PRE2020 3 Group4

Smart lock for AEDs

Group members

Name	Student ID	Department	Email address
Mark den Besten	1022231	Automotive Technology	m.d.besten@student.tue.nl
Yann van Eijk	1454447	Mechanical Engineering	j.h.m.v.eijk@student.tue.nl
Ilse Schuckman	1239641	Electrical Engineering	i.schuckman@student.tue.nl
Rik Schutte	1005841	Psychology & Technology	r.j.r.schutte@student.tue.nl
Roel Wijnands	1235389	Electrical Engineering	r.r.c.wijnands@student.tue.nl

Problem statement

Every minute is crucial when providing aid to someone who has an out-of-hospital cardiac arrest (OHCA). The highest survival rates are achieved when the patients receive defibrillation as fast as possible (Jafri, 2021). The time it takes for an ambulance to arrive, however, is usually too long to ensure good survival odds for the patient (Solanki, 2017). To reduce the average time between an arrest and defibrillation, automatic external defibrillators (AEDs) are placed in public areas. These AEDs are placed in cabinets that are often locked, so they are protected from acts like vandalism. The people that can access them are usually the owner(s), their employees, and voluntary first responders. Voluntary first responders (in dutch: burgerhulpverleners) are people who have followed CPR and AED courses and are registered in the national database of the HartslagNu foundation. When an OHCA occurs, nearby voluntary first responders are alerted via either text message, or via the app, to the location of the closest AED and the place of the incident (Burgerhulpverlening 2021). Within the original alert message, the pincode of the nearby AED cabinet is provided, this only happens when alerted via the application. This means that people who only get notified via SMS do not get access to an AED. Moreover, because the situation is both stressful and extremely time-sensitive, the time and effort it takes to read and input a code correctly is unwanted. This project aims to remove this process and thereby reduce the average time between an arrest and defibrillation. In order to do so, the users have to be identified and contacted in order to hear their opinions on the projected solution. The technical aspect of the solution also has to be worked out in order to provide a complete conclusion at the end.

How it currently works

HartslagNu is the paging system that calls up voluntary first responders via a message on the mobile phone to go to the scene in case of possible resuscitation. In many cases, voluntary first responders are able to respond quicker to an emergency situation than the emergency services, because voluntary first responders are more wide-spread than the emergency services. This means that roughly every neighborhood in the Netherlands has at least one voluntary first responder who can be at the emergency location in a relatively short amount of time. It is important to be able to notify the emergency responders as quickly as possible, to keep this time advantage.

The current state of the HartslagNu system in the Netherlands is well seen from a technological standpoint. The user gets to enter their home and/or work address online or within the HartslagNu app, and based on that information the user can be contacted via SMS in the case of an emergency. In addition to this basic feature, the application also allows the user to turn on their location and let the application track their GPS signal (HartslagNu, 2021). When the user is near an emergency location, the application sends them a notification containing the location of a possible resuscitation. It is important to make the distinction between the SMS-based function and the application-based function, these two can work separately from each other and in addition to each other. HartslagNu thus assumes that everyone has a mobile phone, but not everyone (always) has an internet connection on their phone. To use the application the user has to have an internet connection and allow for the app to send push notifications. Using the application also has the advantage that it provides the pincode to get access to an AED.

Voluntary first responders are trained to start up the resuscitation, preferably using an AED. These AEDs are publicly available all over the Netherlands and are usually protected by a pincode in a so-called AED cabinet. The voluntary first responders are able to get access to the AED by retrieving the pincode, which can only be done with the application, and not via SMS. This, however, takes time and sometimes does not work at all. The time loss can result in the loss of lives or the cause of brain damage at the victims' end. This is where an adaption to the current system can be made and what this project is all about. Once the emergency services have arrived, the voluntary first responders return the AED to its cabinet where, at a later time, the AED can be routinely checked for any malfunction. The owner of the AED also gets notified that their AED has been used, so they can contact HartslagNu or any other local heart foundation to get the AED checked.

Objectives

The objective of this project is to optimize the current system of retrieving an AED for voluntary first responders. The way this is envisioned is to allow the cabinets to be opened with less user interaction. In addition, the new solution should not discriminate between the different forms of notification receiving (via the app or SMS). The different ideas can be found in the section 'design options'. At the end of this project, it should be clear which option is the best, and a demonstration of this option should be given.

State of the Art

[1] Cho, H., Ji, J., Chen, Z., Park, H., Lee, W. (2015). Measuring a Distance between Things with Improved Accuracy. Procedia Computer Science, 52, 1083-1088. Retrieved from: https://doi.org/10.1016/j.procs.2015.05.119

A Bluetooth based design for measuring distance between objects using Bluetooth Low-Energy (BLE). The results showed that it is possible to attain an accuracy of 90% in a radius of 1.5 meter.

[2] Drawil, N. M., Amar, H. M., Basir, O. A. (2013). GPS Localization Accuracy Classification: A Context-Based Approach. IEEE Transactions on Intelligent Transportation Systems, 14(1), 262-273. Retrieved from: https://ieeexplore.ieee.org/document/6295661

A study that compares the pros and cons of different GPS localization methods.

[3] Eisenberg, M. S., Horwood, B. T., Cummins, R. O., Reynolds-Haertle, R., & Hearne, T. R. (1990). Cardiac arrest and resuscitation: A tale of 29 cities. Annals of Emergency Medicine, 19(2), 179-186. doi:10.1016/s0196-0644(05)81805-0

Old article at the base of fast-response knowledge investigating the survival rates in 29 different cities depending on the rate and speed of response including which method was used by whom. So option like direct CPR or some time passing before someone with AED arrives. Concludes faster CPR is key to survival but still defibrillation is needed.

[4] Everything you need to know about smart locks. (2021, February 01). Retrieved February 12, 2021, from https://www.safety.com/smart-locks/guide/

Article about safety surrounding smartlocks. This article discussed the opportunities as well as the pros and cons.

[5] H. Oguma, N. Nobata, K. Nawa, T. Mizota and M. Shinagawa, "Passive keyless entry system for long term operation," 2011 IEEE International Symposium on a World of Wireless, Mobile and Multimedia Networks, Lucca, 2011, pp. 1-3, doi: 10.1109/WoWMoM.2011.5986125.

Talks about keyless entry systems of vehicles. Makes use of synchronized keys.

[6] Jeong, H. J., Lee, W., Lim, J., & Hyun, W. (2015). Utilizing a bluetooth remote lock system for a smartphone. Pervasive and Mobile Computing, 24, 150-165. doi:10.1016/j.pmcj.2015.07.010

Talks about a Bluetooth lock and how this should work with an smartphone app. However the phone first has to pair with the smartphone so in this state does not work for multiple unknown devices.

[7] Koster, R. W. (2013). Modern BLS, dispatch and AED concepts. Best practice & research. Clinical anaesthesiology, 27(3), 327–334. Retrieved from: https://doi.org/10.1016/j.bpa.2013.07.005

A study that presents the current strategies for OHCAs. It concludes that AEDs are of critical importance when providing aid, especially before the arrival of an ambulance.

[8] Mair, N., Mahmoud, Q. H. (2012). A collaborative Bluetooth-based approach to localization of mobile devices. 8th International Conference on Collaborative Computing: Networking, Applications and Worksharing (CollaborateCom), 363-371. Retrieved from: https://eudl.eu/doi/10.4108/icst.collaboratecom.2012.250437

Another design based on Bluetooth to determine position. It uses the combined Bluetooth connections of multiple devices. The results also show that, while it is slower than GPS, Bluetooth localization is faster than Wi-Fi or cellular methods.

[9] R. Karani et al., "Implementation and design issues for using Bluetooth low energy in passive keyless entry systems," 2016 IEEE Annual India Conference (INDICON), Bangalore, 2016, pp. 1-6, doi: 10.1109/INDICON.2016.7838978.

Another Bluetooth lock. This one with a focus on low-power Bluetooth. It uses a whitelist for checking which devices may unlock. The white list is controlled by an admin terminal.

[10] Smith, C. M., Lim Choi Keung, S. N., Khan, M. O., Arvanitis, T. N., Fothergill, R., Hartley-Sharpe, C., . . . Perkins, G. D. (2017). Barriers and facilitators to public access defibrillation in out-of-hospital cardiac arrest: A systematic review. European Heart Journal - Quality of Care and Clinical Outcomes, 3(4), 264-273. doi:10.1093/ehjqcco/qcx023

This article explores the different areas which affect the response in assisting in defibrillation and partaking in accessing defibrillators. The study concludes that there is very little evidence in this field and more research should be done into these different factors.

