신발 장착 관성 센서 기반 보행자 추측 항법의 센서 대역폭 및 샘플링 주파수에 따른 성능 분석

621;

공과대학 기계항공공학부 ;

University:서울대학교 대학원

关键词: 621;   

Others  :  http://s-space.snu.ac.kr/bitstream/10371/150681/1/000000154209.pdf

美国|英语

来源: Seoul National University Open Repository

【 摘 要 】

There is a high interest in technologies that provide various services based on personal location. These service is called personal navigation system (PNS), and it is used in a variety of fields such as providing location information during the mission of firefighters or soldiers, and providing customers’ location in shopping malls.Pedestrian dead reckoning (PDR), a personal navigation system, uses an inertial sensor to estimate the position of a pedestrian. PDR with foot-mounted IMU is based on inertial navigation system (INS). In order to reduce the estimation error caused by gyro and accelerometer bias and white noise, INS is combined with zero-velocity update (ZUPT). The ZUPT assumes that the velocity of the pedestrian shoe is zero in the stance phase. Error due to bias and white noise can be reduced by using ZUPT, but error still remain due to other factors.The movement of the shoe during walking is represented by various frequency components. In particular, the gyro and accelerometer signal change suddenly and contain high-frequency components in the heel-strike, when the impact occurs as the shoe touches the ground. The IMU determines the frequency range that can be measured by the bandwidth and the sampling rate. If the bandwidth or sampling rate is narrow, the frequency components exceeding boundary can not measured and position error occurs because the measurement accuracy is lowered. In this paper, I propose a frequency range to analyze the signal measured by the sensor during the gait cycle and to prevent the degradation of the accuracy of the signal. The validity of the proposed frequency range is verified by comparing the position estimation performance of the PDR system with various bandwidth and sampling rate.