Home health monitoring for elderly : Development of a health monitoring device for bathing

设计了一种无约束心电图（ECG）和心率监测器，用于记录浴缸中的ECG测量结果。该仪器由模拟部分和数字部分组成。在模拟部分，从放置在浴缸内的电极获得的心电图信号被放大。电极由不锈钢制成，并且包括吸收性海绵覆盖物的尺寸为104*74* 0.5mm。将工作电极放置在离长浴缸壁底部20 cm处，靠近浸入躯干的位置，将参比电极放置在离侧壁底部5 cm处，靠近脚趾位置。在系统的数字部分，心率是从每个ECG脉冲的R-R间期计算的，并且计算的数据被存储在随机存取存储器中。这些存储数据通过RS232 C接口传输到个人计算机并进行分析。浴缸心电图的频率特性随浴缸尺寸和水的电导率而变化。使用该装置，可以容易地和非侵入性地获得心率测量，但是心电图信号的解释需要非常小心。因此，该监测器可以用作家庭健康心率监测器。对于家庭健康护理，当其他家庭成员也使用监测设备时，需要进行个人识别。在床上，个人身份识别很容易，因为大多数受试者使用自己的床。在浴缸中，提出了使用ECG波形的个人识别[7，8]。首先获取一个心动周期的心电图波形，然后采用Daubechies-4小波变换对心电图数据进行小波压缩，并引入一个三层神经网络对家庭成员中的每个个体进行识别。

An unconstrained electrocardiograph (ECG) and heart rate monitor was designed to record ECG measurements in the bathtub. The apparatus consisted of analog and digital parts. In the analog part, electrocardiographic signals obtained from electrodes placed inside the bathtub were amplified. The electrodes were made of stainless steel and were 104*74*0.5 mm in size including the absorbent sponge covers. Working electrodes were placed 20 cm from the bottom of the long bathtub wall, close to the position of the immersed trunk, and a reference electrode was placed 5 cm from the bottom of the side wall close to the toe position. In the digital part of the system the heart rate was calculated from the R-R interval of each ECG pulse and the calculated data were stored in a random access memory. These store data were transferred to a personal computer through an RS 232C interface and analyzed. The heart rates obtained from the bathtub ECG agreed with those obtained from a body surface ECG.The frequency characteristics of the bathtub ECG varied with the bathtub dimensions and the electrical conductivity of the water. Using this apparatus, heart rate measurement can be obtained easily and noninvasively, but interpretation of the electrocardiographic signal requires great care. This monitor could thus be used as a home health heart rate monitor.For home health care, personal identification was required when other family members also used monitoring devices. In bed personal identification was easy because most subjects use their own bed. In the bathtub, personal identification using an ECG waveform was proposed [7,8]. The ECG waveform of one cardiac cycle was obtained and wavelet transform in which the mother wavelet was Daubechies-4 was applied to reduce the ECG data and a three-layr neural network was introduced to identify each subject among the family members.