[11] Smith, C. M., Wilson, M. H., & Perkins, G. D. (2018). Reply to: Letter by Derkenne et al. regarding the Article, ‘the use of trained volunteers in the response TO Out-of-hospital cardiac Arrest – The GoodSAM experience.’. Resuscitation, 125, E4. doi:10.1016/j.resuscitation.2018.02.002

The investigation looking at GoodSAM in England, a system somewhat similar to HartslagNu in the Netherlands investigating the acceptance of the system and willingness of people to actually partake in resuscitation including accessing an AED.

[12] Smith, C. M., Wilson, M. H., Ghorbangholi, A., Hartley-Sharpe, C., Gwinnutt, C., Dicker, B., & Perkins, G. D. (2017). The use of trained volunteers in the response to out-of-hospital cardiac arrest – the goodsam experience. Resuscitation, 121, 123-126. doi:10.1016/j.resuscitation.2017.10.020

Article exploring the effectiveness of more participation of ‘bystanders’ in performing in a resuscitation. This participation is achieved via the GoodSAM application which alerts volunteers nearby of a resuscitation request.

[13] Stieglis, R., Zijlstra, J. A., Riedijk, F., Smeekes, M., van der Worp, W. E., & Koster, R. W. (2020). AED and text message responders density in residential areas for rapid response in out-of-hospital cardiac arrest. Resuscitation, 150, 170–177. Retrieved from: https://doi.org/10.1016/j.resuscitation.2020.01.031

A study that shows the effects of increasing the density of AEDs and Text Message responders in residential areas. It concludes that the text message system is most effective when there are at least two AEDs and ten TM-responders per square kilometer. The number of patients who were defibrillated in under 6 minutes doubled when these two conditions were met.

[14] Waar zijn nog AED's Nodig? (2020, September 10). Retrieved February 12, 2021, from https://hartslagnu.nl/aed/waar-in-nederland-zijn-nog-aeds-nodig/

Site with an interactive map showcasing the current spaces where there isn’t an AED in the near vicinity. Based on a history of resuscitation attempts. The rule of thumb here is that there should always be an AED within a 500m radius.

Current cabinet designs

There are a plethora of different cabinets currently in use. For this project, the type of cabinets that are of interest are those in outdoor public places that require a pincode to open. These cabinets do share some functionality, namely they are all ventilated, illuminated and have an alarm. The remaining functionality differences are summarized in the table below.

Manufacturer	Pyrescom	GT Medicare B.V.
Cabinet	AIVIA	SmartCase	SixCase
Type	210	SC1240	SC1340	SC1340P	SC1340M	SC1340PM	SC1440	SC1440P	SC1440M	SC1440PM
Pin	Touchscreen	Rotary knob	Touchscreen	Touchscreen	Touchscreen	Touchscreen	Touchscreen	Touchscreen	Touchscreen	Touchscreen
Mechanical lock	-	-	-	-	Yes	Yes	-	-	Yes	Yes
Material	PC/ABS	SST	Aluminium	Aluminium	Aluminium	Aluminium	SST	SST	SST	SST
Temperature	-20° C	-20° C	-15° C	-25° C	-15° C	-25° C	-15° C	-25° C	-15° C	-25° C

All three cabinet designs open in a different way. The AIVIA has two horizontally oriented hinges at the bottom left and bottom right corners. The SmartCase is opened like any general locker, with two hinges oriented vertically at the lower and upper right corners. The SixCase has a compartment at the bottom that slides down when the right pincode is entered.

Users

There are four main users that have to interact with the AED cabinet.

Firstly, the first responders to an emergency, the so-called voluntary first responders. A voluntary first responder gets a notification on their cellphone and has to, as quickly as possible, get an AED to the person that is having an emergency. An AED should get to the victim within 6 minutes (bhvnederland, 2018), every second counts. A voluntary first responder should therefore have quick and easy access to the AED. Quickly, because lives are at stake. Easily, because in the heat of the moment a voluntary first responder might not be able to think coherently. The voluntary first responder should not have to think about getting a door unlocked.

Secondly, the owners and maintainers of the AED and its cabinet. Placing an AED outside is a voluntary decision. Therefore, when an AED is damaged or lost it is a loss for the owner. An AED should be sufficiently protected from malicious users. At this point in time, pincode is commonly used for this (HartslagNu, 2021). The maintainers also would like to keep their costs and maintenance low. So it should not cost a lot extra. Additionally, these owners already have a locker that functions well. They do not want to buy a new one before the old one is written off.

Additionally, there is the owner of the app and database of all the AEDs in the Netherlands, HarstslagNu. The improvements on the interaction between the AED cabinets and their app should be easy to implement. There should also be some new entry for their database, namely of what type of cabinet it is. Especially in the beginning when not all cabinets have been upgraded. The app should know what kind of locking mechanism the cabinet uses.

Lastly, the manufacturers of the AED cabinets also should be taken into account. If a solution is found, the manufacturers have to agree to make it adaptable to their product. If they do not want to cooperate, the solution becomes infeasible. It is therefore important to keep the cost of manufacturing in mind, as well as the compatibility with the current AED cabinet designs since the manufacturers will determine their cooperation based on these factors.

To further detail the user requirements, the following user scenarios were created.

Dirk

Dirk is a 32-year-old man living in Helmond. He is a licensed aid worker and is part of the NIBHV network (the national aid worker network). On a given Saturday, he was taking a walk in the park when he got a message on his phone. A notification from HartslagNu, that someone in the area is suffering a heart attack. He opens the HartslagNu app and accepts the call for action. The app shows him the route to the afflicted person as well as the nearest AED for him to take with him. The app also gives him the code to open the cabinet that houses the AED that he is assigned to. He hurries to the AED as quickly as possible and gets there in 3 minutes. The route shows that it is another 2 minutes to get to the victim. As he knows from his training, the resuscitation should begin within 6 minutes as the victim might end up with permanent brain damage otherwise. This gives him 20 seconds at most to get the AED, since he also needs to set it up it at the patient.

However, when he arrives at the location of the AED, there is a problem. The cabinet is locked and requires a pincode to open. He remembers that the app should have a code for him to unlock the cabinet, so he grabs his phone and opens the app. After looking at the app, he cannot find the code. In the meantime, another responder shows up and starts looking for the code too. After 40 seconds, they finally find the code and unlock the cabinet. After grabbing the AED, they hurry towards the victim together. They arrive at the scene 7 minutes after the call was made. When they want to start helping the victim, they hear sirens coming closer. The ambulance shows up and takes over the care for the victim. So, in the end, Dirk and the other responder could not do anything and go back to return the AED.

Mandy

Mandy is a 43-year-old woman living in Rotterdam. She owns a small store in the area and is also a certified company aid worker. She is part of the HartslagNu network and was looking through the AEDs in the area when she noticed a lack of AEDs in the area around her store. Because of this, she had decided to invest in an AED at the shop. She intends for the AED to be placed just outside the shop so it can be utilized for general emergencies. One of her worries is that certain people, say a drunk man after a night out, would steal the AED, so she decides to get a cabinet with a lock. The Dutch heart foundation advises her to get one with a pincode as it is more resilient against weather conditions.

To fund this endeavor, she approaches the municipality. She poses her problem and requests some financial aid to make this happen, as it would benefit the entire neighborhood. The council in Rotterdam agrees and wants to cover half the costs of the AED. With half the costs dealt with, Mandy can afford to get the AED installed. To make it easier on herself, she arranges one with self-testing capabilities. This causes the AED to send a message to her over Wi-Fi when something is wrong. Apart from that, maintenance only involves changing out batteries which is done by the company she bought the AED from.

One afternoon, someone rushed over to her AED, and took it out of the cabinet. After about 20 minutes it was returned without issue. That same night, she saw a story on the local news about some being reanimated using an AED. This proved to Mandy that having an AED installed is definitely worth it.

User requirements

Brain damage can occur within 6 minutes of cardiac arrest (bhvnederland, 2018). Considering the average response time for voluntary first responders is between 5.5 and 7.5 minutes (Dagblad van het Noorden, 2019), every second counts. As such, the cabinet housing the AED should open in no less than 15 seconds to ensure quick access.

To make it easy to open the cabinet, it should not require any extra operations from the user. So, during an emergency, opening the app should not be required to unlock the cabinet.

The users also require the AED to still be in the cabinet. This involves creating a tamper-resilient cabinet with a robust locking mechanism that keeps unwanted people out while maintaining ease of use for the authorized users. Since it is connected to an online system, this framework also requires tight security such that people cannot falsely authorize themselves and gain access to the AED.

