lineLogic

Your smart pass through TSA—our advanced AI to predict wait times at airports.

Comment

Inspiration

lineLogic's colleague Jeremy, a frequent traveler, often arrives at the airport excessively early due to the unpredictable nature of TSA security wait times. This results in him squandering precious time waiting for his gate to become available. To address this issue, we've opted to leverage technology for a more efficient solution.

What it does

Utilizing advanced deep learning algorithms, lineLogic predicts TSA security queue durations at airports. This innovative technology can also be adapted to forecast wait times in various other scenarios involving lengthy queues.

How we built it

Our team developed a sophisticated visual analysis algorithm using Python and OpenCV, capable of tracking individuals and differentiating between obstacles and separate queues. Throughout the event, we participated in various activities, primarily meal gatherings. To simulate real airport scenarios, we employed a compact 4K drone at a previous hackathon to capture sample queue footage from multiple angles, as actual airport video was unavailable.

The project's technical stack incorporated MongoDB, React, TensorFlow, Scikit-learn, Matplotlib, and several other technologies to create and deploy our website, including its user interface and experience design. Our web platform processes queue footage to estimate approximate wait times.

The primary objective of our website is to assist travelers in efficiently planning and organizing their departures, ultimately enhancing the overall airport experience.

Challenges we ran into

In developing our project, we faced significant challenges in creating 3D objects to enhance the UI experience. These included dealing with complex geometries and mesh errors, optimizing performance for smooth rendering across devices, ensuring cross-platform compatibility, and balancing visual appeal with usability. Our team had to employ advanced 3D modeling techniques and implement fallback options to address these issues effectively.

Analyzing empirical airport data and developing accurate wait time predictions presented another set of obstacles. We grappled with data scrubbing and normalization from various sources, determining the crucial "golden ratio" between our sample footage and real-life scenarios, ensuring diverse and representative sample sizes, and implementing efficient real-time data processing. Overcoming these challenges required extensive analysis, testing, and optimization of our deep learning models to balance accuracy with processing speed, ultimately resulting in a solution that combines immersive 3D visualization with practical, data-driven insights.

Accomplishments that we're proud of

Firstly, we successfully overcame numerous challenges to create an exceptionally clean and intuitive user interface. This achievement reflects our commitment to user-centric design and our ability to translate complex data into a visually appealing and easily navigable format. The result is a platform that enhances user experience and makes wait time information readily accessible to travelers.

Secondly, we developed an impressive deep-learning algorithm capable of accurate human detection. This sophisticated system forms the core of our wait time prediction model, demonstrating our technical prowess in machine learning and computer vision. The algorithm's ability to efficiently process video footage and distinguish individuals in various airport scenarios showcases our innovative approach to solving real-world problems through advanced technology.

What we learned

Our collaborative experience proved invaluable as we embraced the role of full-stack developers. We adopted a flexible approach, alternating between backend and frontend tasks to maintain a well-rounded perspective. This strategy proved particularly effective when confronting challenging issues, especially those related to JavaScript, Tailwind, and CSS formatting. By frequently switching roles, we gained fresh insights into persistent problems, ultimately leading to more innovative and efficient solutions.

What's next for lineLogic

Moving forward, we aspire to partner with a nearby airport, such as BWI, to gain access to real-world insights from airport personnel and obtain high-volume test data. We believe our AI-driven technology has significant potential for widespread adoption across airports nationwide. By leveraging actual airport environments and expertise, we aim to refine and scale our solution, potentially revolutionizing queue management and enhancing the travel experience for passengers across the country.

Built With

Share this project:

Updates