• 로봇학회논문지
• /
• 제17권1호
• /
• 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.

• Ocean and Polar Research
• /
• 제42권1호
• /
• pp.33-47
• /
• 2020

The bulk and partition geochemistry was studied in two sediment cores collected from the axial valley of the north-central Fonualei Rift and Spreading Center (FRSC), Lau back-arc Basin, southwest Pacific. The sediments consist of mostly volcanic ash, although minor amounts of biogenic and other components were present in some intervals. The major element composition of bulk sediments recalculated to a carbonate-free basis was in good agreement with the magma compositions of the adjacent Tofua Arc and the FRSC, with only significant difference in Mn content. The enrichment of Mn and other associated elements (e.g. Cu, Co, Ni, and P) is attributed to hydrothermal input to the sediments, as evidenced by their significant partitioning in the non-detrital phases according to the partition geochemistry. Hydrogenetic and diagenetic inputs were assessed to be relatively insignificant. Estimated hydrothermal Mn fluxes during the Holocene ranged between 5.0 and 37.1 mg cm^-2 kyr^-1, with the higher values in younger sediments, suggesting enhanced hydrothermal activity. The hydrothermal Mn fluxes comparable to or higher than those reported from other spreading centers with strong hydrothermal activities may indicate the presence of unknown hydrothermal vent sites and/or topographic restriction on the dispersal of hydrothermal plumes in the northern part of the FRSC.

• 한국통신학회논문지
• /
• 제39C권9호
• /
• pp.820-827
• /
• 2014

One of the basic problems in Wireless Sensor Networks (WSNs) is the localization of the sensor nodes based on the known location of numerous anchor nodes. WSNs generally consist of a large number of sensor nodes and recording the location of each sensor nodes becomes a difficult task. On the other hand, based on the application environment, the nodes may be subject to mobility and their location changes with time. Therefore, a scheme that will autonomously estimate or calculate the position of the sensor nodes is desirable. This paper presents an intelligent localization scheme, which is an artificial neural network (ANN) based localization scheme used to estimate the position of the unknown nodes. In the proposed method, three anchors nodes are used. The mobile or deployed sensor nodes request a beacon from the anchor nodes and utilizes the received signal strength indicator (RSSI) of the beacons received. The RSSI values vary depending on the distance between the mobile and the anchor nodes. The three RSSI values are used as the input to the ANN in order to estimate the location of the sensor nodes. A feed-forward artificial neural network with back propagation method for training has been employed. An average Euclidian distance error of 0.70 m has been achieved using a ANN having 3 inputs, two hidden layers, and two outputs (x and y coordinates of the position).

• 한국산학기술학회논문지
• /
• 제15권2호
• /
• pp.1020-1026
• /
• 2014

슬라이딩모드 제어의 문제는 제어입력에 포함된 알려지지 않은 외란 등 비선형 입력으로 인한 채터링의 발생이다. 본 연구는 채터링 발생의 문제를 해결하기 위한 제어 알고리듬에 대하여 고찰하고 채터링을 억제하는 슬라이딩모드 제어기를 설계하고자 한다. 슬라이딩모드 제어 시 발생하는 채터링을 억제하기 위해 알려지지 않은 외란을 포함한 비선형 입력에 대하여 평활함수를 적용한다. 이 방법에 의하여 도립진자 시스템의 동적 방정식에 포함된 비선형 요소와 외란에 의한 문제가 해결될 수 있음을 보인다. 또 슬라이딩모드 제어를 위한 제어 입력을 시스템에 적용하였고, 제안한 제어기의 제어성능을 검증하기 위하여 도립진자를 대상으로 컴퓨터 모의실험을 수행하였다. 모의실험 결과 제어입력에 포함된 큰 폭의 채터링이 제거되었음을 확인할 수 있다.

• 한국시뮬레이션학회논문지
• /
• 제32권1호
• /
• pp.13-21
• /
• 2023

본 논문은 실제 시스템의 빅데이터가 확보되었고 시스템에 대한 정보를 일부 알고 있을 때 파라미터를 가진 그레이박스 혹은 블랙박스 형태의 가설모델을 설정하고 기계학습을 통해 모델을 자동 생성하는 기법을 제안하였다. 제안된 프레임워크를 구현하고 다양한 가설모델에 대한 실험을 통해 학습된 모델의 정합도와 가설모델의 학습에 소요되는 비용에 대해 분석하였다. 실험결과 제안된 가설모델 기반 기계학습 기법으로 상미분방정식으로 기술될 수 있은 연속시스템의 그레이박스 혹은 화이트 박스 가설모델과 주어진 빅데이터를 이용하여 모델링을 했을 때 상당히 좋은 성능과 정확도를 보인 모델을 찾아낼 수 있음을 확인하였다. 이 기법은 최근 생성된 빅데이터를 이용하여 디지털트윈 모델의 일치성을 자동 갱신하거나 새로운 입력에 대한 출력을 예측하는 목적으로도 잘 활용될 수 있을 것으로 기대된다.

