Sleep Walking Prevention (SWP) Alarm System 1 Kuo-pao Yang, 2 Theresa Beaubouef, 3 Johnnie Morrison 1, 2, 3 Department of Computer Science and Industrial Technology, Southeastern Louisiana University, Hammond, LA 70402 ABSTRACT This paper discusses the development of a Sleep Walking Prevention (SWP) system, designed to prevent sleep walkers from suffering any dangerous consequences. It sounds an alarm if an incidence of sleep walking is detected. The SWP system also alerts a friend or family member in case the alarm alone was unsuccessful in awakening the sleep walker so that help can arrive before any harm befalls the sleepwalker or others. It does this by sending out a text message to the caregiver s cell phone. The sleep walking prevention alarm system makes use of passive infrared sensors for monitoring the status of a potential sleep walker. A system log is also in place to keep detailed records of sleep walking activities and patterns for reference purposes. This information can be very helpful to the consulting psychiatrist or doctor. Keywords: Sleep walking, alarm system, infrared sensor 1. INTRODUCTION Somnambulism, also more commonly known as sleep walking, is a sleep disorder that can lead to various dangerous consequences. There is no direct cure for this disorder that can assure the patient s health [1]. Furthermore, there is no substantial proof towards to the argument that a sleep walker should not be woken up. The only known problem with waking him up is that he may be slightly disoriented for a minute or two [2]. Since we cannot cure this disorder the question becomes: what is the most effective way of awaking sleep walking persons? Sometimes loud sounds are sufficient. Sometimes, however, it may require the intervention of another person to aid in their waking. The SWP alarm system discussed here provides a solution based on hardware and software co-design for waking sleep walkers. The main reason for waking a sleep walker is so they will not endanger themselves nor put others at risk. The activities of sleep walkers can be tracked by using Passive Infrared (PIR) sensors. After an occurrence of sleep walking has been detected and the alarm sounded, this system provides an integrated software/hardware application with equipment for sending an additional alert to a friend or family member. This failsafe helps to ensure the sleep walker is wakened and safely returned to bed. PIR sensors were selected for use in this application since they produce accurate zoning when detecting a moving object and when a specific event has occurred [3, 4, 5]. These PIR sensors are also compatible with most forms of hardware, allowing for few compatibility problems and promoting reusability if needed. The sensors are passive because that they do not emit their own infrared beam but only take in the infrared image of the target area. An alarm system should inform all needed parties whenever something has gone wrong [6], and the operator of the alarm should have a set of monitors that provides necessary information, thus preventing the anxiety of not being sure what is going on [7]. With the alarm system in place one does not have to worry about sleep walking events happening unexpectedly. The SWP alarm system is initialized by establishing how it should be operated between programmed start and end times. Because the alarm provides instant notification, the procedures needed to correct some problem situation can be initiated rather quickly. Delaying the sounding of an alarm until the system is certain that something has gone wrong is also very crucial to its success [8]. A simple double check system can be implemented in many ways to carry out this task [9]. One of the best ways to make sure about movement being tracked is to use multiple sensors [10]. The SWP system must have confirmation from two different perspectives in order to determine if a problem really has happened. PIR sensors can work in such a manner easily enough. 2. THE PROBLEM The motivation for this project arose from a personal need by one of authors his grandfather suffers from somnambulism. The goal of the project was to research hardware and software in order to design a system that could help out the family as much as possible. Through hardware we can solve most of this problem. The PIR sensors can easily detect movement within the grandfather s room. Then, the hardware board can take in the sensor readings and feed them into a computer. Software can be developed to handle the alert part of the system. The problem solution is not quite that straightforward, however, since we must account for several variations of this problem. In one case, the patient may get up to use the bathroom, causing the alarm to sound. This is one scenario that should be avoided due to its possible frequency. Additionally, the grandfather regularly falls asleep with the television on. Someone must then enter the grandfather s room to turn off the television after he has fallen asleep. It would be undesirable for the alarm to sound every time a caretaker must enter the room. To avoid sounding the alarm in these and possibly other unforeseeable situations, we 657
