• The Journal of Korea Robotics Society
• /
• v.17 no.1
• /
• pp.76-85
• /
• 2022

One of the most fundamental challenges when designing controllers for dynamic systems is the adjustment of controller parameters. Usually the system model is used to get the initial controller, but eventually the controller parameters must be manually adjusted in the real system to achieve the best performance. To avoid this manual tuning step, data-driven methods such as machine learning were used. Recently, reinforcement learning became one alternative of this problem to be considered as an agent learns policies in large state space with trial-and-error Markov Decision Process (MDP) which is widely used in the field of robotics. However, on initial training step, as an agent tries to explore to the new state space with random action and acts directly on the controller parameters in real systems, MDP can lead the system safety-critical system failures. Therefore, the issue of 'safe exploration' became important. In this paper we meet 'safe exploration' condition with Control Barrier Function (CBF) which converts direct constraints on the state space to the implicit constraint of the control inputs. Given an initial low-performance controller, it automatically optimizes the parameters of the control law while ensuring safety by the CBF so that the agent can learn how to predict and control unknown and often stochastic environments. Simulation results on a quadrotor UAV indicate that the proposed method can safely optimize controller parameters quickly and automatically.

• Proceedings of the Korean Society for Agricultural Machinery Conference
• /
• 1993.10a
• /
• pp.734-743
• /
• 1993

The performances of three same type of vibrating potato diggers were estimated by observing the potato separation and material flow on the bottom plate. Four-bar mechanism were adopted for three diggers and pairs of eccentric cams on both sides of driving shaft were used as driving link of the diggers. Each machine was tested with different amplitudes , frequencies, and travels speeds. Blade performance were observed in three categories : Impossible forward travel , acceptable operation, and unsatisfactory potato digging , but good material flow. Three parameters were used to set marginal values that enable the machines operate for potato digging, and the parameters were compared to select best one. Three parameters are λ, $\rho$, and K.λ is the ratio of vibrating speed to travel speed, $\rho$ is the ratio of blade acceleration to travel speed, and K is the ratio of blade acceleration to gravitational acceleration. K value of 2 or more is suggested to be used as design and evalu tion criterion of the vibrating digger.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.28 no.7
• /
• pp.1038-1046
• /
• 2004

In order to acquire high-speed and high-precision performances in servomechanisms, an integrated design method have been proposed. Based on strict mathematical modeling and analysis of system performance according to design and operating parameters, a nonlinear constrained optimization problem including the relevant subsystem parameters of the servomechanism is formulated. Optimum design results of mechanical and electrical parameters are obtained according to the design parameters specified by designers through the integrated design processes. Motors are optimally selected from the servo motor database. Both the geometric errors referring to Abbe offset and the contour errors are minimized while required constraints such as stability conditions and saturated conditions are satisfied. This design methodology both offers the improved possibility to evaluate and optimize the dynamic motion performance of the servomechanism and improves the quality of the design process to achieve the required performance for high-speed/precision servomechanisms.

• 한국기계연구소 소보
• /
• s.19
• /
• pp.125-139
• /
• 1989

As an adaptive control function generator, the CMAC (Cerebellar Model Arithmetic or Articulated Controller) based learning control has drawn a great attention to realize a rather robust real-time manipulator control under the various uncertainties. There remain, however, inherent problems to be solved in the CMAC application to robot motion control or perception of sensory information. To apply the CMAC to the various unmodeled or modeled systems more efficiently, it is necessary to analyze the effects of the CMAC control parameters on the trained net. Although the CMAC control parameters such as size of the quantizing block, learning gain, input offset, and ranges of input variables play a key role in the learning performance and system memory requirement, these have not been fully investigated yet. These parameters should be determined, of course, considering the shape of the desired function to be trained and learning algorithms applied. In this paper, the interrelation of these parameters with learning performance is investigated under the basic learning schemes presented by authors. Since an analytic approach only seems to be very difficult and even impossible for this purpose, various simulations have been performed with pre specified functions and their results were analyzed. A general step following design guide was set up according to the various simulation results.

• The KIPS Transactions:PartB
• /
• v.14B no.5
• /
• pp.377-382
• /
• 2007

This paper proposes a training algorithm for global optimization of the parameters of radial basis function networks. Since conventional training algorithms usually perform only local optimization, the performance of the network is limited and the final network significantly depends on the initial network parameters. The proposed hybrid simulated annealing algorithm performs global optimization of the network parameters by combining global search capability of simulated annealing and local optimization capability of gradient-based algorithms. Via experiments for function approximation problems, we demonstrate that the proposed algorithm can find networks showing better training and test performance and reduce effects of the initial network parameters on the final results.