In case of the system malfunctioning, an alternative way of opening the cabinet containing the AED needs to be present as well. This is also for users that do not wish to use the app, whether that is because of a lack of technological understanding or a desire to protect all their data. As such, the new way of opening the box needs to be an addition rather than a replacement.

When there is no active emergency, the lock should not open, even when authorized users with the app walk past. This excludes maintenance. This is to prevent the cabinet from randomly opening when users walk past it unknowingly.

When the AED has been removed from the cabinet, it should be easy for the voluntary first responder to return the AED and to close and lock the cabinet. The cabinet should not be locked when the AED is not present in it. This prevents the cabinet from locking itself prematurely when the AED has not yet been returned. This way the owner can perform the mandatory check-up of the AED.

For the owners, it is vital that the system is easy to be installed and requires very little maintenance. High levels of upkeep would be bothersome for the owners and would also discourage people from buying an AED in the first place. While an initial installation has to be done, including the necessary wiring, further upkeep must be kept at a minimum. Since the AED needs to be tested once a year, the lock should also be tested once a year at a minimum.

The price of this newly developed technology should be reduced to encourage buyers to implement this device on their AEDs. The cost for upkeep should also not be increased by more than 3 percent of the existing cost. This is due to the high cost of upkeep as is (Medisol), and people might get deterred from installing an AED if the cost becomes even higher.

Design options

Method	Brief description	Functionality needed for the cabinet*	Additional actions needed to function**	Reliability	Maintenance/Risk
Bluetooth	Anyone with the app can connect to the cabinet via Bluetooth. If a connection is established, the cabinet will open.	-Bluetooth	None	High	All electronics can be equipped inside the cabinet, reducing risk of damage.
GPS	GPS is used to determine the user's position relative to the AED. If they are close, the cabinet receives a signal and opens.	-Internet or SMS module	None	High	The current implementation already makes use of GPS. All electronics can be equipped inside the cabinet, reducing risk of damage.
Wi-Fi Direct	When the user receives an alert via the app, their phone will start searching for devices to pair with via Wi-Fi Direct. The cabinet will open when a connection has been established.	-Wi-Fi -Other	The user has to finalize the pairing between the smartphone and the cabinet in some way. Possible options: Bluetooth, NFC, button press.	High, but lower than Bluetooth or GPS due to the additional connection/method needed.	Most/all electronics can be equipped inside the cabinet, reducing risk of damage.
SMS	When HartslagNu sends out messages to voluntary first responders in the area, an SMS is also sent to nearby cabinets. Upon receiving this SMS, the cabinets will be unlocked.	-SMS module	None	High	The cabinets will be unlocked early, increasing risk of theft. All electronics can be equipped inside the cabinet, reducing risk of damage.
Biometric	Anyone who downloads the app must submit fingerprints/facial picture to a database. The cabinet will open to anyone that matches.	-Internet -Camera or fingerprint scanner	The user has to pass a biometric lock.	Facial recognition is fallible, especially for people with a darker ethnicity. (Klare, 2012)	A database has to be updated regularly. The exteriors of all cabinets have to be equipped with electronics, increasing risk of damage.
NFC	The user transmits the code received through the app via NFC. When installed, each cabinet is given a unique code.	-NFC-tag	The user has to hold their phone to a scanner.	The high stress situation may cause the user's hands to tremble too much.	All cabinets need to be equipped with an NFC tag on the exterior, increasing risk of damage.
QR-code	The cabinet has a camera that can read a QR-code generated by the app. Each cabinet that is installed has its own unique code that does not change.	-Camera	The user has to hold their phone to a camera.	May not be reliable in poor lighting conditions. Due to the high stress situation, the user's hands may tremble too much.	Electronics vulnerable to damage have to be installed on the exterior of all cabinets, increasing risk of damage.

*Assuming the user has a modern smartphone with up-to-date functionality.

**Apart from the user having the app and being in the vicinity.

User survey

In order to test the main plan of opening the AED cabinets from a remote location, the opinions of the users were needed. To reach out to the users, some different methods were tested. The first idea was to simply reach out to HartslagNu and ask if they were willing to spread the survey via their social media accounts. However, HartslagNu was not willing to cooperate since they also use their social media accounts to spread surveys for themselves, and they did not want any confusion under their users by spreading a survey from a third party. This was a bit of a disappointment, and after trying to reach out to local foundations that are in close contact with HartslagNu and getting the same response, we tried to reach out to the users directly. More can be read on this in te section 'contact with the users'.

Voluntary first responders

Q1: How long have you been a voluntary first responder?

Answers from the voluntary first responders on question 1

Q2: How often have you retrieved an AED?

Answers from the voluntary first responders on question 2

Q3: How are you notified by HartslagNu of a possible emergency?

Answers from the voluntary first responders on question 3

Q4: Did you ever have to use an AED where a pincode was needed to access it?

Answers from the voluntary first responders on question 4

Q5: If yes, did you encounter any difficulties while accessing the AED?

Only two responses stating: not applicable.

Q6: Is there anything you would change about the current methods used to access locked AEDs?

One response stating: Our AED cabinets are not locked.

Q7: If you were called to retrieve an AED, do you think the amount of time and/or stress would be significantly reduced if a locked cabinet were unlocked remotely so you no longer have to enter a pincode?

Answers from the voluntary first responders on question 7

AED owners

Q1: For how long have you owned an AED?

Answers from the AED owners on question 1

Q2: Is your AED placed in a cabinet that requires a pincode to open?

Answers from the AED owners on question 2

Q3: Why did or did you not choose this option?

Three responses along the lines of: 'It was the advice of the local heart foundation.'

Two responses along the lines of: 'Because every second counts.'

One explanation: 'From the 72 AED cabinets we manage, only one has been vandalized and one has disappeared over a span of 14 years. Putting the cabinet behind a lock is in our eyes unnecessary.'

Q4: Has your cabinet and/or AED ever been damaged due to (attempted) vandalism or theft?

Answers from the AED owners on question 4

Q5: In regards to the security of your AED, how would you feel if your cabinet was unlocked remotely at the same time as the voluntary first responders are notified of an incident?

One response: 'I am fine with it.'

Survey conclusion

Voluntary first responders

The voluntary first responders reached with the survey seem to be very experienced, with 66.7% of them being a voluntary first responder for 6 years or more. However, it seems that they (luckily) do not have to come into action that much since question 2 indicates that they retrieved an AED only 0 up to 5 times. In hindsight, the answers for this question should be provided differently, since now it is unclear whether they have retrieved an AED even once. This could be improved by adding the option '0 times' as a stand-alone answer. It also seems that not all voluntary first responders are notified by HartslagNu, with 33.3% saying they are not notified, 16.7% saying they no longer are notified, and only 50% being notified either via SMS or via SMS and the app. This can be due to the confusion of the abbreviation 'BHV', in Dutch this commonly stands for 'bedrijfshulpverlener' (emergency response worker), but in this case, it stands for 'burgerhulpverlener' (voluntary first responders). Emergency response workers are employees that are trained to provide first aid in case of an accident on at work. They are not signed up for the HartslagNu paging system. A possible explanation for the number of respondents not signed up for notifications from HartslagNu could be due to this confusion. The answers to question 4 also indicate that the ideal target audience was unfortunately not reached. Questions 5 and 6 are hard to discuss due to the lack of answers to those questions. The main gain from the survey under voluntary first responders is the answers to question 7. Here it can be seen that 83.3% of the respondents are in favor of opening locked cabinets from a remote location, with only 16.7% of the respondents not seeing any gain at all. This means there is no reason to cut the project short by saying that there is no need for it from the side of voluntary first responders. This conclusion is very limited, and only based on a small number of respondents, so it should be taken with a grain of salt.

AED Owners

Most of our AED owner respondents seem to be in contact with Rooi Hartsafe. This is an indication that the answers to this survey are heavily based on the AED owners in the old municipality of Sint-Oedenrode. This, of course, is not ideal as this is not representative of the whole country of the Netherlands. Thus the main conclusion that can be made from this survey is that there are not enough respondents, and the respondents are not wide-spread. This does not mean the survey was useless. The survey did provide the insight that not all local heart foundations see the usefulness of locking the AED. This means that the solution is mainly based on bigger towns and cities, or other busy areas. the heart foundations in several cities were contacted but sadly a response was not forthcoming. When continuing to contact the users, it is important that the users in denser areas are reached. Due to COVID-19 measures and the unwillingness of the people in these areas to cooperate, their opinions have not yet been heard in the survey. It is therefore still unclear what their view on the topic is, based on the survey alone.

