Winning the 2017 AutoMobility LA Hackathon

By Ardy Rahman, Bryan Norden, Chris Loncarich, Simon Tsai & Kelly Moore

The 2017 AutoMobility LA conference, brought together automakers, and tech leaders to discuss the future of transportation. The conference also featured a Hackathon presented by Honda Innovations (from Nov 27–28).

The competition challenged teams to build a solution to solve LA’s traffic congestion issues during major events, using the 2018 Honda Odyssey as a test vehicle. With the 2028 LA Olympics coming up, solutions from this Hackathon would be key in easing common mobility issues that drivers and passengers face in LA.

TribalScale is no stranger to Hackathons. We’ve participated in and hosted multiple across different industries and platforms, including the first Amazon Echo Hackathon in Canada, the enLIGHT HACKference, and the HackuWeather Hackathon that we co-hosted with AccuWeather. Naturally, we decided to take the plunge and participate. Most of us work out of TribalScale’s OC office and know firsthand how terrible traffic can get in LA.

We spent 24 hours competing against some really sharp engineering teams and ended up winning first place with our predictive accident response platform!

To learn more, read the full Press Release of the AutoMobility LA Hackathon here.

Here’s how it happened…

author

How it Started

author

Our team consisted of five members. Three Agile Engineers: Bryan Norden, Chris Loncarich and Simon Tsai. Ardy Rahman was our Product Manager/Data Scientist and Kelly Moore was our researcher. There were over 100 participants, with our team being one of 17 teams who were all there to win the grand prize of $7,500.

All teams were given access to Honda’s vehicle APIs, as well as APIs and SDKs for augmented reality, mapping, points of interest, and payments technologies.

We had a lot of data to work with. Making sense of which data we wanted to use was important, as we only had 24 hours to strategize and build our solution.

Our Winning Solution

Most teams at the Hackathon were working on making apps to help people find parking around LA or develop augmented reality (AR) apps with info about the car or city. However, we were really intrigued by all the traffic accidents and traffic deaths on the streets of LA and wanted to do something to help with the City of LA and Vision Zero’s goal. We knew we had to create a solution that would help the city ease into their goal because it doesn’t just happen overnight.

After a long night, little sleep and a lot of caffeine, we built out a solution we were all very proud of. Our solution was a predictive assistance platform that would allow the City of LA to anticipate and respond to traffic incidents within 30 seconds — a dramatic decrease from current response times (6 mins). Our predictive platform consists of three native apps: an iPad app for the City’s Dispatchers, a Responder or Assistant iOS app for iPhones, and an in-car iOS Helper App for drivers.

We analyzed the City of LA’s traffic collision data from the past four years, while also doing our own data collection on LA sports events from the same time period. Using both sets of data, we created a model that could predict which street and intersections were most likely to experience an accident when given information about a specific sporting event such as: time of day, expected attendance, location, weather, and venue. With this data the City of LA would be able to place response units in those areas to reduce traffic accidents and keep cars moving on the roads.

“Real innovation comes from refining data to make it useful, and layering it with an elegant technical solution to help solve real problems. This is our bread and butter, and this competition perfectly captures the spirit of that.” — Ardy Rahman, Lead Product Manager on TribalScale’s Hackathon team.

author

The Dispatcher app shows a dynamic heat map of areas in the City where accidents are likely to happen during a major sporting event while taking environmental variables into consideration. The Assistant app will alert collision assistants, assigned to designated Assist zones in the City, when a vehicle-related incident has occurred. By interfacing with telematics from the vehicle, the Helper app for the end user/driver will automatically notify the other two apps in real-time, once a vehicle collision or incident is detected. The Dispatcher and Assistant apps are notified without being prompted by the user, which helps to cut down response time, especially during life-threatening incidents.

Trained professionals will be able to arrive on the scene within 30 seconds to help assess the incident and determine which first responder resources are required, if any. Small collisions and incidents, like a flat tire, can easily be mitigated by these low-cost and agile Assistants while allowing first responders to attend to more serious collisions.

Let’s go through a scenario of how our platform will work.

Our platform can tap into the car’s onboard sensors to detect collisions and other vehicle anomalies that warranted attention. All the driver needs is the mobile app and a car with the appropriate sensors in order for our platform to know when something is wrong with the car.

If a driver gets into a car accident, the app will detect that accident based on the car’s collision sensors. Then, it will determine the severity of the accident by using the airbag deployment sensors and other vehicle data, such as speed. The app will automatically send the driver’s current location and crash details to the City’s Dispatcher app.

The dispatcher will then notify the nearest response unit to go to the scene of the incident within 30 secs. If it was a serious accident, it would automatically dispatch the Fire and Police Department.

author

All three apps work seamlessly together to reduce traffic and get drivers the assistance they need quickly, while also allowing the City’s Dispatchers and Responders to be more proactive during traffic incidents.

Our solution will help the City of LA become anticipatory, rather than reactionary, when it comes to dealing with a range of vehicle-related incidents. Currently, it takes first responders an average of six minutes to arrive on the scene of a collision. For every minute it takes a responder to arrive, the likelihood of mortality increases between 7–10%. By using our predictive model, the City of LA can manage its resources throughout the city more intelligently, and make the roads safer for visitors and locals during major sporting events.

author

We can’t believe we took the Grand Prize home! It was an honor to be presented the award by the CEO of Honda Innovations himself, Nick Sugimoto. The whole experience was both exhausting, but exhilarating at the same time.

author

Join our fast growing Tribe and connect with us on Twitter, LinkedIn & Facebook! Learn more about us on our website.

Visit Us on Medium

You might also be interested in…