chose to design the system so that it incorporates not one, but two PIR sensors. These dual sensors can serve as a means for double checking movement patterns to ensure there is an actual occurrence of sleep walking before sounding the alarm. With two sensors in place as part of the system design, the remaining hardware and software components can be integrated to result in an appropriate solution. sensor if needed. The board, as seen in Figure 2, provides a versatile interface that connects a computer to the outside analog and digital signals. This board can be plugged into a computer via a serial port or a USB port. By making use of software libraries in programming languages such as Visual Basic, Visual C++, C#, Delphi, and Java, the board can sense and control the real world. 3. HARDWARE IMPLEMENTATION After considering all of the requirements from the sensors, the first task was to figure out where to place them for optimal effectiveness for the system. The two sensors must be placed in such a way that one double checks the other. Figure 1 shows the chosen placement of the PIR sensors and their coverage areas in blue and red lines. In this diagram, the blue line indicates the vision of sensor one, and the red line indicates the vision of sensor two. The two sensors were placed in the bedroom in such a way that allowed coverage of the two key areas: the bathroom door and door to the living room. One sensor monitors the bed and the pathway to the bathroom. The other sensor monitors the door leading to the living room. With this double sensor system in place, we can have the alarm sound only after both sensors have detected motion. This overcomes the issues discussed earlier, and the alarm is not sounded for a visit to the bathroom or the entrance of a caretaker. The sensors are positioned so that when sensor one is tripped then sensor two becomes armed. Only after sensor two has been armed will it sound the alarm once it is also tripped. This way if the grandfather only trips sensor one by going to the restroom then the alarm is not sounded. Moreover, if the caretaker trips sensor two by coming in to turn off the television at night the alarm is not sounded since it had not been armed by the prior tripping of sensor one. Fig 1: Placement of Two Sensors in Bedroom Fig 2: A GP-3XU Board A passive infrared (PIR) sensor is shown in Figure 3. This sensor detects the motion of the patient and communicates with the connected GP-3XU board. Fig 3: A Passive Infrared (PIR) Sensor 4. SOFTWARE IMPLMENTAION The main purpose of the software implementation in the system is to sound the alarm when needed. It is necessary to implement a customized program application, however, to indicate how and when the alarm shall sound. Moreover, by having a monitoring system in place we can track which sensors are currently active. A simple graphical interface for this monitoring system has been designed as part of the application software, all of which was developed in the C# programming language under the.net framework. Basic sensor information makes up the default software application screen as shown in Figure 4. The warning label is Inactive when the patient is on the bed. The warning label is Active once the patient is off the bed. Whenever sensor one has been tripped, a consequent tripping of sensor two will cause the alarm to sound, so the alarm sensor confirms by going into an active state. This system will issue an alarm if and only if both sensors are active. In case of a situation where the user would like to manually turn off the sensors, the checked box option has been provided to forcibly disable the alarm. This alarm system incorporates a GP-3XU board to interpret data from the sensors and to arm the second 658
This alarm system serves not only to wake the patient but also to alert the caretaker. Even a very loud alarm, however, may fail to wake a sleep walking patient. A caretaker, upon hearing the alarm, can easily wake the sleep walker and put the patient back to bed if he lives in the same building and can hear the alarm. But that is not always possible. The solution is to both sound the alarm and also to send a text message to the caretaker s phone. In this way, the caretaker will be alerted of sleep walking activities all the time. Fig 4: Information for Motion Sensors Figure 6 illustrates how the user can configure how the alarm sounds. He must enter the phone number and the phone carrier information in order for the caretaker to be alerted by a text message when an alarm has been issued. There are other options as well. For example, when the computer and the board sound are off but text message is on, the alarm will remain completely silent, but the caretaker will be personally alerted to the alarm situation. Instead of requiring the user having to set the alarm each night, the system can be run automatically. The user can easily set times for alarm activation and deactivation. Figure 5, for example, shows the alarm system starting at 10:00 PM and stopping at 5:00 AM. Once the user selects the desired times and clicks on the Set Time button, the system saves the timing data and automatically monitors the working environment. This simple enhancement to the system eliminates the many unnecessary trips to personally activate and deactivate the alarm. This alarm system can be also controlled from a remote location. The web interface is designed to access and modify settings for this desktop application remotely. A simple installation of IIS 7 and ASP.NET tools is all that is required for this feature. Once this software is installed on the computer at the remote location, the user can access all parts of the system remotely. Fig 5: Time Settings for Automatic Activation Fig 6: Alert Types The system also keeps a record of the movement and location of the patient between going to bed and waking up in a system log file. This log file may be kept for reference purposes, or it may be printed and given to the consulting psychiatrist or doctor, providing additional information about nightly patient activity. A sample system log file is shown in Figure 7, where the alarm system was set to run automatically from 10:00PM to 5:00AM each day. When the warning sensor (S1) is active and the alarm sensor (S2) is inactive at 12:02:10AM 03/03/2013, the system status indicates the patient is off the bed since the patient is on the bed at 10:00:00PM 03/02/2013. The caretaker comes to check the patient at 01:03:12AM 03/03/2013 and leaves at 01:03:50AM 03/03/2013. When both S1 and S2 are active at 02:27:30AM 03/03/2013, the system sounds an alarm and sends a text message since the patient is not only off the bed but also going out of the room. The caretaker clears out the alarm status and then puts the patient back 659
