• Journal of KIISE:Software and Applications
• /
• v.31 no.8
• /
• pp.1073-1082
• /
• 2004

This paper presents a technique for an agent to adaptively behave to unforeseen and dynamic circumstances. Since the traditional methods utilized the information about an environment to control intelligent agents, they were robust but could not behave adaptively in a complex and dynamic world. A behavior-based method is suitable for generating adaptive behaviors within environments, but it is necessary to devise a hybrid control architecture that incorporates the capabilities of inference, learning and planning for high-level abstract behaviors. This Paper proposes a 2-level control architecture for generating adaptive behaviors to perceive and avoid dynamic moving obstacles as well as static obstacles. The first level is behavior-network for generating reflexive and autonomous behaviors, and the second level is to infer dynamic situation of agents. Through simulation, it has been confirmed that the agent reaches a goal point while avoiding static and moving obstacles with the proposed method.

• Journal of Advanced Navigation Technology
• /
• v.24 no.5
• /
• pp.374-381
• /
• 2020

In order to improve the vertical position accuracy, the advantages of GPS and barometric altimeter are combined and used, but in order to fuse the two sensors, the geoid altitude must be compensated. In this paper, we proposed a technique that can calculate geoid altitude in real time even in low-cost embedded systems applied to drones or autonomous vehicles. Since the reference EGM08 is determined by a polynomial of the 2160th order, real-time calculation is impossible in the embedded system. Therefore, by introducing a linear interpolation technique, the amount of calculation was increased, and the storage space was saved by 75% by using the integer geoid height as a grid point. The accuracy of the proposed technique was evaluated through simulation, and it was confirmed that the accuracy of the maximum error is -1.215 m even in the region where the geoid change is rapid.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.11 no.2
• /
• pp.414-419
• /
• 2010

This paper describes the design and performance of a hovering AUV 'NOAH' constructed at Jeju National University. We analyse the dynamic performance of NOAH using simulation program and carry out depth control test at small basin. The main purpose of NOAH is to carry out fundamental tests on its attitude control and position control. Its configuration is similar to general ROV appearance for underwater works and dimension is $0.75m{\times}0.5m{\times}0.5m$. It has 4 thrusters of 450watt for longitudinal/lateral/vertical propulsion and is equipped with a pressure sensor for measuring water depth and a magnetic compass for measuring heading angle. The navigation of the vehicle is controlled by an on-board Pentium III-class computer, which runs with the help of the Windows XP operating system. These give us an ideal environment for developing various algorithm which are needed for developing and advanced hovering AUV.

• International Journal of Contents
• /
• v.12 no.1
• /
• pp.25-43
• /
• 2016

An agent-centric event planning method is proposed for providing pedagogical experiences in an immersed environment. Two-level planning is required at in a macro-level (i.e., inter-event level) and an intra-event level to provide realistic experiences with the objective of learning declarative knowledge. The inter-event (horizontal) planning is based on search, while intra-event (vertical) planning is based on hierarchical decomposition. The horizontal search is dictated by several realistic types of association between events besides the conventional causality. The resulting schematic plan is further augmented by conditions associated with those agents cast into the roles of the events identified in the plan. Rather than following a main story plot, all the events potentially relevant to accomplishing an initial goal are derived in the final result of our planning. These derived events may progress concurrently or digress toward a new main goal replacing the current goal or event, and the plan could be merged or fragmented according to their respective lead agents' intentions and other conditions. The macro-level coherence across interconnected events is established via their common background world existing a priori. As the pivotal source of event concurrency and intricacy, agents are modeled to not only be autonomous but also independent, i.e., entities with their own beliefs and goals (and subsequent plans) in their respective parts of the world. Additional problems our method addresses for augmenting pedagogical experiences include casting of agents into roles based on their availability, subcontracting of subsidiary events, and failure of multi-agent event entailing fragmentation of a plan. The described planning method was demonstrated by monitoring implementation.

• The KIPS Transactions:PartC
• /
• v.15C no.3
• /
• pp.213-220
• /
• 2008

A MANET is an autonomous, infrastructureless system that consists of mobile nodes. In MANET, on-demand routing protocols are usually used because network topology changes frequently. The current approach in case of broken routes is to flag an error and re-initiate route discovery either at the source or at the intermediate node. Repairing these broken links is a costly affair in terms of routing overhead and delay involved. Therefore, this paper propose a NAODV(Neighbor-aware AODV) protocol that stands on the basis of an AODV. It sets up the route rapidly if it operates for setting the route directly by using sequence number of neighbor nodes without re-search the route when the route to destination node is broken. Also, it reduces loss of packets. We use NS-2 for the computer simulation and validate that the proposed scheme is better than general AODV in terms of packet delivery ratio and average end-to-end delay. Also, when the proposed protocol is applied to the large ad-hoc network with multiple nodes, the performance is more efficient.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.38 no.3
• /
• pp.228-235
• /
• 2010

This paper deals with an autonomous flight controller design problem for a tilt-rotor aircraft in rotary-wing mode. The inner-loop algorithm is designed using the output-based approximate feedback linearization. The model error originated from the feedback linearization is cancelled within allowable tolerance by using single-hidden-layer neural network. According to Lyapunov direct stability theory, the adaptive update law is derived to run the neural network on-line, which is based on the linear observer dynamics. Moreover, the outer-loop algorithm is designed to track the trajectory generated from way-point guidance. Especially, heading and flight-path angle line-of-sight guidance are applied to the outer-loop to improve accuracy of the landing tracking performance. The 6-DOF nonlinear simulation shows that the overall performance of the flight control algorithm is satisfactory even though the collective input response shows instantaneous actuator saturation for a short time due to the lack of the neural network and the saturation protection logic in that loop.

