• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.34 no.4
• /
• pp.383-390
• /
• 2016

Remote sensing data has been widely used in the estimation of crop yields by employing statistical methods such as regression model. Machine learning, which is an efficient empirical method for classification and prediction, is another approach to crop yield estimation. This paper described the corn yield estimation in Iowa State using four machine learning approaches such as SVM (Support Vector Machine), RF (Random Forest), ERT (Extremely Randomized Trees) and DL (Deep Learning). Also, comparisons of the validation statistics among them were presented. To examine the seasonal sensitivities of the corn yields, three period groups were set up: (1) MJJAS (May to September), (2) JA (July and August) and (3) OC (optimal combination of month). In overall, the DL method showed the highest accuracies in terms of the correlation coefficient for the three period groups. The accuracies were relatively favorable in the OC group, which indicates the optimal combination of month can be significant in statistical modeling of crop yields. The differences between our predictions and USDA (United States Department of Agriculture) statistics were about 6-8 %, which shows the machine learning approaches can be a viable option for crop yield modeling. In particular, the DL showed more stable results by overcoming the overfitting problem of generic machine learning methods.

• 제어로봇시스템학회:학술대회논문집
• /
• 2004.08a
• /
• pp.891-895
• /
• 2004

This paper suggests the techniques in determining the values of the steady-state equivalent circuit parameters of a three-phase squirrel-cage induction machine using immune algorithm. The parameter estimation procedure is based on the steady state phase current versus slip and input power versus slip characteristics. The proposed estimation algorithm is of a nonlinear kind based on clonal selection in immune algorithm. The machine parameters are obtained as the solution of a minimization of least-squares cost function by immune algorithm. Simulation shows better results than the conventional approaches.

• Journal of KSNVE
• /
• v.4 no.1
• /
• pp.71-82
• /
• 1994

We propose a systematic method to select the master states, which are retained in the reduced model after the order reduction process. The proposed method is based on the fact that the range space of right eigenvector matrix is spanned by orthogonal base vectors, and tries to keep the orthogonality of the submatrix of the base vector matrix as much as possible during the reduction process. To quentify the skewness of that submatrix, we define "Absolute Singularity Factor(ASF)" based on its singular values. While the degree of observability is concerned with estimation error of state vector and up to n'th order derivatives, ASF is related only to the minimum state estimation error. We can use ASF to evaluate the estimation performance of specific partial measurements compared with the best case in which all the state variables are identified based on the full measurements. A heuristic procedure to find suboptimal master states with reduced computational burden is also proposed. proposed.

• 제어로봇시스템학회:학술대회논문집
• /
• 2000.10a
• /
• pp.130-130
• /
• 2000

This paper presents a robust fault diagnosis and fault tolerant control lot the actuator and sensor faults in the closed-loop systems affected by unknown inputs or disturbances. The fault diagnostic scheme is based on the residual set generation by using robust Parity space approach. Residual set is evaluated through the threshold test and then fault is isolated according to the decision logic table. Once the fault diagnosis module indicates which actuator or sensor is faulty, the fault magnitude is estimated by using the disturbance-decoupled optimal state estimation and a new additive control law is added to the nominal one to override the fault effect on the system. Simulation results show that the method has definite fault diagnosis and fault tolerant control ability against actuator and sensor faults.

• Journal of the Korean Society for Precision Engineering
• /
• v.6 no.1
• /
• pp.52-74
• /
• 1989

In general, the state of system can be effected by external noise and observed only through a noisy channel. Therefore we use the estimation technigue for the information of state of the system effected by noise. There are many filters such as kalman-Buchy filter, kalman filter, Extended Kalman filter algorithm, cononlinear, extended Kalman filter algorithm to the estimation of parameters is very useful and has a long history. Also a considerable number of applications of this method has been reported. In this paper, the robot control system is treated in stochastic optimal control because of the robots doing a complicated and accurate task in inapproate environment. We have conclusion that error covariance is converged and the stability of filtering is obtained.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.61 no.2
• /
• pp.276-283
• /
• 2012

