Kyung-Hoon Jung

Positioning accuracy improvement of laser navigation using UKF and FIS

This paper presents positioning improvement of a laser navigation system (LNS) using the unscented Kalman filter (UKF) and fuzzy inference system (FIS) for an automatic guided vehicle (AGV). The existing AGVs mainly used a magnetic system or an inductive system as a guidance system. However, those systems have high initial facility cost and are difficult to maintain according to changes of environment, and it can drive only the designated path by sensors which are installed on. The laser guidance system is developed to solve these problems, but it also has limitations which are slow response time and low accuracy. Therefore, we propose a sensor fusion method for the AGV. The sensors used in this paper are encoders, a gyro and the LNS, and they are fused by the UKF and FIS. To analyze the performance of the proposed system, we designed a fork-type AGV for ourselves and performed the experiment that was repeated 10 times under the same working conditions. In experimental results, we verified that the proposed method could improve positioning accuracy of the LNS effectively. In addition, it was appropriate to apply a real AGV system for autonomous driving.

Line Tracking Method of AGV using Sensor Fusion

*** Abstract This paper present to stu dy the guidance system as localizat ion t echnique using sensor fusion and line t racking technique using virtual line for AG V(autonomous gu ided vehicle) . An existing AG V could drive on decided line only. And representative guidance systems of such guidance system are magnet-gyro guidance and wired guidance. However, t hose have had the high cost of inst allation and maintenance, and the difficulty of syst em change according to variation of working environment. To solve such problems, we make the localization syst em which is fused with a laser navigat ion and gyro, encoder. The system is robust against noise, and flexible according to working environment through sensor fusion. For line tracking of laser navigation withou t wire guidance, we set the virt ual line in program, and design t he driving controller based on difference of angle and distance between AG Vs position and decided virtual line. To experiment, we use the AG V which is made by ourselves, and experiment t he line tracking repeatedly on same experimental environment . In result, maximum distance error between decided virtual line and AG Vs position was less than 49.93mm, and we verified that t he proposed system is effic ient for line tracking of actu al A GV .

Velocity Control Method of AGV for Heavy Material Transport

This paper presents to study the velocity control method of AGV for heavy material transport. Generally, in the industries, fork-type AGV using path tracking requires high stop-precision with performing operations for 20 hours. To obtain the high stop-precision of AGV for heavy material transport, AGV requires driving technic during low speed. Hence, we use encoder with keeping the speed of AGV and study the velocity control method to improve for the stop-precision of AGV. To experiment the proposed the velocity control method, we performed the experiments engaging the pallet located 4m in front of the AGV. In the experimental result, the maximum error of stop-precision was less than 18.64mm, and we verified that the proposed method is able to control stable.

Positioning Accuracy Improvement of Laser Navigation Using Unscented Kalman Filter

This paper presents positioning improvement of a laser navigation system (LNS) using unscented Kalman filter (UKF) for an automatic guided vehicle (AGV). The existing AGVs mainly used a magnetic system or an inductive system as a guidance system. However, those have high cost and difficult maintenance according to change of environment, and can drive only the designated path which sensors are installed on. The laser guidance system is developed to solve those problems, but it has also problems which is slow response time and low accuracy. Therefore, we propose a sensor fusion method for the AGV. The sensors used in this paper are encoders, a gyro and the LNS, and they are fused by UKF. To analyze the performance of the proposed system, we designed a fork-type AGV for ourselves, and performed the experiment that was repeated 5 times under the same working conditions. In experimental results, we verified that the proposed method could improve positioning accuracy of the LNS effectively. In addition, it was appropriate to apply a real AGV system for autonomous driving.

Positioning Accuracy Improvement of Analog-type Magnetic Positioning System using Fuzzy Inference System

This paper presents a development of an analog type magnetic positioning system and its positioning accuracy improvement using fuzzy inference system. As the magnetic positioning system used on a magnet-gyro guidance system for AGV(automatic guided vehicle), it measures a position of magnet embedded in floor of the work place. The existing product of the magnetic positioning system is very expensive in Korea because it is being sold in a foreign country exclusively. Moreover, the positioning accuracy of the product is low because it uses digital type unipolar hall sensors. Hence, we developed the magnetic positioning system by ourselves and improved the positioning accuracy of the developed magnetic positioning system using fuzzy inference system. For experiment, we used the analog type magnetic positioning system which we have developed, and compared the performance of the proposed method with the performance of the existing positioning method for the magnetic positioning system. In experimental results, we verified that the proposed method improved the positioning accuracy of the magnetic positioning system.

