Sleep Sound

Considered Concepts

 

This section will focus on the determination of the most suitable sensor to meet the design metrics. The remaining two components of the design are the base unit (information hub) and the remote receiver for the parents room. Although these components are important, their basic designs can easily be modeled after pre existing devices. The concept design phase was focused primarily on finding the best way to gather information about the infants' breathing patterns and send said information to the base unit.


Secondary research was first conducted to identify a range of sensors that could be used for the design. Pulse oximetry, respiratory impedance, CO2 detection, heart rate detection, movement monitors and thermal imaging were the found results from this research.


Pulse oximetry is a method for determining the oxygen saturation of blood. It functions by using three different LED's along with signal processing techniques. Two LED's - One which is Red (~660nm) and one which is Infrared(~910nm) are shown through a site with oxygenated blood from arteries. The blood will absorb differing amounts of light depending on it's oxygenation at the time. Concerns with using pulse oximetry at home were expressed by the staff of the Gritman Medical Sleep Center as the following. Many false alarms are due to this sensor merely because it must attach to the infants' foot and therefore is often pulled out of place or kicked off completely.


A respiratory impedance sensor detects the changing impedance of the thoractic cavity of the chest. This impedance is a function of volume of air in the chest. In general, cavity impedance will increase during inhalation and decrease during exhalation, leading to a cyclic pattern that is present during normal breathing. This principle could be used to determine when an infant stops breathing by employing a small high frequency (in the range of 50 kHz to 500 KHz) current between two surface electrodes to detect when the impedance stops changing, and thus breathing has stopped. The cost of the electrodes for this monitoring would be around $70. The invasiveness of this sensor raises concern since electrical current would be conducted directly through the infant.


There are two methods of CO2 detection, nondispersive infared (NDIR) and by chemical reaction. NDIR sensors measure the absorbtion of the characteristic wavelength of infared light as it passes through a light tube. As CO2 filters into the tube, the specific change in measured absorbtion signifies the presence of CO2. These sensors typically have sensitivities of 20-50 parts per million and usually cost between $100 and $1,000 in the (US). NIDR sensors are currently the most reliable method for detecting CO2

Chemical sensors utilize thin layers of polymer or heteropolysiloxane to detect CO2. Though smaller and size and power consumption, these sensors typically have a shorter life span than NDIR sensors. Unfortunately, each of these sensor types are limited to detect the presence of a gas and not the gas flow. If breathing stops there will be a delay period in which the sensor will continue to detect CO2. Therefore, CO2 sensors should not be relied upon to provide second-to-second updates on infant breathing patterns.


Heart rate monitors are typically used in two different applications; medical and recreational. Both of which utilize a different method of detecting the heart rate. For the purposes of the design, the heart rate of an infant can provide important information about breathing patterns.

The 12-lead electrocardiograph (ECG) monitor is the standard monitor used by hospitals around the world to measure a patient's heartbeat. These machines record from 12 different points of view the electrical activity that results when the muscle cells in the heart contract. The monitor then displays these readings as a waveform. Measuring from 12 different points of view makes these monitors very accurate because no single point on the heart will give the same reading as another. However, for a reliable, accurate monitor the price range varies from $1,500 to $5,000. A cheaper alternative also utilized in medical practice detects the heart beat by observing the change in impedance of the chest cavity similar to the respiratory impedance sensor. The same concern of conducting current through the infant is also present with this heart rate monitor.

The sensors used in recreation track the heart rate by monitoring the pulse. Heart rate monitors that use pulse meters measure the mechanical pulse of blood flow through the capillaries. These tend to be much less accurate than monitors using electrical activity. However, these monitors can be found in strapless forms (no chest-strap) and can be used in wristwatches and other attachment methods. Although generally cheap and effective, these monitors are designed for adults. This is a problem because a newborn infant's heart rate is much larger than that of an adult; 100-160 beats per minute for a newborn compared to 60-100 beats per minute for an adult. However, the same principles used in these devices can still be used in our device to monitor an infant's heart rate.


Three types of movement sensors were researched: motion sensors, strain gauges and weight scales. For the purposes of this design, the sensors would focus on the movement of the infants' chest as it breathes.


Motion Sensors use cameras to take successive picture of an area and then compare them to a reference image, pixel by pixel. If the number of pixels that are different exceeds some threshold, motion is occurring and the device will act on that motion. Processing is needed to distinguish the degree of change to determine whether motion is occurring because of false-positives associated with varying light, CCD dark currents and other minute factors. These devices are very cheap ranging from $20.00 and up and have a wide variety of sensitivity. These devices could be used to see the babies slight change in movement during inhaling and exhaling. The down side of these devices is the inability to see motion going on under sheets or blankets so the infants' motion could be blocked by its beading, giving a false positive to the alarm.


