# PROJECT IDEA:
The system idea is to detect accidents on the road and the probability
of it happening in the future through analysis, and in case of
occurrence of such an event this will inform the appropriate
authorities and other interested parties with real-time information.
also detect severity automated.
# AI Model Used:
-We used the CNN model
to indicate whether an
accident occurred or not.
-We used the YOLO
model to detect Severity
of the accident
# STAKEHOLDERS:
- Normal user
-Tracker
# FUNCTIONAL REQUIREMENTS
-Ability to detect accidents in real-time
using various sensor inputs such as
cameras.
-Ability to inform the appropriate
authorities by the tracker (take action).
-The ability to generate real-time traffic
flow data, including traffic congestion
and delays, to help manage traffic and
reduce the risk of accidents occurrence
(GPS API).
-The ability to store, retrieve and provide
accidents data reports and statistics to
authorities and other interested parties
for future analysis and reporting.
# NON-FUNCTIONAL REQUIREMENTS
-Performance
-Reliability
-Security
-Maintainability
-Interoperability
-Usability
-Portability
-Availability
# USED SOFTWARE DEVELOPMENT LIFE CYLCE (SDLC):
-In our project, we have adopted the Agile Scrum methodology
for the development of AI-powered solutions in safe
transportation for smart cities.
-By embracing Agile Scrum, we delivered high-quality results,
foster collaboration, and adapt to evolving requirements
throughout the project lifecycle.
# Security:
-JWT (JSON web Token) is created
after authentication.
-JWT is used to authorize users.
-We apply validation to our token to
complete the user's request.
-Each token has expiry date and
revoke attributes.
-We used HTTPs for further security .
-We have white- and blacklists for APIs
to restrict access of users based on
authority and authentication status.
# DELIVERABLES:
-Real-time accident detection with
its severity and optimized traffic
flow are the key outcomes of our
AI-powered safe transportation
solutions for smart cities.
# Recommendations:
-Proactive Incident Detection:
Developing advanced AI models to
predict and detect potential accidents in advance, enabling preventive
measures and timely interventions.
# USED TECHNOLOGIES:
- Angular - YOLO
- Router - TensorFlow
- Spring boot - MySQL
- Python - Flask
- Microservices
# Dataset:
-We used dataset provided by
DELL