In this paper, a practical quasi-optimal DOA(direction of arrival) estimator is proposed in order to develop a one-axis gimbaled ultrasonic source tracker for mobile robot applications. With help of the gimbal structure, the ultrasonic moving source tracking problem can be simply reduced to the DOA estimation. The DOA estimation is known as one of the representative long-pending nonlinear filtering problems, but the conventional nonlinear filters might be restrictive in many actual situations because it cannot guarantee the reliable performance due to the use of nonlinear signal model. This motivates us to reformulate the DOA estimation problem in the linear robust state estimation setting. Based on the assumption that the received ultrasonic signals are noisy sinusoids satisfying linear prediction property, a linear uncertain measurement model is newly derived. To avoid the DOA estimation performance degradation caused by the stochastic parameter uncertainty contained in the linear measurement model, the recently developed NCRKF (non-conservative robust Kalman filter) scheme [1] is utilized. The proposed linear DOA estimator provides excellent DOA estimation performance and it is suitable for real-time implementation for its linear recursive filter structure. The effectiveness of the suggested DOA estimation scheme is demonstrated through simulations and experiments.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.16 no.4
• /
• pp.270-275
• /
• 2016

We present an approach for acceleration actor-critic algorithm for reinforcement learning with continuous action space. Actor-critic algorithm has already proved its robustness to the infinitely large action spaces in various high dimensional environments. Despite that success, the main problem of the actor-critic algorithm remains the same-speed of convergence to the optimal policy. In high dimensional state and action space, a searching for the correct action in each state takes enormously long time. Therefore, in this paper we suggest a search accelerating function that allows to leverage speed of algorithm convergence and reach optimal policy faster. In our method, we assume that actions may have their own distribution of preference, that independent on the state. Since in the beginning of learning agent act randomly in the environment, it would be more efficient if actions were taken according to the some heuristic function. We demonstrate that heuristically-accelerated actor-critic algorithm learns optimal policy faster, using Educational Process Mining dataset with records of students' course learning process and their grades.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.30B no.8
• /
• pp.54-63
• /
• 1993

This paper proposes a method for improving the performances of the neural network for optimization by an optimal estimation of initial states. The optimal initial state that leads to the global minimum is estimated by using the stochastic approximation. And then the update rule of Hopfield model, which is the high speed deterministic algorithm using the steepest descent rule, is applied to speed up the optimization. The proposed method has been applied to the tavelling salesman problems and an optimal task partition problems to evaluate the performances. The simulation results show that the convergence speed of the proposed method is higher than conventinal Hopfield model. Abe's method and Boltzmann machine with random initial neuron output setting, and the convergence rate to the global minimum is guaranteed with probability of 1. The proposed method gives better result as the problem size increases where it is more difficult for the randomized initial setting to give a good convergence.

• Nuclear Engineering and Technology
• /
• v.21 no.2
• /
• pp.99-110
• /
• 1989

The optimal control of xenon concentration in a nuclear reactor is posed as a linear quadratic regulator problem with state feedback control. Since it is not possible to measure the state variables such as xenon and iodine concentrations directly, implementation of the optimal state feedback control law requires estimation of the unmeasurable state variables. The estimation method used is based on the Luenberger observer. The set of the reactor kinetics equations is a stiff system. This singularly perturbed system arises from the interaction of slow dynamic modes (iodine and xenon concentrations) and fast dynamic modes (neutron flux, fuel and coolant temperatures). The singular perturbation technique is used to overcome this stiffness problem. The observer-based controller of the original system is effected by separate design of the observer and controller of the reduced subsystem and the fast subsystem. In particular, since in the reactor kinetics control problem analyzed in the study the fast mode dies out quickly, we need only design the observer for the reduced slow subsystem. The results of the test problems demonstrated that the state feedback control of the xenon oscillation can be accomplished efficiently and without sacrificing accuracy by using the observer combined with the singular perturbation method.

• International Journal of Control, Automation, and Systems
• /
• v.1 no.2
• /
• pp.163-170
• /
• 2003

The optimal Bayesian filter for a single target is known to provide the best tracking performance in a cluttered environment. However, its main drawback is the increase in memory size and computation quantity over time. In this paper, the inevitable predicament of the optimal Bayesian filter is resolved in a suboptimal fashion through the use of a receding horizon strategy. As a result, the problems of memory and computational requirements are diminished. As a priori information, the horizon initial state is estimated from the validated measurements on the receding horizon. Consequently, the suboptimal algorithm proposed allows for real time implementation.