• Journal of Institute of Control, Robotics and Systems
• /
• v.14 no.8
• /
• pp.800-808
• /
• 2008

This paper presents a localization algorithm of the outdoor wheeled mobile robot using the sensor fusion method based on indirect Kalman filter(IKF). The wheeled mobile robot considered with in this paper is approximated to the two wheeled mobile robot. The mobile robot has the IMU and encoder sensor for inertia positioning system and GPS. Because the IMU and encoder sensor have bias errors, divergence of the estimated position from the measured data can occur when the mobile robot moves for a long time. Because of many natural and artificial conditions (i.e. atmosphere or GPS body itself), GPS has the maximum error about $10{\sim}20m$ when the mobile robot moves for a short time. Thus, the fusion algorithm of IMU, encoder sensor and GPS is needed. For the sensor fusion algorithm, we use IKF that estimates the errors of the position of the mobile robot. IKF proposed in this paper can be used other autonomous agents (i.e. UAV, UGV) because IKF in this paper use the position errors of the mobile robot. We can show the stability of the proposed sensor fusion method, using the fact that the covariance of error state of the IKF is bounded. To evaluate the performance of proposed algorithm, simulation and experimental results of IKF for the position(x-axis position, y-axis position, and yaw angle) of the outdoor wheeled mobile robot are presented.

• Proceedings of the KSME Conference
• /
• 2002.08a
• /
• pp.201-204
• /
• 2002

This paper describes the hydrodynamic characteristics of a test-bed AUV SNUUV-I constructed at Seoul National University. The main purpose of the AUV is to carry out fundamental control and hydrodynamic experiments. Its configuration is basically a long cylinder of 1.35m in length and 0.25m in diameter with delta-type wings near its rear end. On the edge of each wing, a thruster of 1/4HP is mounted, which is used for both drive and turn the vehicle for horizontal movement as the output control power is varied. A pair of control surfaces installed near its font part generates pitch moments for vertical movement. The 6 DOF mathematical model of SNUUV-I contains hydrodynamic forces and moments expressed in terms of a set of hydrodynamic coefficients. These coefficients can be classified into linear damping coefficients, linear inertial coefficients and nonlinear damping coefficients. It is important to estimate the exact value of these coefficients to control the vehicle precisely. Among these, the linear coefficients are known to affect the motion of the vehicle dominantly. The linear damping coefficients are estimated by using Extended Kalman Filter. The responses of the vehicle to input signals are used to estimate the hydrodynamic coefficients, which can be inferred from output signals measured from an IMU (inertial motion unit) sensor, while the linear inertial coefficients are calculated by a potential code. By using these coefficients estimated as described above, a simulation program is constructed using Matlab.

• Journal of Institute of Control, Robotics and Systems
• /
• v.21 no.3
• /
• pp.199-209
• /
• 2015

This paper presents a model predictive control (MPC) approach to control the steering angle in an autonomous vehicle. In designing a highly automated driving control algorithm, one of the research issues is to cope with probable risky situations for enhancement of safety. While human drivers maneuver the vehicle, they determine the appropriate steering angle and acceleration based on the predictable trajectories of surrounding vehicles. Likewise, it is required that the automated driving control algorithm should determine the desired steering angle and acceleration with the consideration of not only the current states of surrounding vehicles but also their predictable behaviors. Then, in order to guarantee safety to the possible change of traffic situation surrounding the subject vehicle during a finite time-horizon, we define a safe driving envelope with the consideration of probable risky behaviors among the predicted probable behaviors of surrounding vehicles over a finite prediction horizon. For the control of the vehicle while satisfying the safe driving envelope and system constraints over a finite prediction horizon, a MPC approach is used in this research. At each time step, MPC based controller computes the desired steering angle to keep the subject vehicle in the safe driving envelope over a finite prediction horizon. Simulation and experimental tests show the effectiveness of the proposed algorithm.

• Journal of the Korean Institute of Telematics and Electronics C
• /
• v.35C no.11
• /
• pp.76-88
• /
• 1998

While most of the research on reinforcement learning assumed a discrete control space, many of the real world control problems need to have continuous output. This can be achieved by using continuous mapping functions for the value and action functions of the reinforcement learning architecture. Two questions arise here however. One is what sort of function representation to use and the other is how to determine the amount of noise for search in action space. The ubiquitous neural network is used here to learn the value and policy functions. Next, the reinforcement predictor that is intended to predict the next reinforcement is introduced that also determines the amount of noise to add to the controller output. The proposed reinforcement learning architecture is found to have a sound on-line learning control performance especially at high-speed road following of high curvature road. Both computer simulation and actual experiments on a test vehicle have been performed and their efficiency and effectiveness has been verified.