Related image


Please, however, refrain from taking off all your clothes.

Have you ever been making bacon or boiling water and had to stop to remind your smoke detector that you are alive and well?

How about having to do a post-shower workout to grab a chair and use a manual, makeshift fan (aka a towel or pillow) because the detector is too sensitive to steam?

In the next year or so many of us will be getting jobs (yikes), moving into our own homes (small apartments), and gain new responsibilities–this could very well be one of them.

What if you had a smart detector? One that “knew” you were showering or cooking? Could a detector be equipped with a humidity monitor or motion sensor that would notify the machine there is no need to blare a false positive? Don’t get me wrong. Smoke detectors are a necessary evil and save thousands of lives each year. But, maybe we could save our beloved firefighters a wasted trip or two. Before we dive into the various ways this system can be improved, let’s discuss how it works as of now.


What is a smoke detector and how does it work?

The overall system is pretty simple and consists of a detector as well as an  audible or visual

response to an imbalance. There are essentially two forms of smoke detection that differ in how they detect smoke particles. One uses ionization and the other utilizes photoelectric technology.  Rather than talk you through the details, please enjoy the video below that summarizes the the functionality of both types in a tasteful, yet ambiguous accent (feel free to stop watching at the bonus facts at 4:36):

Now, the details on how each individual detection system works are not super significant because the best (aka the most expensive) smoke detectors utilize both technologies. However, in case you skipped the video or are not a visual learner here’s what you need to know.

Ionization: air is made conductive within a detection chamber via a small amount of radiation from a man made metal called Americium. Inside the chamber are two metallic plates with opposing charges (thanks to the voltage sent out by a battery one should change twice a year!). Alpha particles emitted by Americium collide and interact with Nitrogen and Oxygen molecules, splitting them into electrons and ions. The electrons travel to the positively charged top plate and the ions to the negatively charged bottom plate creating a current. When small smoke particles enter the detection chamber it causes a deviation from the constant current and this disruption causes the alarm to be triggered.

Now, a photoelectric detector is much more short and sweet, but needs a little visual assistance.

The T shaped chamber contains a light that travels horizontally. When smoke enters the chamber, the large particles disrupt the light’s linear path and light is refracted and detected at the bottom photocell–signaling the alarm.

Now as we established, the best detectors utilize both technologies. This is because large smoke particles (resulting from slow burning, smoldering fires) do not disrupt the electric current in an ionization detector as much as small particles (the small electrons and ions are able  go around the smoke essentially maintaining a constant current). Small smoke particles result from a quickly burning open flame and do not cause the light in a photoelectric detector to refract as dramatically. All in all, if you can’t spring for the dual sensor, photoelectric is generally the way to go. But because we can’t be held liable, here are some people more certified to help you make an informed decision:


The decision is not a trivial one. A majority of household fire related injuries and deaths result from smoke inhalation and not from the flames themselves. Disclaimer: the following video describes a tragedy and an interview of maybe the most uncompassionate person you’ll ever encounter (5:19).


All in all, a false negative is bad at any time and a false positive is only bad sometimes. As stated before, thousands of people are saved by smoke detectors and systems every year. Many people are annoyed by false fire alarms because they lack specificity. Whether it be a reaction to boiling water, a steaming shower, or flat-iron singed hair, it is not uncommon for a detector to trigger a false warning response.

Our solution builds on a newer model for smoke detection made by Nest and is referred to as the Smart Smoke detector equipped with an app for your phone, speaking functions, a motion sensor, humidity monitor (no more steam alarms!), cool lights and everything. The video is over 20 minutes long and due to the fact that ain’t nobody got time for that we’ll walk you through our control process.


Why does it need control?

So essentially, we need to control when the alarm sounds. This is critical, because as we mentioned before, an alarm that sounds when it shouldn’t is annoying, but one that doesn’t sound when it should is disastrous. When the alarm sounds will have two aspects to it: an automatic aspect, where the alarm will sound regardless of human input, and a manual aspect, where the human can preemptively abort an alarm sequence. In the former instance, the alarm will sound when the ion detector sends a positive signal (indicative of flames). In the second instance, the photo detector (which reacts to smoke/water vapor particles) will signal for a 30 second warning indication during which a person can press a button, or wave towards the detector to initiate the cancellation of the alarm. Any lack of action on the user’s part will result in the alarm going off as normal.

As you can guess, what we’re really manipulating here is the user input. As the device is configured, it’s the only thing we actually have control of. We posit two ways to accomplish this, both fairly convenient. Either the press of a button (most likely located near the stove which we anticipate to be the source of the possible alarm) or a wave to a motion detector would  send the override signal.

Our disturbance variable, which can fluctuate, is the air sample/ quality. This would have two facets to it. If the air sample contains enough smoke to trigger the photoelectric detector, a warning signal will be produced. Upon hearing this signal, the user then can choose whether or not to allow the alarm sequence to continue by changing the manipulated variable, the cancellation signal. If a cancel signal is sent, the sequence ends. If no signal is sent (either by choice or by absence), the sequence proceeds as normal, and the alarm blares.