Accuracy improvement of magnetic guidance sensor using fuzzy inference system

This paper proposes performance improvement of a magnetic guidance sensor using fuzzy inference system. The magnetic guidance sensor is used in measuring the location of AGV (automated guided vehicle). The magnetic guidance sensor had low positioning accuracy and it is inefficient because it is generally on/off digital type and can measure only N magnetic polarity. To solve those problems, it is improvement of accuracy using by fuzzy inference system, in addition, facilitates in two polarities to measure using by analog magnetic guidance sensor produced directly. For evaluate the improvement of accuracy of proposed method, analog magnetic sensor is necessary to compare it with digital magnetic guidance sensor presently. As a result, the performance of proposed method is more accuracy improvement and analog magnetic guidance sensor is possible to measure in two polarities.

Vision guidance system for AGV using ANFIS

This paper presents a study of vision guidance system of AGV using ANFIS(adaptive network fussy inference system). The vision guidance system is based on driving method that recognize obvious characteristic object like driving line and landmark. It has an advantage in its ability to get more data points than other induction sensor of AGV. However, it is hard to build such a system because the camera used for vision guidance system is severely affected by disturbance factor caused by varying brightness of light. Therefore, we have designed and created a dark-room environment to minimize this disturbance factor. However, due to the reduction of viewing-angle by minimized dark-room design, it is difficult to control using PID which is commonly used in driving control, on fast converted driving line. Therefore, this paper proposes vision guidance method of AGV using ANFIS. AGV modeling is done through kinematic analysis for camera and created dark-room environment. Steering angle by double wheel input is revised through FIS. This data is trained by hybrid ANFIS training method, and it is used for driving control. To do performance test of proposed method, we have conducted series of experiments by creating a simulation model of AGV. We also conducted a comparative analysis of proposed method with PID control.

Accuracy Improvement of Laser Navigation System using FIS and Reliability

This paper presents to study the accuracy improvement of the laser navigation using FIS(fuzzy inference system) and the reliability. As wireless guidance system, the top-mounted laser with the laser navigation can rotate with phototransistor or other optical sensors that read the return signal from reflectors mounted at the perimeter of the workspace. The type of major existing guidance systems is a wire guidance system. Because they have high accuracy and fast response time, they are used to most industries. However, their installation cost is very expensive and maintenance is very difficult because their sensors are placed approximately 1 inch below the ground or embedded in the floor. To solve those problems, the laser navigation was developed as a wire guidance system. It does not need to reconstruct a floor or ground. And it can reduce costs of installation and maintenance because changing the layout is easy. However, it is difficult to apply to an industrial field because it is easily affected by disturbances which cause loss and damage of data, and has slow respond time. Therefore, we study the accuracy improvement of the laser navigation. The proposed method is a correction method using reliability of the laser navigation. here, reliability is calculated by FIS which is designed with the analyzed characteristics of the laser navigation. For performance comparison, we use original position data form the laser navigation and position data corrected by original reliability from the laser navigation. In experimental result, we verified that the performance of the proposed method compared the others is improved by about 50% or more.

Design and Evaluation of Multitasking-Based Software Communications Architecture for Real-Time Sensor Networking Platforms

Real-time wireless sensor networking platform should satisfy requirements of sensor networking features - efficient exploitation of limited resources and energy efficiency, real-time characteristics and scalability. Existing studies on sensor networking platforms, such as TinyOS and DCOS, have been focused on solving the maximization problem of exploitation of limited resources and the energy efficiency. Envisioned application scenarios applied in sensor networking platforms need to perform real-time sensing activity as well as provide good response time with task communication, and support the scalability to improve productivity. The proposed method is able to support task scheduling which meets deadlines of all the periodic tasks and provides aperiodic tasks with good average response time. It also supports the scalability by dividing the functionalities in sensor node into periodic and aperiodic tasks. Experimental results show that our method improves real-time characteristics of sensor networking platforms by means of guaranteeing to complete the execution of all the periodic tasks within their deadlines and having a good response time for aperiodic tasks.

A Feasible Approach to Assigning System Components to Hybrid Task Sets in Real-Time Sensor Networking Platforms

This paper studies an efficient periodic and aperiodic task decomposition technique for real-time sensor networking platforms from two perspectives. First, wireless sensor networking platforms available in literature have not addressed the concept of real-time scheduling for hybrid task sets where both types of periodic and aperiodic tasks exist. Second, individual system components in real-time sensor networking platforms must be assigned to either a periodic or aperiodic task in order to provide the concept of fully optimal real-time scheduling given real-time requirements of user applications. The proposed approach guarantees the feasible scheduling of hybrid task sets while meeting deadlines of all the periodic tasks. A case study based on real system experiments is conducted to illustrate the application and efficiency of the proposed technique. It shows that our method yields efficient performance in terms of guaranteeing the deadlines of all the periodic tasks and aiming to provide aperiodic tasks with good average response time.