퍼지로직과 칼만필터를 이용한 자세추정시스템 개발 = Development of attitude heading reference system using Fuzzy logic and Kalman filter

다국어 초록 (Multilingual Abstract)

Recently, the development in micro-electro-mechanical system (MEMS) technology has made it possible to make cheap chip like accelerometer and gyro sensors, which have been adopted into many applications but traditionally inertial sensors have been too costly. Attitude Heading Reference System (AHRS) estimates body’s roll, pitch and yaw. It is widely used in robot, aircraft, ship and so on. Particularly, there is getting popular robot market. Humanoid and military robots have been interesting topics for the past decade. Therefore, inexpensive and high efficient attitude sensor is necessary for an effective attitude control.
Most of the AHRS is consist of accelerometer, gyro and magnetometer. Accelerometer measures acceleration of the body and gyro measures angular rate of the body. Before time, attitude calculation is possible by simple integration using gyro. However, there is difficult to attitude calculation as using low cost MEMS sensors. MEMS sensors have errors of bias and drift. It have an effect on accumulated error of gyro output. Therefore, there is recently study about calibrated method of gyro error using accelerometer. Magnetometer is used to yaw estimation.
Fusion method of different sensors uses widely Kalman filter. But it has trend that unstable filter by parameter of Kalman filter which is difficult to estimate. We had known through experiment that if variation of accelerometer is large then Kalman filter error is large. Using this fact, this paper proposed calibration method that controls parameter of Kalman filter using fuzzy logic.
I confirmed the AHRS through experiment. I assumed that true value is NAV420CA-100. It is commercial sensor having GPS. From the results of experiments, I known that Kalman filter algorithm results controlling parameter is better than Kalman filter fixing parameter algorithm results. When I compared developed AHRS with NAV420CA-100, it has not only error of roll and pitch within 3 degree but also yaw's error with 10 degree. Future work for increasing the accuracy is fusion algorithm with AHRS and GPS.

목차 (Table of Contents)

• 1. 서 론
• 1.1 연구 배경 및 연구 동향
• 1.2 연구 내용
• 2. 자세추정시스템
• 2.1 자세추정시스템 구성
• 2.2 가속도센서
• 2.2.1 가속도센서 보정
• 2.2.2 가속도센서를 이용한 자세추정
• 2.3 자이로센서
• 2.3.1 자이로센서 보정
• 2.3.2 자이로센서를 이용한 자세추정
• 2.4 지자기센서
• 2.4.1 지자기센서 보정
• 2.4.2 지자기센서를 이용한 자세추정
• 3. 칼만필터를 이용한 자세추정시스템
• 3.1 칼만필터 개요
• 3.2 시스템 모델
• 3.2.1 롤과 피치 추정을 위한 시스템 모델
• 3.2.2 요 추정을 위한 시스템 모델
• 3.3 관측 모델
• 3.3.1 롤과 피치 추정을 위한 관측 모델
• 3.3.2 요 추정을 위한 관측 모델
• 4. 퍼지로직를 이용한 칼만필터 계수 조절
• 4.1 퍼지로직 개요
• 4.2 퍼지로직을 이용한 칼만필터 계수 조절 알고리즘
• 4.2.1 칼만필터 계수 조절을 위한 퍼지화기
• 4.2.2 칼만필터 계수 조절을 위한 퍼지 추론기
• 4.2.3 칼만필터 계수 조절을 위한 비퍼지화기
• 5. 실험 및 결과
• 5.1 실험환경
• 5.2 자세각 측정 실험
• 6. 결 론
• 참 고 문 헌
• Abstract