• Smart Structures and Systems
• /
• 제33권4호
• /
• pp.291-300
• /
• 2024

Uncertainty of the model, system delay and drive dynamics can be considered as normal uncertainties, and the main source of uncertainty in the seismic control system is related to the nature of the simulated seismic error. In this case, optimizing the management strategy for one particular seismic record will not yield the best results for another. In this article, we propose a framework for online management of active structural management systems with seismic uncertainty. For this purpose, the concept of reinforcement learning is used for online optimization of active crowd management software. The controller consists of a differential controller, an unplanned gain ratio, the gain of which is enhanced using an online reinforcement learning algorithm. In addition, the proposed controller includes a dynamic status forecaster to solve the delay problem. To evaluate the performance of the proposed controllers, thousands of ground motion data sets were processed and grouped according to their spectrum using fuzzy clustering techniques with spatial hazard estimation. Finally, the controller is implemented in a laboratory scale configuration and its operation is simulated on a vibration table using cluster location and some actual seismic data. The test results show that the proposed controller effectively withstands strong seismic interference with delay. The goals of this paper are towards access to adequate, safe and affordable housing and basic services, promotion of inclusive and sustainable urbanization and participation, implementation of sustainable and disaster-resilient buildings, sustainable human settlement planning and manage. Simulation results is believed to achieved in the near future by the ongoing development of AI and control theory.

• 한국지하수토양환경학회지:지하수토양환경
• /
• 제12권3호
• /
• pp.53-65
• /
• 2007

지하수 모델 개발 시 수문 경계를 실제 지하수계에 부합되도록 개념화하는 것은 모델의 신뢰도를 결정하는데 매우 중요하다. 본 논문에서는 지하수 분수령, 하천, 대수층의 하부 경계면 등의 수문 경계를 모델에서 개념화할 때 수반되는 불확실성이 모델 결과에 미치는 영향을 고찰하였다. 첫째, 연구지역을 대상으로 현장시험을 수행하여 모델 입력 자료를 확보하였으며, Visual Modflow 프로그램을 이용하여 연구지역에 대한 지하수 흐름 모델을 개발하였다. 지하수 함양량을 모델 보정 인자로 설정하였으며, 현장에서 관측된 지하수위 자료를 이용하여 모델을 보정하였다. 둘째, 민감도 분석을 통하여 지하수 분수령의 위치, 하천 지류들의 경계조건 설정 여부, 암반의 하부 경계면의 위치 등이 모델 결과에 미치는 영향을 정량적으로 분석하였다. 셋째, 민감도 분석 결과에 근거하여, 국내 지하수계를 대상으로 신뢰성 있는 개념 모델을 개발하고자 할 때 요구되는 주요 내용들을 토의하였으며, 현장조사 단계에서 부지특성화를 위해 필요한 효과적인 전략을 제시하였다.

• The Korean Journal of Physiology and Pharmacology
• /
• 제1권3호
• /
• pp.251-262
• /
• 1997

Peripheral nerve injury sometimes leads to neuropathic pain and depletion of calcitonin gene related-peptide (CGRP) and substance P (SP) in the spinal cord. However, the pathophysiological mechanisms for depletion of CGRP and SP following the neurorathic injury are still unknown. This study was performed to see whether the distribution of immunoreactivity for CGRP and SP in the superficial dorsal horn and dorsal root ganglia(DRG) was related to the distance between the DRG and injury site. To this aim, we compared two groups of rats; one group was subjected to unilateral inferior and superior caudal trunk transections at the level between the S3 and S4 spinal nerves (S34 group) and the other group at the levels between the S1 and S2, between S2 and S3 and between S3 and S4 spinal nerve (S123 group). The transections in both groups equally eliminated the inputs from the tail to the S1-3 DRG, but the distance from the S1/S2 DRG to the injury site was different between the two groups. Immunostaining with SP and CGRP antibody was done in the S1-S3 spinal cord and DRG of the two groups 1 and 12 weeks after the injury. The results obtained are as follows: 1. The immunoreactivity for CGRP and SP in the ipsilateral superficial dorsal horn and DRG decreased 1 and 12 weeks after neuropathic nerve injury. 2. The immunoreactive area of SP and CGRP in the S1 dorsal horn was smaller in the S123 group than in the S34 group, whereas that in the S3 dorsal horn was not significantly different between the two groups. The number of SP-immunoreactive DRG cells decreased on the neuropathic side as compared to the sham group's in all DRGs of experimental groups except the S1 DRG of the S34 group. These results suggest that the amounts of SP and CGRP in the dorsal horn and DRG following neuropathic injury inversely decrease according to the distance between the DRG and injury site.