• Korean Journal of Air-Conditioning and Refrigeration Engineering
• /
• v.27 no.7
• /
• pp.354-361
• /
• 2015

The building energy audit is an important process when collecting basic information for improving the energy performance of existing buildings. Audit parameters should be associated with the energy performance of the building. Such audit parameters will vary according to an individual building's characteristics and energy consumption patterns, but most building energy audits are performed in the same way. The sensitivity analysis (SA) is a statistical method to quantify the correlation between inputs and outputs that can determine which input is influential to which output. Therefore, an SA can identify influential parameters when applied to building energy analysis. In this paper, we adopted the Morris method to identify building energy audit parameters and performed a Monte Carlo simulation for uncertainty analysis. As a result, this method was able to identify an influential parameter for building energy audits and reduce uncertainty in energy consumption in buildings.

• Smart Structures and Systems
• /
• v.18 no.2
• /
• pp.293-315
• /
• 2016

Seismic isolation is often used in protecting mission-critical structures including hospitals, data centers, telecommunication buildings, etc. Such structures typically house vibration-sensitive equipment which has to provide continued service but may fail in case sustained accelerations during earthquakes exceed threshold limit values. Thus, peak floor acceleration is one of the two main parameters that control the design of such structures while the other one is peak base displacement since the overall safety of the structure depends on the safety of the isolation system. And in case peak base displacement exceeds the design base displacement during an earthquake, rupture and/or buckling of isolators as well as bumping against stops around the seismic gap may occur. Therefore, obtaining accurate peak floor accelerations and peak base displacement is vital. However, although nominal design values for isolation system and superstructure parameters are calculated in order to meet target peak design base displacement and peak floor accelerations, their actual values may potentially deviate from these nominal design values. In this study, the sensitivity of the seismic performance of structures equipped with linear and nonlinear seismic isolation systems to the aforementioned potential deviations is assessed in the context of a benchmark shear building under different earthquake records with near-fault and far-fault characteristics. The results put forth the degree of sensitivity of peak top floor acceleration and peak base displacement to superstructure parameters including mass, stiffness, and damping and isolation system parameters including stiffness, damping, yield strength, yield displacement, and post-yield to pre-yield stiffness ratio.

• Structural Engineering and Mechanics
• /
• v.43 no.1
• /
• pp.127-145
• /
• 2012

One of the most important and challenging steps in seismic vulnerability and performance assessment of highway bridges is the determination of the bridge component damage parameters and their corresponding limit states. These parameters are very essential for defining bridge damage state as well as determining the performance of highway bridges under a seismic event. Therefore, realistic damage limit states are required in the development of reliable fragility curves, which are employed in the seismic risk assessment packages for mitigation purposes. In this article, qualitative damage assessment criteria for ordinary highway bridges are taken into account considering the critical bridge components in terms of proper engineering demand parameters (EDPs). Seismic damage of bridges is strongly related to the deformation of bridge components as well as member internal forces imposed due to seismic actions. A simple approach is proposed for determining the acceptance criteria and damage limit states for use in seismic performance and vulnerability assessment of ordinary highway bridges in Turkey constructed after the 1990s. Physical damage of bridge components is represented by three damage limit states: serviceability, damage control, and collapse prevention. Inelastic deformation and shear force demand of the bent components (column and cap beam), and superstructure displacement are the most common causes for the seismic damage of the highway bridges. Each damage limit state is quantified with respect to the EDPs: i.e. curvature and shear force demand of RC bent components and superstructure relative displacement.

• Journal of Power Electronics
• /
• v.14 no.4
• /
• pp.806-813
• /
• 2014

Planar electromagnetic interference (EMI) filters are widely used to restrain the conducted EMI of switching power supplies. Such filters are characterized by small size, low parasitic parameters, and better high-frequency performance than the passive discrete EMI filter. However, EMI filter performance cannot be exactly predicted by using existing methods. Therefore, this paper proposes a method to use scattering parameters (S-parameters) for the measurement of EMI filter performance. A planar EMI filter sample is established. From this sample, the relationship between S-parameters and insertion gain (IG) of EMI filter is derived. To determine the IG under different impedances, the EMI filter is theoretically calculated and practically measured. The differential structure of the near-field coupling model is also deduced, and the IG is calculated under standard impedance conditions. The calculated results and actual measurements are compared to verify the feasibility of the theory.

• Journal of Sensor Science and Technology
• /
• v.19 no.6
• /
• pp.490-496
• /
• 2010

In this study, in order to overcome the weakness of acoustic field analysis which is generally used for ultrasonic probe performance evaluation, automatic acoustic field measurement system and evaluation parameters were proposed. The comparisons between acoustic field simulation and measured acoustic distribution data of normal and abnormal channels were conducted to evaluate the availability of proposed system and evaluation parameters. First, the impulse response characteristic of sample probe was investigated to classify the normal elements and abnormal elements. And then, normal channels and abnormal channels with abnormal element were chosen. The suggested 12 evaluation parameters were calculated using the acoustic fields of these channels. The availability of proposed automatic acoustic field measurement system and evaluation parameters was confirmed. And the performance evaluation of ultrasonic probe using acoustic field analysis could be easier and faster.