SIREN

Sounding the alarm on school shootings by detecting gunshots, streamlining communication, and reducing response times.

180 school days x 18 school years = 3240 possible days for a student to undergo a school shooting. School shootings are a significant public safety issue that has been spiking in occurrence throughout the past couple of years. In 2022, the United States recorded a record high of 250 school shootings according to the K-12 School Shooting Database [1]. Most schools spend time out of academic education during the school year to practice lockdowns in case of a shooter [2]. Despite this pressing matter, there exists no thorough advancement that mitigates school shooting casualties and offers students a reassuring sense of security in a supposedly safe environment.

The Police Executive Research Forum notes that during an active shooting, the first call to local law enforcement happens three to five minutes after the first shot is fired. This late notice leads to inaccurate and incomplete information, limiting police and EMT’s ability to effectively respond and save lives [3]. Moreover, the lack of effective communication between key parties limits the ability of those at school – students, faculty, and administrators – to make the best decisions to escape to safety.

As high school students, we could not let more innocent students be targeted. By focusing on reducing the response times for both law enforcement and school community, we restlessly developed SIREN to sound the alarm on school shootings.

By reducing response times in active school shootings, SIREN is a cost-efficient innovation that protects students’ lives and school communities, especially those at financially disadvantaged Title I schools. This life-saving technology can be housed in a compact 3 inch by 6 inch case (resembling a smoke detector) and can be easily installed on ceilings using only four screws. 

The device utilizes an omnidirectional microphone to detect gunshots and seamlessly communicates this information to registered users through calls and texts. Users – students, parents, and school faculty – can register for these alerts by obtaining specialized codes from partnered schools. 

Once the audio gathered by the microphone is classified as a gunshot through the use of advanced machine learning algorithms, the device – powered by a Raspberry Pi – will instantly send text messages to students and parents, as well as calls to law enforcement and school faculty. Additionally, the location of the shooter will be indicated on an uploaded school map through text, based on the frequencies picked up by multiple sensors placed throughout the school.

SIREN serves not only students, but parents, teachers, administrators, and law enforcement in order to keep them safe from school shootings. SIREN aids law enforcement and administrators in locating the shooter through calls while students, parents, and teachers are all informed of the situation through text.

There are detrimental mental health and educational consequences for those who are affected by school shootings. Antidepressant use increases by 21.3% according to the Stanford Institute for Economic Policy Research. Long term school absence also increases by 27.8%, which could mean students fall behind in their education and would have to repeat a grade [4]. SIREN aims to minimize damage in case a school shooting occurs.

All four of us were victims of false school shootings that plunged our campuses into lockdown. Although our experiences were only scares, others may not have been as lucky. Recently, multiple gunmen injured six adults at Oakland Tech, a high school less than an hour from our own.

SIREN was founded by four high school students jointly motivated to make an impact on the safety of students across the nation, and MIT SOLVE is our opportunity to showcase our potential impact and to immerse ourselves in the world of innovators we admire. 

Caitlin is experienced in hardware and CAD, so she focused on deploying the gunshot model and 3D-printing the case for the microphone and motherboard. Rebecca and Swarnya have pursued machine learning research and perfected the gunshot detection algorithm through Edge Impulse and worked with the Twilio text and call APIs. Audrey is familiar with business models and the web development process, and focused on creating a user-friendly website connected to the Airtable API.

Before finalizing SIREN, we surveyed students, parents, and school faculty of schools from our four school districts spanning throughout the San Francisco Bay Area. 

From all of the elementary, middle, and high schools we asked, there was an overwhelming consensus that an additional layer of protection must be implemented in schools. The survey also indicated that a gunshot detection system would be preferred over detection through manual surveillance of students, which invades students’ privacy and is costly. Camera surveillance systems can cost at least $400, on top of installation and maintenance fees. Another costly option is employing security guards which often results in an authoritarian atmosphere for students. Historically, the presence of security guards has contributed to bias against marginalized students, especially the BIPOC community.

In addition to our website linked above, please view our full 3 minute product video. 

SIREN is the first innovation to streamline gunshot detection and cater automatic communication to all affected parties in two different ways: text and call. We plan to patent SIREN under the micro entity status after our initial funding stage in which we focus on further testing, research, and development.

Currently, there are no effective solutions that offer escape assistance to students during an active shooting to stop the problem before permanent damage is done. Solutions from companies Guardian and ShotSpotter prove to be unaffordable to the most vulnerable schools (Title I) and lack communication with the school community. ShotSpotter faces controversy regarding transparency and privacy concerns as well, since its detectors retain hours of continuous audio. In particular, these companies deploy arrays of acoustic sensors that ultimately rely on their companies’ personnel to differentiate between gunshots and non-gunshots, leading to costly minutes wasted. Additionally, none of their products have a transparent display of costs.

SIREN solves just that by emphasizing our “Triple A” core values: accessibility, affordability, and accuracy. We stand out from the competition since our product is low-cost and easy to use. SIREN is non-intrusive, and the audio data gathered is not stored. The audio gathered takes less than a second for the machine learning model to determine whether there has been a gunshot or not, which instantly triggers SMS text and call communications.

In addition, SIREN has faster notification response times, a larger coverage area for a single unit, and product scalability in schools, as compared to our competitors. Our company’s values and how we reflect them in our product makes us stand out in the current market.

Our goals reflect our “Triple A” core values: accessibility, affordability, and accuracy. 

Accessibility: Through the entire design process, we have looked for ways to empower the most vulnerable of populations, especially those who come from low income, Title I schools. Since notifications are sent by text message as opposed to through a separate app, the complexity of setup is minimized. Users do not need WiFi to receive the notification, allowing for universal compatibility with the SIREN software.