General

All in all the survey did not reach the target group that needed to be reached. This does not mean the survey was useless, but to be able to make some better-substantiated conclusions, the survey should be filled out by more respondents from all over the Netherlands. The explanation on question 3 for the AED owners can be used to substantiate the risk analysis. The survey would be more effective if a cooperation with HartslagNu was possible.

Contact with the users

Yann

Spreading of the survey

To collect as much data as possible we tried to spread the above-mentioned survey to as many people as possible. I myself started by contacting as many people I know via WhatsApp. Via this medium, I must have reached around 500-600 people, mostly via group-chats. This was, however, not very effective since the people who received the message were a random group of the population. This meant that the chance that one of them was either a voluntary first responder or an owner of an AED was very small. After waiting for 24 hours the number of answers on the survey did not increase, so I concluded this had been ineffective.

Afterward, I tried to specifically reach out to AED owners in the region in an attempt to increase the chances of getting answers to the survey questions. For this, I used the sites of the foundations HartVeilig Den Bosch and Rooi Hartsafe. They offer education on resuscitation in the municipalities of s'-Hertogenbosch and Sint-Oedenrode (Meijerijstad), respectively. They also manage the spread of AEDs in their own municipality. A list of where AEDs can be found has been posted on their site, using this list, I contacted around 15 companies/individuals via either e-mail and/or phone. I also contacted the foundations to ask for their own opinion and if they were willing to spread the survey. A list of the people contacted and their responses can be found below. I tried to contact small businesses and people I knew via the local football club, etc. I also called two people who managed AEDs in private locations, Mr. Van Dongen and Mrs. Smits. With respect to their privacy, their contacts will not be included, as requested by themselves. I will include their answers to the basic questions in a later paragraph.

People contacted and reactions

Contact	Reaction (y/n)	Notes
HartslagNu	Yes	They were not willing to spread the survey, I did get answers to the questions that I asked in the mail directly. They will be processed in a later section.
Stichting HartVeilig Den Bosch	No	I never got any reaction to the question of a possible interview.
Stichting Rooi Hartsafe	Yes	Due to interconnection with HartslagNu, they are not allowed to spread the survey. They did fill out the survey. They were not available for an interview.
Goessens podologie	Yes	They replied mentioning that they do not have an AED on their location. They did think it was a good idea.
Fysiopraktijk RompertCentrum	No	Most likely did not fill in the survey.
Gemeente 's-Hertogenbosch	Yes	I got an empty response, probably meaning that they ignored it.
Ergotherapie helftheuvel	Yes	Were not willing to cooperate.
Mister design	No	Most likely did not fill in the survey.
Winkelcentrum arena	No	Most likely did not fill in the survey.
Noorderpoort samen doen	No	Most likely did not fill in the survey.
Sporthal Kienehoef	No	Most likely did not fill in the survey.
Woonmeij	No	Most likely did not fill in the survey.
Horecacentrum de Vriendschap	No	Most likely did not fill in the survey.
Zorgboederij Dommelhoeve	No	Most likely did not fill in the survey.
Camping Kienehoef	No	Most likely did not fill in the survey.
A. Kuis uitvaartverzorging	Yes	Mentioned that they filled in both parts.
Gemeentehuis Meijerijstad locatie Sint-Oedenrode	Yes	Said to contact the foundations active in the municipality. Did not fill in the survey.
Fysiotherapie Marie van Heijst	No	Most likely did not fill in the survey.
Acura	No	Most likely did not fill in the survey.

After sending out e-mails to the last nine locations, the responses for the AED owners went up by three. It could be that this is pure coincidence, but to me, it seems that the people in small-town Sint-Oedenrode were more eager to help out than the people in 's-Hertogenbosch. In the end, I do believe that this method works, but it is very time-consuming.

Interview with Mr. Van Dongen and Mrs. Smits

I was able to interview two people who put an AED on their resident houses. They both mentioned that they were willing to help, so I prepared some simple questions that could be answered quickly.

Q: "How long have you been in possession of an AED?"

Answer Mr. Van Dongen: "I've been in possession of an AED for 2.5 years."

Answer Mrs. Smits: "I've been in possession of an AED for a little over 1 year."

Q: "Why did you decide to obtain an AED?"

A,D: "My brother suffered from cardiac arrest, one of the reasons he could not be saved was the lack of AED availability. I decided to purchase this AED with the help of Hartsafe Rooi out of his legacy."

A,S: "It was brought to my attention by a friend of mine that there was a lack of AEDs in our neighborhood. Together we decided to do something for the district by purchasing one together. It is pure coincidence that the AED ended up at my house."

Q: "Was your AED ever victim of (attempted) theft or vandalism?"

A,D: "No."

A,S: "As far as I know not."

Q: "has your AED ever been retrieved for a potential resuscitation?"

A,D: "I live in a very quiet neighborhood, so luckily it has not been needed since I obtained the AED."

A,S: "No."

At this point, I told them about our idea.

Q: "Do you think that your AED will be at more risk when this solution will be implemented?"

A,D: "I think the real risk will not increase, but it feels like the AED will be very vulnerable, but that is something I will have to get eased to probably."

A,S: "Yes, I think AEDs are very valuable, especially when looking at their cost price. Leaving them open like this must increase the risk."

For Mr. Van Dongen I continued directly to the next question. For Mrs. Smits I briefly explained why we believe the risk will not increase.

Q: "Do you think that giving access to the AED remotely is a good idea from your point of view?"

A,D: "Yes, I do. But I also think that this will not be a smooth transition, I think most owners will not immediately accept this solution."

A,S: "It might be, I am still a bit skeptical. But maybe when the solution has been proven to be effective and the foundations are behind the idea, I can be convinced."

At this point, I thanked them for their time and left my contact info if they had some more comments and/or questions.

What I took away is that both Mr. Van Dongen and Mrs. Smits are hesitant to accept the solution. This doubt can be eliminated by providing our arguments in a digestible way. The owners need a good foundation of arguments.

HartslagNu short interview

In my e-mail to HartslagNu, I tried to get at least some response by asking them directly what they thought of the proposed solution. The person responding to my e-mail, Christel van Huffellen, mentioned that HartslagNu was also looking into the viability of this solution with a third party. This means that they already weighed the pros and cons and came to the conclusion that the pros outweigh the cons. They also believe that the owners of AED's will react positively to this change in their system.

Ilse

New strategy

Since, up to this point, reaching out to relatively big social media pages and asking them to post the survey had yielded no results, I decided to use a more personal approach. I live in a more rural area, with lots of towns close together. Saying that ‘everyone knows everyone’ would be an overstatement, but being able to introduce yourself to someone as a friend or family of so-and-so does make them more inclined to talk to you. First, I asked around my immediate family if anyone knew of someone who was a voluntary first responder or owned an AED. This method was already appearing to be more successful that the social media approach, because I got the contact information to a couple that is heavily involved in the local first aid association, and a former colleague who is an voluntary first responder.

After calling the couple, it turned out that the husband had actually helped install the AEDs in our town and he redirected me to the person who he had installed them with. The wife also knew of someone who was involved in the local AED foundation and provided me with her name. This lead me to Mr. Van der Schaaff and Mrs. Ehlebracht, who are a board member and the secretary of Stichting Polderhart, respectively. Stichting Polderhart manages the AED network in the (former) municipalities Graft-De Rijp and Schermer and nearby towns since 2009. Among other things, they ensure the AEDs are maintained properly, provide reanimation and AED training to citizens, promote the voluntary first responder system and give guest lessons on AEDs to children in primary school. Also noteworthy, not all volunteers of Stichting Polderhart take part in HartslagNu. Instead, when the pincode to the cabinet changes every so often, they are notified by Stichting Polderhart of the new pincode.

The former colleague, Mr. Zwarthoed, is a voluntary first responder from Volendam. All three were happy to be interviewed, and the summaries are provided below.

Interview with Mr. Van der Schaaff

Q: “Are most of the AEDs in a cabinet with or without a pincode?“

A: “Most of them have a pincode. There are a couple of old cabinets that do not have a pincode, but all of the newer ones do.”

Q: “So why did you change from cabinets that don’t need a pincode to ones that do?“

A: “The old cabinets we had were not ideal for outside placement due to poor ventilation. And all the newer cabinets you buy either need a key or a pincode. A pincode is easier because you can send it along with the SMS people receive.”

