving channels and constraint on antenna configuration, a new radiated source localization method using the changing rate of phase difference （CRPD） measured by a long baseline interferometer （LBI） only is studied. To solve the strictly nonlinear problem, a two-stage closed-form solution is proposed. In the first stage, the DOA and its changing rate are estimated from the CRPD of each observer by the pseudolinear least square （PLS） method, and then in the second stage, the source position and velocity are found by another PLS minimiza- tion. The bias of the algorithm caused by the correlation between the measurement matrix and the noise in the second stage is analyzed. To reduce this bias, an instrumental variable （IV） method is derived. A weighted IV estimator is given in order to reduce the estimation variance. The proposed method does not need any initial guess and the computation is small. The Cramer-Rao lower bound （CRLB） and mean square error （MSE） are also analyzed. Simulation results show that the proposed method can be close to the CRLB with moderate Gaussian measurement noise.

Robust Sensor Bias Estimation for Ill-Conditioned Scenarios

Sensor bias estimation is an inherent problem in multi-sensor data fusion systems. Classical methods such as the Generalized Least Squares （GLS） method can have numerical problems with ill-conditioned sets which are common in practical applications. This paper describes an azimuth-GLS method that provides a solution to the ill-conditioning problem while maintaining reasonable accuracy com- pared with the classical GLS method. The mean square error is given for both methods as a criterion to de- termine when to use this azimuth-GLS method. Furthermore, the separation boundary between the azi- muth-GLS favorable region and that of the GLS method is explicitly plotted. Extensive simulations show that the azimuth-GLS approach is preferable in most scenarios.

Trajectory Optimization for Target Localization and Sensor Bias Calibration with Bearing-Only Information

The problem of trajectory optimization of an unmanned aerial vehicle(UAV)for static target localization with biased bearing measurements is considered.The angular bias in sensor measurements is modeled as an additive constant in the observation model and jointly estimated with the position of the target.The necessary conditions for system observability of this estimation problem is first derived analytically with geometrical interpretations provided.The trajectory of UAV is designed based on the Fisher Information Matrix(FIM)considering physical constraints to enhance the system observability.Simulation results with Monte-Carlo runs are presented to demonstrate the improvement in target localization with biased measurements by UAV trajectory optimization.

Mean Shift Registration Algorithm for Dissimilar Sensors

The mean shift registration(MSR) algorithm is proposed to accurately estimate the biases for multiple dissimilar sensors.The new algorithm is a batch optimization procedure.The maximum likelihood estimator is used to estimate the target states,and then the mean shift algorithm is implemented to estimate the sensor biases.Monte Carlo simulations show that the MSR algorithm has significant improvement in performance with reducing the standard deviation and mean of sensor biased estimation error compared with the maximum likelihood registration algorithm.The quantitative analysis and the qualitative analysis show that the MSR algorithm has less computation than the maximum likelihood registration method.

Convergence and error bounds of adaptive filtering under model structure and regressor uncertainties

Adaptive filtering algorithms are investigated when system models are subject to model structure errors and regressor signal perturbations. System models for practical applications are often approximations of high-order or nonlinear systems, introducing model structure uncertainties. Measurement and actuation errors cause signal perturbations, which in turn lead to uncertainties in regressors of adaptive filtering algorithms. Employing ordinary differential equation （ODE） methodologies, we show that convergence properties and estimation bias can be characterized by certain differential inclusions. Conditions to ensure algorithm convergence and bounds on estimation bias are derived. These findings yield better understanding of the robustness of adaptive algorithms against structural and signal uncertainties.