Affordability: We aim to protect as many schools as possible, so we will use funding for further research and development as well as feedback testing. In particular, we plan to partner with at least 5 San Francisco Bay Area schools and manufacture a specialized motherboard and communication APIs to reduce costs. 

 Accuracy: In order to ensure the highest level of accuracy in gunshot detection, we aim to continuously put our deep learning model through rigorous training and evaluation. Our database currently contains over 15,000 audio clips, and by hopefully gaining support from Solv[ED], we hope to increase this number to over 45,000 with live samples of gunshots, fireworks, clapping, and other common false positive sounds. This is crucial for our product to be able to perform effectively in real-world situations and ultimately ensures that all schools have access to dependable and trustworthy products that can immediately respond in the event of a school shooting.

SIREN harnesses the power of hardware and software technologies, including Raspberry Pi and machine learning algorithms. 

First of all, the microphone captures audio in Waveform Audio (WAV) files and feeds them into the convolutional neural network (CNN) deployed on a Raspberry Pi 3 Model B+ for analysis. The CNN identifies gunshots and initiates the communication pipeline (automated texts and calls using Twilio APIs) if necessary. The SMS texts will include a map showing the shooter’s location in the school based on the gunshot detectors’ readings. The SMS notifications allow for universal compatibility with the phones of our clients.

Finally, we developed several iterations of our 3D-printed case of the gunshot detection module, landing on a final 3x6 inch unit to compactly house the microphone, Raspberry Pi, and batteries, similarly resembling a smoke detector. This can be easily and securely installed on school ceilings with four screws and a rotational lock. 

As SIREN is in the prototype phase, we do not serve any customers at the moment.

In regards to our projected outcome in one year, SIREN is meant to be first implemented in San Francisco Bay Area schools. We will partner with our school districts. Upon positive feedback, we will quickly expand throughout the state of California and the nation, taking advantage of the fact that the average total cost of producing a SIREN sensor will decrease as production increases.

Upon reaching our target funding and gaining investors, we will be able to protect over 1 million individuals in school communities in our very first year–and that’s only accounting for the San Francisco Bay Area schools. The students of tomorrow will be able to learn in a safe environment, allowing them to fully reach their potential. Given that a single sensor from SIREN effectively prevents and mitigates school shootings within a single school, SIREN sensors in schools across the country will transform what safety means to American students.

SIREN is a social enterprise whose success is hindered by the lack of financial funding. Despite the pressing need for socially-conscious solutions, banks and financial institutions are often unwilling to support social enterprises, leaving many struggling to secure initial funds and remain in the market. As high school students, we are faced with the daunting reality that our limited financial means impede us from bringing our dream of creating an inclusive and equitable society to fruition. 

However, we refuse to let financial constraints stifle our ambition. We implore Solv[ED] to support us in this endeavor by providing us with the financial means necessary to launch SIREN and bring about positive change. Every investment, no matter the amount, brings us closer to realizing our goal and creating a better world for all.

In the future, as we progress into the research and development phase of SIREN, we anticipate potential legal issues arising from false positives and safety concerns during the testing, and debate over the ethical placement of the sensors. We are committed to ensuring the ethics and accuracy of our technology, so we will work closely with experts to address and avoid any potential risks and contributing to social biases. With the support of Solv[ED], within the next five years, we can bring SIREN to market and make a real impact in the lives of students, families, and entire communities nationwide.

Our business model is business to business (B2B), so we plan to sell units of SIREN commercially to individual schools (typically private schools), school districts, and governments hoping to ensure better preparation for school shootings.

The users will be the students, parents, teachers, and administrations; when a school/district/government purchases SIREN, we will personally consult with them and set up special codes for users to receive notifications (texts and/or calls).

The development of one unit includes a Raspberry Pi Model B+, Twilio SMS text and call APIs, an omnidirectional USB microphone, 3 AA batteries, and a battery box. This totals $105.

Considering a typical high school in the United States has an area of 50,000 square feet and one microphone can detect gunshot sounds within a 50 feet radius, eight units per school would be sufficient to keep the schools safe. SIREN will distribute one free sensor to a list of initial schools. The distribution of free sensors is to promote smoother adaptation of SIREN within the initial schools, who would ideally buy seven or more units following positive reception of the free security protection.

For revenue, if we estimate each unit (sensor and communication alerts pipeline included) costing $105 to produce, including requisite operational costs, we can sell the assembled product from a range of $180 to $220 (a huge profit range of 40% to 50%), depending on the partnership; if a school is affiliated with a partnered district or institution with SIREN, it will receive the discounted price. From a product selling standpoint, this is a major success.

For each sensor, SIREN will include a yearly maintenance fee of $50. The yearly fees will go towards regular software upgrades, hardware tests, and installation checks. With an average of eight sensors per school, we will make a yearly profit of $400 from one school (and subtracting the amount going to manual labor costs).

With increases in demand later on, we plan to invest more into the research and development of sustainable and cheaper components of SIREN such as the Raspberry Pi, as well as additional features that can enhance communication alert time. This will mean we can increase production and earn larger profits.

First, we will raise money to finetune our minimum viable product (MVP) through equity-free grants for research-and-development (R&D) and fellowships with programs like the Federal Commission for School Safety. In addition, we will file for patenting under the micro entity status. This first phase will take $5,000.

After multiple iterations of our prototype, we will show this MVP to potential schools through a demonstration day and sign contract agreements with them.

With this traction, we can then apply to top Silicon Valley accelerator programs like Y Combinator or Berkeley SkyDeck to receive 100k-500k in funding. Through an accelerator, the product will mature. We will hire key personnel for expansion (sales, marketing, engineering) and achieve product market fit. Once we reach the 100k+ ARR (annual recurring revenue), we will raise a seed round from VC firms and angel investors.

The end goal is for schools nationwide to be secured with SIREN.