Q: “Did you ever encounter any vandalism?”

A: “No, we haven’t.”

I then briefly explained our project and asked for his thoughts.

A: “We wouldn’t buy them if they are much more expensive. There is also the problem that if a cabinet is opened remotely, but no one goes to retrieve that specific AED, then that cabinet is just open. Now someone has to actually go there to open the cabinet and retrieve the AED.”

Q: “So you do consider there to be an increased risk if the cabinet were opened longer?”

A: “Yes, I guess. Also, the connection may not work flawlessly, we have never had issues with the current system not working as intended.”

Q: “To your knowledge, have there ever been any issues with opening the cabinets?”

A: “There was one time when someone had remembered an outdated pincode and was unable to open the cabinet. But since people are sent to retrieve multiple AEDs, an AED from another location was retrieved instead.”

Interview with Mrs. Ehlebracht

Q: “Why did you switch from unlocked to locked cabinets?”

A: “After one had been stolen and we wanted to be sure they were secure.”

Q: “One had actually been stolen?”

A: “Not from us, but in Alkmaar*.” (*A bigger city nearby.)

Q: “Did switching to cabinets that are locked with a pincode give you any problems?”

A: “Yes, because some people forget there is an pincode and, in panic, are unable to open the cabinet.”

Q: “Did you ever have any vandalism attempts?”

A: “No, we’ve never had any. We go around the schools to explain to kids about AEDs, what they are far and how to use them and such. I think that helps.”

I then briefly explained our project and asked for her thoughts.

A: “I do believe there are some cabinets that you can have access to via internet, but we don’t have those. They are way too expensive.”

Q: “So are you satisfied with the current pincode system?”

A: “Sort of yes. However, when someone calls 112 because someone needs to be reanimated, I think 112 should have a list with pincodes.”

Q: “Because then they would be able to provide the pincode immediately.”

A: “Yes. Now we have to wait until someone gets a message. I firmly believe this to be a weak point in the system. I have asked HartslagNu if they could change this, but there has been no progress on that front.”

Interview with Mr. Zwarthoed

Q: “For how long have you been a voluntary first responder?”

A: “For as long as the system has existed.”

Q: “Have you had to retrieve an AED often?”

A: “Well, no, because there is one at my front door.”

Q: “And is that cabinet locked?”

A: “No, it is not.”

Q: “Do you know if the other AEDs in Volendam are in locked cabinets or not?”

A: “A few of them are. Those that are privately owned. But all the AEDs that belong to our first aid association are publicly accessible.”

Q: ”Have there ever been any vandalism attempts?”

A: “Yes, twice. One was stolen near a public road, and another was destroyed. But that is since they were deployed, so since 2002.

Q: “So it does not happen often enough to warrant a locked cabinet.”

A: “No, I don’t think so.”

Q: ”Do you know if there have ever been problems opening the privately owned cabinets?”

A: “No, there have never been problems with that.”

Q: “So you are satisfied with how the AEDs are managed in Volendam?”

A: “Yes, absolutely. But I do think more should be unlocked.”

I then briefly explained our project and asked for his thoughts.

A: “Sounds great. What happens now is when a privately owned AED is used, then the battery is refunded by the municipality. It would be great if the added costs would be covered too.”

Short summary

All three interviewees were concerned about the added costs of our idea, stressing the importance of keeping the costs as low as possible. Mr. Zwarthoed firmly believes more cabinets should be unlocked, despite the added risks of vandalism and/or theft. Mrs. Ehlebracht believes bystanders should be able to unlock cabinets as well to reduce the time needed to retrieve an AED. Mr. Van der Schaaff is concerned about the safety of the AEDs if a cabinet were unlocked remotely but no one goes to retrieve it and thus lock it again.

Contact with GT Medicare B.V.

For the demonstration of the prototype, it is ideal if the final design was implemented in an actual cabinet. This way, the real time won with our implementation can also be seen. Furthermore, we wanted to know what their views are on this type of cabinet design. This let us on a quest to try and find companies willing to help us with our project. Something we at first thought would be quite easy turned out to be rather difficult. The main problem with finding companies willing to help us was the fact that there are not many AED cabinet producers to start with. When looking around we found two major companies that were producing cabinets. The first one of these was AIVIA.

AIVIA

AIVIA is one of the biggest suppliers of AED cabinets in the Netherlands. They have a wide range of models ranging from indoor always open cabinets to outdoor locked cabinets. However, once we started investigating further we found out that AIVIA was based in France. Something that does not immediately result in problems but we could run into some translation difficulties. This turned out to be the case between us an AIVIA since we attempted multiple times to get into contact with them to no avail. They were not willing to participate in our design and we unfortunately had to halt this option.

Parallel to AIVIA we also came across SixCase as one of the cabinet producers, this is were GT Medicare B.V. came into the picture.

GT Medicare B.V.

We quickly learned that GT Medicare B.V. is a Dutch company that also produces AED cabinets. We then got into contact with Gise Wit, one of the engineers at GT Medicare. We had a meeting together with Mr. Wit and our group where we discussed how we were going to help each other out in this project. He was enthusiastic about the idea and wanted to help us in our design process. This help was facilitated via extra information about the cabinet, schematics for the PCB as well as the fact that they sent us a cabinet to try our design on.

Selected design

Description

It was decided to use SMS as a communication tool. The lock will get an SMS module and on receiving an SMS with a specific text e.g. "OPEN", the AED cabinet will unlock. The Server of HartslagNu, on receiving an alert from the 'Meldkamer' will, in addition to alerting voluntary first responders, send an SMS to the AED cabinet, which unlocks it. The link below redirects to a video that demonstrates the time that can be saved if the voluntary first responders does not have to insert a pincode. This solution has the added advantage of giving the entire voluntary first responders network of HartslagNu access to an AED. In the current system, only people subscribed via the app will be paged to an AED (HartslagNu, 2020). If there is a situation where only people with SMS notifications get paged, an AED will not be available, thus greatly reducing the chances of survival for the victim. With this solution, an AED will always be available in the direct area of a possible resuscitation, thus increasing the survival rates of out-of-hospital cardiac arrests.

Demo

Motivation for SMS

There are three main reasons this technology was chosen. First, it is known for a fact that the server of the 'meldkamer' and HartslagNu already works with SMS, because they send an SMS to the voluntary first responders to alert them. Second, SMS has a high national coverage, practically 100% in urban regions ( https://www.kpn.com/netwerk/dekkingskaart.htm ). This means our system will work in all places with a mobile network, which is practically everywhere. Third, this is easier for the users of the app. There is no need to communicate via their mobile phone, which would create an extra point of failure, and requires more compliance from the users.

Full System

Diagram for the Full System

The system would integrate an SMS gateway in the existing webserver of HartslagNu. Instead of sending a text message to the voluntary first responders only, a separate text gets sent to the AED cabinet, which houses a microcontroller with a GSM module. This is connected to the locking mechanism of the cabinet, and is able to open it when the correct message has been received.

Risk scenario analysis

The currently proposed solution might make the AED more vulnerable to thefts, because the AED is accessible during a bigger time window. In this section, the risk increasements will be assessed in order to convince all users of the importance of the solution. When conducting the interviews with the private AED owners, it was noticed that these users were worried about the increase of theft. For these users the increase of theft risk has been assessed. When conducting the interviews with people from local foundations, it was noticed that they were worried about the increase in costs. For them, the economic plan has been made.

Numbers of thefts in the current situation

The amount of stolen AEDs is traceable due to the existence of a database for stolen AEDs. This database is for owners of AEDs to report their stolen AED using its unique serial number. The stealing of AEDs is rare according to this database (Gestolenaed, 2019), around five to ten AEDs get stolen each year in the Netherlands. This makes it rare since there are around 100000 AEDs in the Netherlands, making the theft rate around 0,01 percent of AEDs get stolen annually. There are a few reasons why AED theft is uncommon:

1. AEDs contain their own unique serial number, which makes tracking down stolen AEDs much easier (Gestolenaed, 2019).

2. Selling a stolen AED on second-hand websites is not useful since most people prefer to buy a brand new AED (Bos, 2019).

3. The parts that make an AED are not in demand on the second-hand/black market, so trying to sell for the parts is not lucrative (Bos, 2019).

4. The language set on an AED is not easy to change, so selling to foreign countries is not lucrative (Bos, 2019).

5. Currently, AEDs are well protected by an electronic lock.

The AEDs that do get stolen sometimes do end up abroad but are able to be traced back to the original owner. Vandalism is also a reason why some AEDs get stolen. Not every AED will be found after it has been stolen, so the real reason behind every theft is unknown (Bos, 2019).