to the bed at 02:32:50AM 03/03/2013. The system log ends at 5:00:00AM 03/03/2013. The current patient record is appended to next day. Signal Processing and Communications (ICSPC), pp. 1459-1462, Nov. 2007. [3] Ha, K., Lee, K., and Lee, S., "Development of PIR sensor based indoor location detection system for smart home," 2006 SICE-ICASE International Joint Conference pp. 2162 2167, October 2006. [4] Bai, Y., Li, Z., and Xie, Z., "Enhancement of the complement of an embedded surveillance system with PIR sensors and ultrasonic sensors," 2010 IEEE 14th International Symposium on Consumer Electronics (ISCE), pp. 1-6, June 2010. [5] Zhang, Z., Gao, X., Biswas, J., Wu, J., "Moving targets detection and localization in passive infrared sensor networks," 10th International IEEE Conference on Information Fusion, pp. 1-6, July 2007. Fig 7: A System Log File 5. CONCLUSION This paper discussed the design and development of the Sleep Walking Prevention (SWP) system. Its goal is to prevent sleep walking from happening by sounding an audible alarm and also to alert another individual or caretaker to the occurrence of sleep walking. A system log keep in records for medical or reference purposes. The SWP alarm system has been successfully implemented in the home of a patient who suffers from somnambulism, providing greater peace of mind to his family and caregivers. This project serves as a prototype of what can be implemented as part of a smart home environment, which integrates a variety of electrical devices in a house. This particular alarm system can be expanded for other purposes such as security. The dynamic initial set up locations of this system will allow the motion sensors to be easily tripped by intruders. The security system can sound an alarm, send a text message, and record a variety of detected motion activities into the system log. Further expanded to include the installation of web cameras, the current web interface can be easily rewritten in order to access and remotely control the cameras via internet connections, providing an additional tool for the detection and later prosecution of intruders. REFERENCES [1] Bhattacharya, S. S., "Intelligent monitoring systems: smart room for patient's suffering from somnambulism," 2nd International IEEE-EMB Special Topic Conference on Micro technologies in Medicine & Biology, pp. 326-331, May 2002. [2] Atrey, P. K., Hossain, M.A., and El Saddik, A., "A Multimedia-Based System for Monitoring Sleepwalkers," IEEE International Conference on [6] Thunberg, A. and Osvalder, A., "What constitutes a well-designed alarm system?," 2007 IEEE 8th Human Factors and Power Plants and HPRCT 13th Annual Meeting, pp. 85-91, Aug. 2007. [7] Mattiasson, C., "The alarm system from the operator's perspective," International Conference on Human Interfaces in Control Rooms, Cockpits and Command Centres, pp. 217-221, Jun. 1999. [8] Xu, J. and Wang, J., "Averaged alarm delay and systematic design for alarm systems," 49th IEEE Conference on Decision and Control (CDC), pp. 6821-6826, Dec. 2010. [9] Abu Sajana, R., Subramanian, R., Kumar, P.V., Krishnan, S., Amrutur, B., Sebastian, J., Hegde, M., and Anand, S., "A low-complexity algorithm for intrusion detection in a PIR-based Wireless Sensor Network," 5th IEEE International Conference on Intelligent Sensors, Sensor Networks and Information Processing (ISSNIP), pp. 337-342, Dec. 2009. [10] Jisha, R.C., Ramesh, M.V., and Lekshmi, G.S., "Intruder tracking using wireless sensor network," 2010 IEEE International Conference on Computational Intelligence and Computing Research (ICCIC), pp. 1-5, Dec. 2010. AUTHOR PROFILES Kuo-pao Yang is an Associate Professor at Southeastern Louisiana University. He works in the Computer Science and Industrial Technology department. He received his B.S. degree in Computer Science at Tamkang University, Taipei, Taiwan, R.O.C. and earned his M.S. degree in Computer Science from Illinois Institute of Technology. He received Ph.D. degree in Computer science from Illinois Institute of Technology in 2003. His research 660
interests include Computer Architecture, Programming Languages, and Expert Systems. university s President s Award for Excellence in Research. Theresa Beaubouef has held positions with the Naval Research Laboratory and Xavier University, and is currently a professor at Southeastern Louisiana University, since earning a Ph.D. in computer science from Tulane University in 1994. Her research interests include the areas of uncertainty in databases, spatial databases, and data mining. In 2007 she was awarded her Johnnie Morrison received his B.S. degree in Computer Science at Southeastern Louisiana University in 2011. He has worked on numerous webpages for various clients ranging from gaming clans to small businesses. He currently resides in the small town of Lacombe, Louisiana. 661