http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
- 中文 을 입력하시려면 zhongwen을 입력하시고 space를누르시면됩니다.
- 北京 을 입력하시려면 beijing을 입력하시고 space를 누르시면 됩니다.
RISS 인기검색어
- 검색키워드
- 주제어 : 자율 수중 운동체 (검색결과 4 건)
-
확장된 3D 벡터필드히스토그램에 기반한 자율수중운동체의 충돌회피시스템
The study for underwater vehicle has matured significantly over the last three decades since 1970, which is applied to industrial and military fields ranging from deep-ocean exploration to detection of sea mine. Underwater vehicle is generally classified into two types, Remotely Operated Vehicle(ROV) and Autonomous Underwater Vehicle(AUV), in terms of control method. ROV is an unmanned underwater vehicle that is controlled by the tether cable from a mother ship. The other hand, AUV possesses an intelligent control-software, called intelligent navigation system, performing high-level functions such as cognition, decision and action on the same level as domain experts hold to overcome the shortcoming of ROV by which ROV has a constraint in the dynamic capacity of vehicle. Most intelligent navigation systems use global path planning and local path planning to find a safe and reasonable path to the goal. Global path planning aims at deriving way-points from a start position to the goal position based on pre-known environment information, and local path planning aims at deriving a new path between way-points when AUV meets with unknown obstacle or unexpected mission change happens. A heuristic search using evaluation function is generally chosen for the local path planning. In this paper we suggest a collision avoidance system of AUV based on expanded 3D vector field histogram(VFH). VFH method is developed for the real-time collision avoidance method for mobile robots by J. Borenstein. This method uses a two-dimensional Cartesian histogram grid as a world model, which is reduced to a one-dimensional polar histogram that is constructed around the robot's momentary location. In this paper, we expand this polar histogram into the three-dimensional polar histogram which is used as a candidate-cell set. In order to derive a reasonable and safe path to the goal position, we adopt the heuristic search based on the fuzzy relational products and the evaluation function representing heuristic information of domain experts. Fuzzy relational products are used as the mathematical implement for the analysis and synthesis of relations between obstacles and available candidate cell which could be selected as next node to move. The suggested collision avoidance system consists of four independent modules such as virtual obstacle module, relative grid module, virtual plane module, and heuristic search module. A simulation is conducted with scenarios in view of safety so as to verify the performance of proposed collision avoidance system. For this purpose, we develop a 3D simulator consists of environment manager, object, and 3D viewer by using OpenGL and Visual C++. Six scenarios are composed of different encounter situations between AUV and obstacles. The analysis of simulation result shows that the collision avoidance system of AUV has excellent performance with respect to safety of derived path.
-
-
서경철 Graduate School, Yonsei University 2007 국내석사
The dynamic model of an autonomous underwater vehicle (AUV) is highly nonlinear and has many unknown parameters due to its shape, low cruising speed and ocean environment. Therefore, the control systems of the AUV are required to learn and adapt to the variation of dynamics and hydrodynamics coefficients of vehicles in order to provide desired performance. Various control strategies have been suggested for the motion control of the AUVs, which were an adaptive control scheme, sliding mode control (SMC) and adaptive robust control techniques, etc. However, these literatures had restricting conditions and were not enough adaptations for unknown bounds of network’s weights. This thesis proposes a self recurrent wavelet neural network (SRWNN) based adaptive backstepping design technique for the robust steering control of an AUV with unknown model uncertainties and external disturbance. The SRWNN, which has the properties such as fast convergence and simple structure, is used as the uncertainty observer of the steering model of an AUV. The adaptation laws for the weights of SRWNN and reconstruction error compensator are induced from the Lyapunov stability theorem, which are used for the on-line control of AUV. Finally, simulation results for steering control of an AUV with unknown model uncertainties and external disturbance are included to illustrate the robustness against disturbance of the proposed method. 수중 자율 운동체의 다이나믹스는 수중 자율 운동체의 형태, 낮은 운항 속도 및 해양 환경 등으로 인하여 매우 비선형적이고 미지의 많은 파라미터 값들을 가지고 있다. 그러므로 이러한 수중 자율 운동체의 제어 시스템은 원하는 동작을 제공하기 위한 운동체의 유체 역학 계수 및 다이나믹스의 변화에 적응 및 학습할 수 있는 능력이 있어야 한다. 현재까지, 적응 제어 기법, 슬라이딩 제어 및 적응 강인 제어 기법 등과 같은 다양한 제어 기법들이 제안되어 왔으나, 이들 기법들은 많은 제한적 사항과 회로망 가중치의 알 수 없는 경계(bound)에 대한 충분치 못한 적응력을 가지고 있다. 따라서 본 논문에서는 자기회귀 웨이블릿 신경 회로망을 이용한 불확실성과 외란을 갖는 수중 자율 운동체의 방향 제어를 위한 적응 백스테핑 (backstepping) 제어 방법을 제안하였다. 빠른 수렴성과 간단한 구조를 갖는 자기회귀 웨이블릿 신경 회로망은 수중 자율 운동체의 방향 제어를 위한 불확실성 관측기로 사용 되어지고, 자기회귀 웨이블릿 신경회로망의 가중치를 위한 적응 법칙과 재구성 오차 보상기는 수중 자율 운동체의 온라인 제어를 위해 사용되는 르야프노프 안정도 이론으로부터 유도한다. 마지막으로, 불확실성과 외란을 가진 수중 자율 운동체의 방향 제어 모의 실험을 통해 본 논문에서 제안한 제어기법의 외란에 대한 강인성을 입증하였다.
-
자기 회귀 웨이블릿 신경 회로망을 이용한 자율 수중 운동체의 방향 제어에 관한 연구
In this paper, we propose a new method for designing the steering controller of autonomous underwater vehicle(AUV) using a self-recurrent wavelet neural network(SRWNN). The proposed control method is based on a direct adaptive control technique, and a SRWNN is used for the controller of horizontal motion of a AUV. A SRWNN is tuned to minimize errors between the SRWNN outputs and the outputs of AUV via the gradient descent(GD) method. Through the computer simulations, we compare the performance of the propose controller with that of the MLP based controller to verify and effectiveness of the propose controller. The results obtained from simulations are summarized as follows: 1) From the simulation results, it is shown that the SRWNN based controller is stable and has faster convergence than the MLP based controller. 2) It is shown that the MSEs of the MLP based controller and the SRWNN based controller are 0.0028, and 0.0019, respectively. From the results of MSE, it is shown that the SRWNN based controller has the least error.
연관 검색어 추천
이 검색어로 많이 본 자료
활용도 높은 자료