The second facet of the disturbance variable is reliant on the ionization detector. If the air sample has enough smoke to trigger the ionization detector, this signal is given priority above all else, and the alarm sequence will proceed, regardless of all other inputs to the system. The logic behind this is that if the situation has reached that point, the user is either definitely absent or out of control of the situation. In either case the alarm is warranted. A simple block diagram of the process is seen here.

Now, all this based on the assumption that the alarm functions normally. Most household smoke detectors are battery powered, so another aspect of the system to consider is whether the alarm is receiving enough power to operate. There aren’t many ways to ensure this beyond connecting the alarm to the house’s power grid, or regularly changing batteries. Large fluctuations in current, or a tripped circuit breaker will cause most smoke detectors to go off. Even worse is a case where an alarm runs out of battery power, and fails to go off. The united states fire administration recommends testing alarms monthly to ensure their proper operation.

So, in conclusion, give us money to fund our startup. And also:




  1. Freshmen year I lived in Gates. During my year there the fire alarm went off 3 times because of people being in a shower, 2 times due to popcorn being burnt and a couple other times due to people burning food on the stove top. It truly was frustrating having to stand outside in the rain or the cold waiting for the fire department to show up just to tell you nothing actually happened. Due to these reasons I can definitely see myself potentially investing in the device as it would not only save time but this way I wouldn’t have to suffer due to a stupid malfunction or miss reading.
    The controlled variable is whether or not the alarm goes off. Based on what was said in the article this makes the most sense since it would be hard to control anything else. The manipulated variables seems to be the way that the readings from the alarm are interpreted. In other words, if the input to turn on the alarm is followed through or not. The disturbance variable is the air either in the sense of smoke or current which causes the alarm to go off. While the process makes sense my worry would be that with a motion activated machine, the alarm could accidentally be turned off, at least initially, by a person waving their hand in panic which could cause late arrival of the fire department and also harm to people, so I believe a button system would be more effective as it reduces the human disturbance that also needs to be accounted for in a motion detection device.
    While the system might be a little more expensive I still see strong value in the application of this device. I believe that socially the concept of safety is very important and as mentioned fire alarms save a lot of lives. Due to this even if it is slightly more costly, the value of what it brings to the table outweighs this issue and it could potentially be found in many houses in the future.

  2. Making a better smoke alarm is absolutely a useful product. When I’m at home, the smoke detector in our front hall often goes off when someone is cooking, even if there’s relatively little smoke. It gets to be quite a pain when you’re cooking something delicate and need to run to the front of the house every five minutes to wave a fan at the smoke detector to get it to stop going off. In addition, having the smoke alarm go off for the smallest thing would probably result in people starting to ignore it, assuming that it is only someone taking a shower or someone burning popcorn, which in a real emergency, could be deadly. All in all, a better smoke alarm that won’t go off after a false positive can only be a positive change for everyone.

    On the whole, this control scheme seems reasonable. Having the alarm go off based off of two sensors should give it a wider range of detection, and having a user input would certainly help prevent the annoyance of having a loud alarm go off in the middle of cooking when there’s no real danger to be had. One thing I would suggest to add would be some sort of monitoring function for the power system, as the article mentions that a sudden change in current could cause it to go off (potentially add another cancel signal if the monitor notices a surge in current), and that a dead battery will cause it to not go off even if it is needed. Since many people don’t change the smoke detector batteries as often as they should, this would be a very user-friendly function to have some sort of reminder to change the battery if it begins to run low.

    On the whole, this system seems like a product that would work, however, I see a couple of problems with it. The first would be issues like the cost of the system and people’s momentum towards not changing something until it breaks. Depending on how expensive this system is, people may or may not have an incentive to go about replacing all of the smoke alarms in their house with this new system to fix something that is, in the grand scheme of things, a relatively small annoyance caused by false positives. In addition, unless a person experiences a smoke alarm not going off when it should have, they’ll generally consider what they have to be “good enough,” unless they are particularly safety-conscious and have the money to spare on a whole new system.

    The one selling point I see of this system is the ability to easily turn off the alarm before it begins blaring. However, the article states that any signal from the ionization sensor will override any human interaction. While I do agree that certain inputs from the sensor should be able to override the human element, perhaps saying that getting a signal from the ionization sensor means that the human is not in control of the situation is generalizing too much. Presumably the ionization sensor would still be set off by someone burning food while cooking, and without any way to abort the alarm sequence, the alarm will still go off, causing the same annoying problem that this device is trying to fix. I personally don’t know too much about how ionization sensors work, but if there was a way to set it to go off after receiving a certain level of signal, perhaps this issue may not be so big of a deal.

    On the whole, this seems like a well thought out product that there is probably a market for, there are just a few quality of life issues with it that may minorly impact its ability to be picked up quickly by the general populace.