Increase of theft risk in the new situation

Risk of theft

In the new situation, AED boxes will be open for a longer period of time, the owners of the AED might see this as an increase of risk on their side. To put the owners at ease, two main points are made why the risk will not increase:

1. Stealing an AED, in general, is not lucrative for criminals. This does not mean the AED can be put outside unprotected. This does however make it viable for a solution where the AED box is open for a bigger time window in the case of an emergency. Assuming an emergency happens every 24 hours and the AED box is remotely opened for 6 minutes, this means that the AED will be accessible 0.42% of the time. The nature of an emergency makes it impossible for a potential criminal to take advantage of this new system since the effort to gain ratio is way too high. The rate at which an AED is needed in an area however, is way lower. It is more likely in the region of once every one or two months. making the percentage go down way lower.

2. Theft usually happens with brute force, where the AED box is being demolished to gain access to the AED. This does not change from the old situation to the new one, so increased theft rates are unlikely to happen due to the current situation.

Due to the current consensus on why AEDs generally do not get stolen, the assumption can be made that the risk will not significantly increase when the new solution will be implemented.

Risks of failure

According to research done by P3 and Connect in 2019, the mobile network in the Netherlands is quite good, especially considering down-times (P3 & Connect, 2019). Down-times are time windows where the mobile network is unavailable. The worst provider was down for 17 hours over 9 months, but the provider used by the emergency services, KPN, was only down for 1 hour over 9 months. This makes the risk of failure for the solution considerable. However, the nature of the solution helps in these cases. The solution has always been an addition to the current situation. The pincode used now to open the AED cabinet will not be removed. This is because, for maintenance reasons, the AED still should be accessible without using a remote-controlled option. This means that the app still can send out the pincode of the cabinet in case of a network outage, or a software/hardware error. A network outage also causes the emergency number to not be available. In this case, the whole network of voluntary first responders will crumble before the chosen solution comes into play. The scope of this project, however, is smaller than solving this more general problem. Due to the nature of the solution being an addition, the risk of failure will not increase with respect to the original situation.

Cost analysis

In this section, the additional costs of the proposed solution will be analyzed. At first, the additional cost of the webserver will be analyzed. After, the additional cost of the module on the AED cabinet will be approximated.

Extra costs for the webserver

In the proposed solution, the intention is to adjust the webserver currently used to page the voluntary first responders, so it is also able to unlock the AED cabinets in the area of a possible resuscitation. This requires the webserver to undergo a slight rebuild to allow for this feature. Since the idea is to make an add-on to the code that is running on the HartslagNu webserver, the costs should not increase for the users. A webserver can cost a high-traffic domain like the national control room up to approximately €600 per month (timothy_stiffler et al., 2008). Because of the high costs associated with introducing a new webserver, it is important that this addition is possible. If this addition takes a lot of RAM to run, it might be needed to upgrade the current webserver. Since information on the webserver of HartslagNu is limited, a real approximation cannot be made. Therefore the assumption now is that the addition is compatible with the current system, and the costs will not rise.

Extra costs for the AED cabinet module

If the final product will be built from the same materials as the prototype, the total cost that will be added to the box will be around €18 (Bol.com and Amazon). Since Phillips sells AEDs with additional cabinets at a price of approximately €1600 (Procardio), the conclusion can be made that the purchase price will increase by a maximum of 1.1%. This is most likely the maximum since the price of the components is now calculated if they were bought piece by piece. Buying the components in bulk will drive the price further down. The Arduino Nano is most likely too powerful for this application, a less powerful processor might also drive the cost further down. The only component that might increase the price is the GSM module. Later on, it will be explained why this particular version did not suffice, and thus should be replaced with a more modern module. The prices for these modules are not available for public and could not be requested at will. The exact change in pricing is thus unknown.

Prototype

Due to certain limitations on pricing, scaling and availability of parts, our demonstration prototype will differ from the design in several ways. One such a way is the use of an Arduino Nano inside the cabinet itself. When implemented on a bigger scale, cheaper and smaller microcontrollers should be used to lower the cost and to take up less space. However, due to the ease of use of an Arduino Nano, its availability, and its large selection of libraries, it will be used for the demonstration. Functionality wise, this changes nothing.

Adding the SMS functionality with the gateway through an ISP to the server of HartslagNu is a viable option because of the scale of the operation and all the information the server already has access to. However, for demonstrative purposes, a server on a Raspberry Pi will be set up with an external GSM module for sending the text messages. This is due to the small scale of the prototype, although the functionality is the same. Especially since the Raspberry Pi also has webserver capabilities if that becomes necessary.

By using separate GSM modules instead of the webserver, double the number of SIM cards is required, though only the one used for sending texts needs credit. Because of the scale, which prepaid card is being used is of very little importance as long as the ISP it belongs to covers all corners of the country. Pricing is less of a factor because of this.

System Diagram for the Prototype

Prototype Assembly Process

The prototype can be divided into three parts: the Pi with the server and SIM800L module, the Arduino Nano with the other SIM800L module that gets connected to the cabinet, and the cabinet itself.

Raspberry Pi

The reason a Raspberry Pi was chosen to act as the initiator is because of the relative ease to set up a simple webserver on it. It also has physical serial ports, allowing the GSM module to be directly connected. The server side was set up using the apache framework, creating an IP to which data could be sent. The GSM module was connected to the Pi as follows:

Wiring for Pi Prototype https://raspberrypi-nederland.nl/sim800l-gsm-gps-framboos/

The RX and TX pins on the Raspberry Pi are the serial ports that allows the Pi to communicate with the GSM module. This GSM module was chosen for its small size, low price and availability, though it came with a small issue. The GPIO pins on the Pi only output 3.3V or 5V, but the GSM module required anywhere between 3.7V and 4.4V. To deal with this disparity, a diode was added that dropped the 5V down to 4.4V. The capacitor was added to provide more stability. This way, the Pi could power the GSM module directly.

Through the Minicom library for Linux, the Pi would be able to utilize the GSM module. However, some troubles arose while using this library. After being installed on the Debian image on the Pi, Minicom would not take any keyboard input. Only after turning on the local echo did it accept any input. However, no connection could be made the GSM module. After further inspection, the GSM module did not seem to receive any power, with or without the diode. After further testing, including wiring the GSM module to an Arduino instead of a Pi to access the direct serial monitor in the Arduino IDE, the GSM module seemed to be malfunctioning, indicating a bad unit. Another SIM800L module was acquired and this one did work. Despite the diode and capacitor, the GSM module was still overvolting though it was functioning.

After activating the SIM cards separately, the module was ready to be connected to. The LED indicator showed that, although the module was on, there was no connection to a network.

Initializing...

AT

OK

AT+CSQ

+CSQ: 16,0

OK

AT+CCID

8931080520037043401f

OK

AT+CREG?

+CREG: 0,3

OK

AT+COPS?

+COPS: 0

OK

OVER-VOLTAGE WARNNING

AT+CPIN?

+CPIN: READY

OK AT+COPS=?

+COPS: (2,"KPN MOBILE","NL KPN","20408"),(1,"vodafone","voda

The above monitor output shows some basic diagnostics about the SIM card and GSM module. It shows that a connection was made with the GSM module, it recognizes the SIM card number and shows strong signal strength (20,0). However, the (0,3) implies no connection to a network was made, so it was impossible to send any messages. When checking for networks, the GSM module did find the networks of KPN and Vodafone. When tested in an actual phone, the SIM card was able to connect to a network and send a text, which excluded it from being the cause of the issue.

Combining prototype and cabinet

At the heart of the received AED cabinet is an Atmega32 microcontroller. The version of this microcontroller is about the same version as is used in the Arduino range. Since the final design should be embedded as small as possible, an Arduino Nano was chosen to use in the prototyping stage. As is discussed in the previous section with the Raspberry Pi, a SIM800L module will also be used for the receiving side. There are a few things to tackle in this design. These are: the power, the communication and the implementation. These will now be discussed. In this discussion it is important to remember that these different topics have to eventually be included in the existing cabinets. Where needed we will be looking into the current designs of the cabinets and make design recommendations for the next PCB version.

The power

The SIM800L is quite a power sipping module when not in use. It will consume some milliampères of power. However, when this module has to transmit something it needs a burst of electricity to generate a high enough electric field in order to reach the local GSM towers. That is why the module needs to be able to consume two ampères of electricity at around 3.7V at once in order to function properly. In our testing it was indeed concluded that the module does 'function' with a less capable supply but will not be able to actively communicate with a network since it cannot produce a strong enough electric field.

