本文利用智慧型手機內建之三軸加速感測器和地磁感測器來計算走路步數，最終目的為老年人之跌倒偵測，高齡族群使用智慧型手機接受度提高，使用智慧型手機內建之感測器進行偵測，可降低感測器大小及配戴多種感測器之不便，以手機內建的多種感測器得知使用者肢體動態計算步數與分析跌倒狀況是否發生，防止當跌倒發生時，無人在旁協助治療之情況。
本文計步器演算法在分析三軸加速度量值於平地行走可達９６％準確率，上樓梯可達９３％準確率，下樓梯可達８３％準確率，並可偵測使用者目前肢體動作以達更高之準確率，跌倒偵測改良MaaritKangas等多位學者演算法，加入肢體動作分析，增加準確度避免因較激烈日常活動而產生跌倒誤判，增加使用者困擾。 using smartphone acceptancehave increased. Usingsensorsbuilt-insmartphonesto detectcan reduce the size of the sensor and the inconvenience of wearing a variety of sensors.Using sensors built-in smartphone todynamic calculationsuserslimbsand analysis of the number of stepsto walkto detectionwhether a fall downsituation.To prevent fall down occur when no one beside help treat of the condition.
This pedometer algorithm for the analysis of three-axis acceleration values of up to 96% accuracy rate of walkingin a flatplace.On the stairs up to 93% accuracy rate, and down the stairs up to 83% accuracy rate. It can detect the user's current physical action to achieve a higher rate of accuracy.Improved Maarit Kangas’ fall detectionalgorithmadd limbs motion analysis.Increased accuracy to avoid the intense daily activities resulting Fall down false positivesto avoid adding users plague.
Relation:
2013全國計算機會議=2013 National Computer Symposium (NCS 2013)，7頁