Strain Gauges use the principal of electrical resistance to detect whether there is deformation on the object the strain gauge is attached hence sensing a force. Resistance is equal to (L*P)/A. L is the length in the conductor, A is the cross-sectional area and P is the electrical resistivity of the material. The strain gauge is composed of an insulating flexible backing, which supports a metallic foil pattern. The strain gauge is attached to the object using any suitable means. When a force strong enough to cause any deformation in the material is supplied, it deforms the strain gauge as well. This deformation of the strain gauge alters the cross-sectional area and length of the conductor in it causing an overall change in resistance. This change in resistance can be measured by many devices by simply running a very small current through it to determine the resistivity of the device. problem that I can foresee would be the deterioration and possible fault of the strain gauge over time. This measuring device could be used around the chest of the infant, or preferably on the surface where the baby is sleeping. The price of this type of sensor approximately $5.00.


Weight Scales simply determine the force being applied by a certain object. When the baby inhales there will be a change in force on the mattress due to the acceleration of the chest cavity up and down. The forces can easily be measured by a variety of different weight scales having two main categories, digital and analog. Sub categories include: spring, hydraulic or pneumatic and balance. The most appealing weight scale, because of size, price, accuracy (1/100 grams) and ease of use, would be an electronic mass balance scale. The price of the needed balance off the shelf would be $60.00 and up, depending on the exact quality and accuracy wanted.


Thermal imaging equipment is sensitive to only infared light which is emitted from any source of heat. The thermal imaging camera could be used to observe the motion of chest movements from the sleeping infant similar to the motion sensor. Although thermal imaging would reduce the issue of not being able to see through sheets, camera prices starting at $2,000 ultimately keep this sensor option from being cost effective.


The second phase of sensor determination was to conduct primary research on as many of the aforementioned sensor options as possible. Each sensor was to be tested for sensitivity when subjected to the expected conditions of monitor operation.


A new sensor concept was developed by the team around this time and was included in the second phase. This concept is based on using a chest strap integrated with a zig-zag pattern of wire which can act as a variable inductor. As the chest expands, the distance between the loops of wire will increase, resulting in a decrease of inductance. This variable inductance strap can be connected as a frequency determinant component within an oscillator circuit. A change in inductance (L1) will translate to a change in generated frequency as seen in the following characteristic equation for a standard Colpitts oscillator circuit.


Eqn 1.


This variable frequency signal can then be amplified and transmitted to a receiver on the base unit. Once the received signal is converted to a digital signal, it can be read as clock pulses by a micro controller. Infant breathing will be confirmed by the varying number of clock pulses per unit of time of the input signal to the micro controller. This sinusoidal change represents the expansion and contraction of the infants' chest. The micro controller can be programmed to sound an alarm in the continued absence of change in the number of clock pulses per unit time.


A test inductor strap prototype was built (see Fig.1). 74 loops of 32 AWG coated wire were sewn into a one inch wide piece of elastic. The average loop length (L) was 5/8 inches. The average distance between each loop (X1) was 1/32 inches and the total distance of the loop section (X2) was 13/4 inches.


Figure 1. Variable Inductance Strap Prototype.

(Actual number of loops not shown)


To test the sensor prototype, the 32 AWG wire leads extending from the elastic were sanded at the tips to remove the coating. They were then connected to a Fluke PM6306 LCR meter. The frequency of the meter was set to one Mega Hertz to simulate the expected high frequency of operation. Figure 2. shows the change in inductance of the strap between no stretch and full stretch.


 

Figure 2.


This 17.8% change in inductance was decidedly large enough to cause a significant change in the generated frequency of an oscillator circuit.

 


To test the sensitivity of the strain gauge option, a metal foil strain gauge was glued directly to the center of a Thermarest air mattress. Two 32AWG wires were soldered to the plus and minus contacts of the strain gauge. The wires were connected to equipment leads which were then connected to an HP 34401A Digital Multi Meter. Approximately three pounds of pressure was applied by hand to either end of the mattress to simulate even pressure distribution. The change in resistance as read from the multi meter was 0.02 Ohms.

The result from this experiment indicated that a strain gauge would not be an accurate sensor for this design. This conclusion is based on the fact that the final design would be sensing the change in mattress tension due to chest expansion and contraction of an infant. This change in pressure would be much less than three pounds.



The third tested option was the weight scale. For this design, a grid of small digital scales could be integrated into a sleeping pad. Each scale would contribute to a single output voltage. Ideally, this voltage will be in continual flux as a result of chest movement from the infant. This output signal can be hardwired to a micro controller within the base unit that can be programmed to look for a changing voltage. An alarm would then sound in the continued absence of a change.

To determine the relative sensitivity of a pre existing digital scale, two team members traveled to the local recycling center. One lied on the scale while the other observed the change in weight due to breathing. The results from this experiment show that such a scale exists that will show a 0.1 lb change due to relaxed breathing of an adult. 0.1 pounds is approximately 45.36 grams. Given that weight scale sensitivities can be around 1/100 grams, this sensor option was decidedly sensitive enough for our purposes.

Site

Changes
Index
Search

 

User

 

Log In
Register

 
 

Last Modified 2/5/07 4:02 AM