  3. I put this article as my first choice, because I expected a wonderful piece from Rojhae, and he delivered (Wait for it ……… psych!!!!!). Actually, this write up was very simple to understand and explained the concepts for a fire alarm very well; I enjoyed it. As you stated, and as many people, recognize, fire alarms are very important household devices which save lives every day and prevents loss of property. However, as nice as this devices can be, they do often go off inappropriate times, as we’ve all probably experienced (I lived in Ruef during my freshman year, trust me, I understand this!!!) The idea for a method to prevent this inappropriate alarm is one which could be very useful, and could save people cost, time and also save from unnecessary deafening sounds.

    The proposed system of preventing inappropriate alarm involved manipulating human signal to control whether the alarm went off or not. Your team also listed the disturbance variable as the quality of air. I think the proposed controlled variable is quite appropriate, especially considering the fact that you’re trying to model a control system for a rigid alarm system whose components you have little control over, however, another potentially manipulated variable is the supply of air to the fire alarm. Since the fire alarm can’t go off unless it is taking up bad quality air, then it also wouldn’t go off it it’s not taking up any air at all. Thus, when a person before a person starts cooking, they could cover up the air inlet of the fire alarm and then open it up when their done. This could be a cheaper alternative to a motion signal. One problem with this suggested manipulated variable is that the person cooking could forget to open up the air inlet when they’re done and this could be dangerous. To prevent this, the air-tight sealed formed around the inlet should have an alarm/timer on it, so that once the time is up, it retracts by itself and if the cook needs more time, they could seal the air inlet once again.

    The disturbance variable listed doesn’t really apply to the situation in my opinion. The air quality is probably not a very appropriate disturbance variable in this application. If the air quality is good, the alarm won’t go off, it it’s bad, the alarm would give a warning first; it’s that simple. The degree of contaminants in the air doesn’t really matter as long as the detectors can detect something wrong. A possible disturbance variable could be lack of attention of the cook when the warning alarm goes off; that directly perturbs the efficacy of the control loop, since human input is needed.

    The idea is a very practical one. It’s relatively inexpensive and I think it should be pursued further. Future applications could extend from cooking to indoor smoking (yeah Rojhae, we all know that’s why you’re so caring about fire alarms), and even to bathrooms (steam from showering, also offsets fire alarms often).

  4. I think smart smoke detectors are extremely useful and relevant in today’s society where people are upgrading to smart home devices. These novel multi-sensor detectors are highly efficient than the standalone old school smoke detectors. The inclusion of photoelectric and ionization sensors and there capability to be controlled remotely by a smartphone are there most useful applications. Typical smoke detectors in contrast to the smart smoke detectors are a required nuisance. Checking batteries every six months, shrill chirping alarms going off every time you burn some popcorn that no matter how frantically you wave the towel you can’t get it to turn off, and the random beeping that comes from a dead battery and you can’t find it and it won’t go away even though you’ve checked every battery in the whole house.

    Understandably, anyone would hate the false alarm going off every now and then, and would want to consider other options to make life simpler in this fast paced world.

    The control scheme that governs the smart smoke detector operation seems practical and makes sense. Smart smoke detectors have fewer false alarms due to the presence of two sensors that detect both smoke and carbon monoxide. The ability to manually turn off an alarm sequence in the case of a false positive with the help of pressing the button or making use of the motion sensor seems to counteract the disturbance variable mentioned- the air quality. For serious situations, the reliance on the ionization sensor to trigger the alarm when help is warranted also seems logical.

    Although a decent job was done in identifying the controlled, disturbance and manipulated variables but from a control vernacular, it might have been helpful if you guys had delved a little into what kind of control loop would the smart smoke detector incorporate. It would have helped to better understand the functioning of the smart smoke detector. Even though I realize that the smart smoke detectors have many convenient features, but from an economical perspective these come with a higher price tag and there is a gap that needs to be offset when considering launching this product-‘cost’. Most people won’t be willing to spend a huge sum of money~ $130 in order to install these smart smoke devices inside all the bedrooms and in hallways outside of bedrooms compared to typical smoke detectors that cost only $10. In addition to the disturbances mentioned, an extremely important unmeasured disturbance variable that I can think of using these smart detector devices that was hinted upon is the loss of internet service, power outages, and an unreliable Wi-Fi signal which can mean that the smart smoke detector device won’t be able to send emergency alerts to the phone when you need them the most. Detectors in essential areas may also not function if they happen to be in Wi-Fi signal weak spots. Moreover, these smoke detectors alerts the user’s smartphone when the alarm goes off (or when batteries is low). They are also connected to each other through a network, meaning when one alarm sounds, all of them do. It can add a layer of safety when there’s a fire in one part of the house, but that doesn’t necessarily means that the users can’t experience false positives during the worst times especially when they are sleeping. The time of 30 seconds for user to react to the alarm is also sometimes not sufficient enough.

    Overall, the product is market-worthy in today’s society and is of growing interest to smart home owners. However, considering the price that’s associated with smart smoke detectors, it is a major constraint for global users.

Leave a Reply

Your email address will not be published. Required fields are marked *