摘要
针对行走距离估计问题,提出基于单加速度传感器的方法.将单个三轴加速度传感器固定在步行者小腿上,根据腿部状态(静止或运动)将读到的连续加速度值进行分步,并重积分运动状态下的加速度值获得行走距离.在原有阈值分步法基础上采用新的分步方法——自适应分步法进行分步计算,它根据步行者当前行走状态(步速、姿态等)对分步参数进行自适应调整.数据显示自适应分步受初始阈值影响小,具有较好鲁棒性,其平均分步误差为1步,平均距离误差在近匀速运动和变速运动情况下分别为15.18%和22.34%;而阈值分步的平均距离误差在近匀速运动和变速运动下则分别为31.08%和49.82%.实验表明:自适应分步法的结果更加准确且鲁棒性强.
A method based on single accelerometer was proposed to estimate walking distance. With a 3-x accelerometer attached to user＇s crus, the method divided continuous accelerations into steps in terms of the state of crus, stance or swing, and then integrated the accelerations during swing phase twice to get the walking distance. A new self-adaptive step detection algorithm, based on the threshold step detection algorithm, could adjust walking parameters in terms of the walker＇s current statue （e. g. rate, stance ere）. Experimental results indicated that the proposed algorithm was robust, and the effects of initial thresholds were limited. The average step error of the self-adaptive step detection algorithm was one pace. And the average walking distance errors of the self-adaptive step detection algorithm were 15.18% with constant pace and 22.34% with variable pace. While the average walking distance errors of the threshold step detection algorithm under the two situations were 31.08% and 49.82% respectively. Experimental results show that the self-adaptive step detection algorithm is more exact and robust.
出处
《浙江大学学报（工学版）》
EI
CAS
CSCD
北大核心
2010年第9期1681-1686,共6页
Journal of Zhejiang University：Engineering Science
基金
国家自然科学基金资助项目(60703040)
浙江省科技计划优先主题资助项目(2007C13019)
浙江省自然科学基金资助项目(Y107178)
关键词
加速度传感器
行走距离
普适计算
accelerometer
walking distance
pervasive computing