In order to produce enough current in our test circuit, a lithium-ion battery was utilized since this would be able to supply the needed current. The current cabinet however does not come with this battery since its power is supplied by the power grid. The power of all the SoCs (systems-on-a-chip) is supplied by a LM317EMP, set at 3.6V. This is a rather capable voltage regulator that unfortunately does have one caveat, this regulator can only supply a maximum current of 1.5A. Which is too little for this application, let alone the whole circuit. Our recommendation on this topic is thus to choose a more capable voltage regulator able to supply around 2.5A in order to leave enough room for other applications.

The communication

The communication between the Arduino and the SIM800L module will be facilitated via AT-commands, a command set widely used in GSM SoCs. The RX and the TX pins that will be used for this communication can be chosen arbitrarily, because the SoftwareSerial library will take care of this for us. The commands that will be sent are standardized, meaning different commands for different SoCs can be reused. For the prototype the SIM800L module was chosen due to its compact size and 2G functionality, resulting in a wide coverage, and the fact that it is widely available for consumers to buy. For the final design inside the PCB we would recommend going with a slightly more capable SoC with more than 2G coverage in order to future-proof the system, because when considering the results of the user questionnaires it can be noticed that people do keep their AED for quite a long time. Currently 2G will be supported until 2025 meaning that at some point it will become obsolete. Which is not desirable for a long-term purchase. Of course the final choice of SoC will be decided by the manufacturers themselves.

When looking at the current schematic of the PCB there are plenty of I/O pins left to attach this PCB to, so this should not pose a problem in the implementation.

The implementation

Integrating the design into the actual PCB currently in the cabinet is not an easy feat. The design of the current PCB is fairly compact and does not have room for upgrades. The only possible solution would be to make use of the ICSP pins. In Circuit Serial Programming pins are pins that are used to (re)program the IC. The benefit of using these pins is the fact that they are already there on the circuit board, after the IC has been programmed. This will, however, introduce an extra step in production and the board cannot be programmed afterwards. A more feasible option would be to produce a new PCB with all the components currently on it and swap these boards.

Another thing to keep in mind for the specific use case of GT Medicare B.V. is the fact that most of the cabinet is made out of metal except the front glass. Since there is also metal shielding the PCB it is wise to use external antennas in order to keep the signal as strong as possible.

The prototype

For the prototype we were going to reprogram the Atmega32 processor that was already present on the PCB of the cabinet via the ICSP pins and then connect the GSM module to these pins. Before attempting this, the module had to be tried out first. During this step problems were encountered with the GSM module, since the module at first would not function properly. After some troubleshooting this boiled down to the fact that the module did not receive enough power, which was resolved via directly powering the module with a lithium-ion battery. The module now communicated properly with the Arduino, however, it still could not connect with the network. Despite lots of troubleshooting, different GSM modules, different power supplies, reflashing its firmware and lots of further troubleshooting, the module still refused to connect to a network. Even though this type of module worked a couple of years ago. This lead us to assume that maybe the module has become out of date. To try and remedy this we investigated if we could buy a newer model. However these were either quite expensive for consumers or not available. Because of this we would advise GT Medicare B.V. to look into these modules, since the production models are cheaper and more future proof by having 3G or 4G networking capabilities.

Conclusion

Even though we were unable to develop a proper working prototype, we are still confident the solution could work. Most of the people that were interviewed are positive about the idea and think it might contribute to saving time and thus saving lives. The proposed design solves many problems, while not adding many of its own. Currently, people that receive an SMS-notification only, i.e. they do not use the app, have no way of knowing what the pincode of a locked cabinet is. Our design gives them, as well as bystanders, access to locked AEDs.

As demonstrated in the demonstration video, our system saves a significant amount of time in retrieving the AED. SMS-notification introduces no new points of failure. Even if something locally goes wrong, it will not be worse than the current system. There are, however, some negative aspects. Implementing the design in current lockers is difficult to do. Manufacturers and the 'meldkamer' both need to make adjustments, which is more of a political than technical challenge. Also, the costs and the power usage will increase slightly. All in all, we believe that this concept has a lot of potential and, when correctly developed and implemented, will help save lives.

Discussion

The biggest limitation of our proposed solution is the lack of uniformity among AED cabinets. This causes the implementation of the module to be significantly more difficult and more time consuming depending on which type of cabinet is considered. This depends on the chip used in these cabinets, their respective reachability, the amount of space available, and also the way the cabinet opens. As such, certain cabinets may need more work to be done to them before the module can be properly implemented.

Similarly to the uniformity among AED cabinets, there was too much optimism in connecting our module to the existing PCB and micro-controller. Because of the proprietary nature of these PCBs, accessing any part of it was a greater challenge than expected. This also hinders the general applicability as all AED cabinets are likely to have different systems and PCBs inside.

The GSM modules we used also caused problems. The reasons for choosing this particular SIM800L module were costs and consumer availability. However, higher grade components outside of general consumer markets should be considered when implementing it properly. Making sure these modules support 3G and 4G is also highly recommended as this increases the longevity of the system.

Regarding the user's needs, we lack data to justify the proposed solution in the first place. While the answers we did receive seemed promising, we did not receive enough responses to properly justify this implementation. More cooperation with larger foundations that involve our target audience would be beneficial in acquiring this data. However, as it stands, very few companies were inclined to respond, let alone spread our survey, including HartslagNu itself.

This lack of data is especially true for AED owners. The amount of responses we got from such parties is incredibly low. A bigger time window needs to be considered to acquire this data and they need to be involved earlier on.

Other user responses we received were skeptical about our solution and if this was even a problem in the first place. This might be caused by the lack of initial information given about the problem in the survey we spread. As such, respondents were not very well informed about why this is such an issue that needs solving. Future surveys should include more specifics about the problem and implementation to allow the respondents to give more informed feedback.

When conducting research, it seemed as though no one was looking into a solution to our proposed problem. However, after getting in contact with several parties, it came to our attention that a similar fix was already being investigated. While most of these parties were unwilling to cooperate with us, it would still be beneficial to start searching for people working on similar ideas.

Planning

Week	Task 1	Task 2	Task 3	Milestones (end of the week)
Week 1	Pick subject	Collect background information	Make a planning	Subject chosen
Week 2	Discuss different types of connections	Collect background information		Formulate possible solutions
Week 3	Work out final design details	Reach out to manufacturers		Final design has been chosen
Week 4	Bill of materials	Research technical details design	Reach out to users	Spread user survey
Week 5	Incorporate user feedback	Receive components	Conduct user interviews	Start building prototype
Week 6	Build the prototype	Troubleshooting	Test the prototype	Finish building prototype
Week 7	Film the demo	Edit the demo	Write/prepare presentation	Presentation/demonstration is finsihed
Week 8	Presentation	Peer review	Finalize the wiki page	Complete wiki page

Work done per week

Week 1

Name	Total [h]	Break-down
Mark den Besten	7	Meetings [3h], Starting lecture [1h], Research and writing User part [3h]
Yann van Eijk	7	Meetings [3h], Starting lecture [1h], Research and writing Approach, milestones and deliverables [3h]
Ilse Schuckman	8.5	Meetings [3h], Starting lecture [1h], Research and writing Problem statements & objectives [4.5h]
Rik Schutte	8	Meetings [3h], Starting lecture [1h], Research and writing User requirements [4h]
Roel Wijnands	7	Meetings [3h], Starting lecture [1h], Research and checking references [3h]

Week 2

Name	Total [h]	Break-down
Mark den Besten	10	Meetings [4h], Research Design options [3h], Researching SMS API [3h]
Yann van Eijk	10	Meetings [4h], Rewriting User requirements [2h], Implementing work done in Week 1 on the wiki page [1h], Fixing SotA and References [3h]
Ilse Schuckman	7.5	Meetings [4h], Research and Design options [3h], updating wiki [0.5h]
Rik Schutte	9	Meetings [4h], Rewriting User requirements [2h], Research SMS API + Options [3h]
Roel Wijnands	9	Meetings [4h], Orderning papers [2h], research component choices [2h], research communication options [1h]

Week 3

Name	Total [h]	Break-down
Mark den Besten	7	Meetings [4h], Meeting with producer [2h], Selected design [1h]
Yann van Eijk	8	Meetings [4h], Survey [1.5h], How it currently works [1.5h], Reference searching [1h]
Ilse Schuckman	8	Meetings [4h], Current cabinet designs [2.5h], Survey [1.5h]
Rik Schutte	9	Meetings [4h], Creating user scenarios [2h], GSM module + RaPi webserver research [3h]
Roel Wijnands	9	Meetings [4h], Contacting AED cabinet companies [3h], meeting with GT Medicare B.V. [2h]

