Fall detection and notification system design for elderly people

Falls are one of the most frequently observed events in elderly or unhealthy people. Falling can sometimes take place on the basis of internal and sometimes external factors, that is a threat to their health, especially for the elderly. Falling event for the elderly is considered as the reason for serious injuries, prolongation of recovery time, increased chance of the next fall and unfortunately, deaths in some cases. Falling has similar characteristics for all people, causing psychological trauma and various psychological problems. Moreover, the best action cannot be predicted, for the care and treatment of an elderly who has had a fall event, the fall event usually causes both the work loss of the elderly person, the work force of the person performing the care or treatment work and the need for all kinds of treatment materials. Most of the elderly do not care about falling cases however, some falling events cause lost of consciousness and mostly, they do not share the fall event that they have experienced except for any emergency with nobody. For this reason, it is very important for the elderly to follow the falling tendencies for providing them necessary care or delivery of treatment in time. Independent of the GSM network location but the geographical location, the fall location will be sent as message to person concerned is thought to be beneficial for the timely delivery of treatment to the falling elderly and for the scheduling of treatment and care needs. In this context, it is necessary to determine needs and system requirements. For the requirements of the design of the fall detection and information system, we interviewed a group of people with non-engineering backgrounds who have expertise in the field of health and also in engineering side we have been working with TESODEV on improving the design and testing. For the wearable device it is decided that the device can be worn on the belt or be suitable for carrying the waist, it should be lighter than 100 grams and has the smallest possible body structure. The current literature for sensors to be used in fall detection has been examined and no hybrid structure has been found that uses accelerometer, gyroscope and altimeter sensors together in wearable fall detection designs. For this reason, the hybrid structure mentioned in the design part is used. It has been noted that the electronic card is resistant to the military device shock test and that the biomedical device is designed in accordance with the EMI / EMC directives. Many of sensors investigated and compared each other. There are many accelerometer sensors in the market but our demands are limiting the sensors in the meaning of acceleration range, resolution and physical dimensions. Like the accelerometer sensors, there are many gyro meters in the market. Gyro meter candidates also eliminated to their gyro range, resolution and physical dimensions. Moreover altimeter sensors determination was too hard, because already being sensors resolution was very weak. For the best sensor option basically limits the resolution value down to 10cm. So low number of sensors find in the market and eliminated each of them regarding to physical dimensions, resolution and cycle time. As a result it was very attractive to use multi sensor in the same chip regarding to separate one meaning of small PCB design. This device is also embellished by GNSS sensor. All sensors and Bluetooth communication interface connected to a microcontroller that is collecting data and detecting the fall actions after that send GNSS location and person situation to cell phone via Bluetooth. Device sourced by a Lithium Polymer battery. Device software has developed by using C programming language. After drivers of sensors are established fall detection algorithm constructed and developed. As it is a wearable device its physical dimensions and weight design should be handled carefully. Because it impossible to wear this device if its volume, weight and body structure disturbs person. Also device wearing method is so important. According to decisions device can be worn with a belt. And its dimensions and also volume had to be small. For the conscious patients two more buttons assigned to emergency calls. It is obligated to pass some legislation for this device. Electronics and mechanical design was structured onto EMI/ EMC directives and environmental constrains. For the PCB and schematic design no more voltage higher than 4.0 volt and grounding was satisfied, switching regulator frequencies and resonance frequencies handled carefully for low electromagnetic interference. Many of the electromagnetic compatibility conditions determined tightly. But the environmental design obligations basically lean over the ingredient materials and human body interaction. Moreover for this design it is also worked on the operating and storage temperature grades and mechanical shock resistance. After functional verification of the device's electronic card, card and mechanical case assembling was conducted. Fall tests were carried out by taking necessary precautions and under the supervision of a medical doctor in the laboratory to test the fall algorithm which is the basic component of the fall detection and information system. Medical doctor informed volunteer people about the falling and laboratory precautions. Total of 1000 tests were administered, 40 people ranging in age from 21 to 32 and 25 tests per person. It is important to share there was no accident. The test results were evaluated using the confusion matrix method. Confusion matrix table contains actual and estimated conditions of an experiment. Confusion matrix table mainly divided into true and false rows and columns. After the table filled by the test results than sensitivity, accuracy and error rate values of test to expose systems statistical quality. According to the result of the experiment, it is estimated that 89.8% of the detection sensitivity, 83.9% of the accuracy, and 23.4% of the error rate are the detection sensitivity of hybrid construction. As a result, the high sensitivity of the device for falling detection indicates that the device is very promising. New tests should be established and algorithm should continue to be developed. Also older volunteers are looked for to carry fall detection and notification system. Fall detection and notification system will records motion data to Micro SD flash memory and it will be help understand to older people movement. To sum up the system is worked on for improving, developing the Smartphone software that will work with the device, reducing the power consumption even further, and making it a low cost, easily accessible commercial older health care product. By the meaning of this device people can take care of their older relatives without limiting the freedom and harming privacy of them.