DEV-ON

College Name: MIT Academy Of Engineering

Team Name: House_Stark

​

Team Leader Name: Ronak Nikam

Theme: Computer Vision based modular ADAS

Team Members Name: Sanskar Waghmare, Abhishek Jawade

Problem Statement

India boasts nearly 70 million four-wheeled vehicles, yet a mere 10% incorporate optional ADAS features. Financial constraints and lower spending capacity hinder ADAS adoption into Indian households. Despite automotive advancements, a transformative modular solution for universal entry-level ADAS remains absent.

Data from Indian ministry of Road and Transport and Highways report on “Road Accidents in India 2021”

Solution

To tackle this issue, our aim is to create an affordable, universally adaptable camera-based Advanced Driver Assistance System (ADAS).

This innovative project focuses on crafting a modular device capable of seamlessly integrating vital safety features like lane-keep assist and driver drowsiness detection into any vehicle by retrofitting it on the vehicle’s dashboard.

Working & Techstack

Working-

Camera Inputs:

1. External Cameras: CNNs make it possible to detect blind spots and provide lane assistance.

2. Internal Camera: Detects driver fatigue using facial recognition using RNN or LSTMs.

​

Information Processing:

1. Computer Vision: CNNs use external feeds to detect lanes and potential threats.

2. Driver Monitoring: Using internal camera video, AI evaluates sleepiness.

3. Hardware: The computing will be done on a Raspberry Pi 3 or 4, with enough compute capabilities.

​

Quick Reacting actuators:

1. For lane-keep assist, an external powerful brushless DC motor will be used and will be directly mounted on the steering shaft.

2. A HUD display will display the proximity and position of the vehicle in driver's blind spot.

Techstack-

Python programming : OpenCV.

AI Frameworks: Raspberry Pi optimized models can be used with TensorFlow Lite.

Integration of Cameras:

Hardware: The Raspberry Pi can be used with both internal and external cameras.

Software: Raspberry Pi libraries for video processing and camera interface.

Edge Technology & Implementation:

Edge Device: For local ML model execution, use a Raspberry Pi 3 or 4.

Deployment: Flask/Django for the edge device backend, Docker for containerization.

Uniqueness of your idea

Modularity: The gadget increases safety by integrating ADAS capabilities into any type of vehicle.

Affordability: With the use of Raspberry Pi technology, enabling more owners to access safety. Without persistent internet access.

Localized AI processing: Guaranteed real-time decision-making.

All-encompassing safety protection featuring both internal and external cameras to identify potential hazards and analyze driving behavior.

Innovation in ADAS retrofitting for older cars without integrated systems.

MarketPlace & Impact

Market:

This idea aims to close a sizeable gap in the market for automobile safety. The market is made up of a wide range:

Owners of Vehicles: Fleet management firms and private drivers alike are looking for reasonably priced safety upgrades.

Automotive Industry: Possible joint ventures with automakers or suppliers of aftermarket accessories.

Safety Solutions Sector: Providing an exclusive, customizable ADAS module can establish a niche market within the larger safety solutions sector.

​

​

Possible Impact:

Improved Safety Standards: A greater number of reasonably priced ADAS installations can greatly increase general road safety and possibly lower the number of collisions and fatalities.

Accessibility: This project may make advanced safety features available to a larger group of car owners by providing entry-level ADAS functionality.

Industry Disruption: By offering affordable safety enhancements for current cars, a modular, retrofit-friendly ADAS system has the potential to upend the current industry.

Economic Impact: Lowering the cost of safety features may have an impact on consumer choices, increasing the demand for ADAS-equipped cars and the desire for retrofitted safety measures.

Industry Collaboration: We may be able to expand the integration of your technology into the automotive market by partnering with aftermarket suppliers or auto manufacturers.

​

The impact of this initiative goes beyond the safety of a single vehicle; by facilitating the accessibility and adaptability of sophisticated safety measures, it may influence industry standards and market dynamics