Week 4

Name	Total [h]	Break-down
Mark den Besten	6	Meetings [4h], Extra SOTA [2h]
Yann van Eijk	8	Meetings [4h], researching/writing risk analysis [3h], Editing wiki (grammar/spelling) [1h]
Ilse Schuckman	6	Meetings [4h], Adjust survey [1h], Contacting social media [1h]
Rik Schutte	8	Meetings [4h], Setup RasPi Webserver [3h], Create Diagrams [1h]
Roel Wijnands	8	Meetings [4h], Setting up company agreement [1h], Contact with GT Medicare B.V. [1h], Researching current PCB design [2h]

Week 5

Name	Total [h]	Break-down
Mark den Besten	4.5	Meetings [3h], Preparing for prototype [1.5h]
Yann van Eijk	5	Meetings [3h], Adjusting risk analysis [2h],
Ilse Schuckman	5	Meetings [3h], Contacting social media [2h]
Rik Schutte	7	Meetings [3h], Researching Minicom [2h], Start writing and researching prototype [2h]
Roel Wijnands	7	Meetings [3h], Investigating cabinet [2h], Researching PCB module options [2h]

Week 6

Name	Total [h]	Break-down
Mark den Besten	12	Meetings[3h], Collecting components [1h], Creating prototype [2h], Troubleshooting and trying to make the prototype work [6h]
Yann van Eijk	12.5	Meetings [3h], Searching for AED owners [2h], Contacting AED owners via e-mail [3h], Interviews via phone [1.5h], Adding contact with user part to wiki [2h], Spreading survey via WhatsApp [0.5h], Bullet point list discussion section [0.5h]
Ilse Schuckman	10.5	Meetings [3h], Asking around for relevant users [1.5h], Research on AED foundations nearby [1h], Mail contact with users [1h], Conducting interviews [2h], Updating wiki [2h]
Rik Schutte	16.5	Meetings [3h], Acquiring Components [1h], Setting up Debian on Pi with Minicom [2h], Creating physical Prototype (Pi) [2h], Testing and troubleshooting Prototype [7h], Updating wiki [1h], Spreading survey on WhatsApp [0.5h]
Roel Wijnands	7	Meetings [3h], Pick up components [1h], Creating physical prototype (Arduino) [1h], Researching design implementation [2h]

Week 7

Name	Total [h]	Break-down
Mark den Besten	13	Meetings [5h], SIM module last tries [3h], Recording Video [3h], writing wiki [2h]
Yann van Eijk	11	Meetings [5h], Making/practising presentation [1h], Editing wiki-page [4h], Researching costs [1h]
Ilse Schuckman	10.5	Meetings [5h], Writing presentation [3h], Editing video [2.5h]
Rik Schutte	14	Meetings [5h], Recording video [3h], Trying to fix SIM module [5h], writing documentation[1h]
Roel Wijnands	21	Meetings [5h], Recording video [3h], Troubleshooting SIM module [9h], Writing documentation [4h]

Week 8

Name	Total [h]	Break-down
Mark den Besten	5	Presentations [2h], Peer review meeting [1h], Updating Wiki [2h]
Yann van Eijk	6	Presentations [2h], Peer review meeting [1h], Updating Wiki [3h]
Ilse Schuckman	7.5	Presentations [2h], Updating wiki [4.5h], Peer review meeting [1h]
Rik Schutte	6.5	Presentations [2h], Peer review meeting [1h], Updating Wiki [3.5h]
Roel Wijnands	5	Presentations [2h], Updating Wiki [2h], Peer review meeting [1h]

Peer review

Matrix

Name	Result
Mark den Besten	0
Yann van Eijk	0
Ilse Schuckman	0
Rik Schutte	0
Roel Wijnands	0

Elaboration

We had some discussions about who deserves more and who deserves less. Some people did more tedious tasks that were very frustrating to do. For example, the people working on the GSM module did not have the best of times. In contrast, the people working on the user end of the project had to undergo a lot of rejection, which was also very frustrating. In the end, we think everyone delivered their work properly and no-one was running away from their tasks. We therefore based the results on the effort everyone has put in. To us, everybody in the group put in just as much effort as anyone else in the group, thus no one should have a point deduction. This also means no one will gain points, because of the fact that the sum of the result has to be zero. In the end, we feel confident carrying the result of our work equally.

References

[1] Amazon. (n.d.). AZDelivery SIM800L GSM GPRS-module met Antenne COMPATIBEL met Arduino Inclusief e-book! Retrieved April 01, 2021, from https://www.amazon.nl/dp/B07D361WZV?tag=pinederland-21&keywords=sim800&geniuslink=true

[2] Bhvnederland. (2018, July 27). Yearly 5,300 people die in The Netherlands due to heart attack. Retrieved February 21, 2021, from https://www.bhvnederland.nl/bhv-nieuws/aed-infographic-en

[3] Bol.com. (n.d.). Arduino Nano 3.0. Retrieved April 01, 2021, from https://www.bol.com/nl/p/arduino-nano-3-0-16mhz-atmega328p/9300000028694555/?Referrer=ADVNLGOO002013-G-118359660147-S-1005228050775-9300000028694555&gclid=CjwKCAjw3pWDBhB3EiwAV1c5rKGbKXzafjk6XgF5TH-e9l3zq5WOyJ-3qv7wQyKxMi1VzXRBe2Fh1BoCd-QQAvD_BwE

[4] Bos, M. (2019, December 14). Wat kunnen dieven eigenlijk met een gestolen aed? Retrieved March 02, 2021, from https://www.ad.nl/dordrecht/wat-kunnen-dieven-eigenlijk-met-een-gestolen-aed~a3cb6357/

[5] Burgerhulpverlening. (2021, February 18). Retrieved February 21, 2021, from https://nl.wikipedia.org/wiki/Burgerhulpverlening

[6] Dagblad van het Noorden. (2019, October 07). Hartstichting: ,,MEER AEDS nodig om EMMEN hartveilig te maken". Retrieved February 21, 2021, from https://www.dvhn.nl/drenthe/emmen/Hartstichting-Meer-AED%E2%80%99s-nodig-om-Emmen-hartveilig-te-maken-24897372.html

[7] Gestolenaed. (2019, December 18). Retrieved March 02, 2021, from https://gestolenaed.nl/

[8] HartslagNu. (2020, October 13). Alarmering - Hoe werkt de alarmering van HartslagNu? Retrieved April 02, 2021, from https://hartslagnu.nl/veelgestelde_vragen/alarmering/

[9] HartslagNu. (2021, January 21). HartslagNu app. Retrieved February 21, 2021, from https://hartslagnu.nl/burgerhulpverlening/hartslagnu-app/

[10] Jafri, S. (2021, January 30). Rapid defibrillation important for survival: Avive solutions. Retrieved February 21, 2021, from https://avive.life/guides/importance-of-time-to-defibrillation/

[11] Klare, B.F., Burge, M.J., Klontz, J.C., Vorder Bruegge, R.W., Jain, A.K. (2012). Face Recognition Performance: Role of Demographic Information. IEEE Transactions on Information Forensics and Security, vol. 7, no. 6.

[12] Medisol. (n.d.). Onderhoudskosten Van de AED. Retrieved February 21, 2021, from https://www.aedwinkel.nl/onderhoudskosten-aed

[13] P3, & Connect. (2019). The 2019 Mobile Network Test In The Netherlands. Retrieved April 01, 2021, from https://tweakers.net/files/upload/NL%20Benchmark%202019%20V5[1].pdf

[14] Procardio. (n.d.). Philips aed: Aed KASTEN en Beugels. Retrieved April 01, 2021, from https://www.procardio.nl/shop/c/aed-kasten-en-beugels/c08

[15] Solanki, M. (2017, October 28). Two-thirds of DUTCH ambulance services do not respond within the set time limit. Retrieved February 21, 2021, from https://www.iamexpat.nl/expat-info/dutch-expat-news/two-thirds-dutch-ambulance-services-do-not-respond-within-set-time-limit

[16] timothy_stiffler, Dan_Grossman, Son_Nguyen, Diligenthost, Beley, & Ezitis. (2008, August 24). Server costs - high traffic websites. Retrieved March 25, 2021, from https://www.sitepoint.com/community/t/server-costs-high-traffic-websites/4148