nick
par Nikhil KolekarHello everyone. Welcome to this presentation on our engineering project titled: IoT-Based Intelligent Vehicle with Automated Path Correction System for Enhanced Safety. This project explores the integration of IoT, edge computing, and sensor-based automation in vehicle navigation and obstacle avoidance.
With the rapid advancement of intelligent transportation systems, vehicle safety remains a top concern globally. Human error, distractions, and delayed response times are major contributors to road accidents. To address these issues, we developed an IoT-enabled autonomous vehicle prototype that features real-time obstacle detection and automated path correction.
Our goal is to demonstrate how affordable hardware, combined with edge computing, can significantly enhance vehicular safety even without relying on cloud services or GPS navigation.
Despite existing safety technologies like GPS-based navigation and lane detection systems, current implementations face several challenges. These include dependency on network connectivity, limited sensor coverage, high costs, and latency due to cloud processing.
Our project seeks to address these limitations by creating a cost-effective, multi-directional obstacle detection and correction system using ultrasonic and infrared sensors, all processed locally via a microcontroller.
The key objectives of our system are:
Real-time detection of obstacles in all four directions.
Automated steering and speed adjustments for path correction.
Edge computing for instantaneous decision-making.
Seamless integration of sensors and actuators.
A scalable and low-cost platform that can evolve with additional features like artificial intelligence and vehicle-to-everything communication.
Our vehicle is built on a two-wheel drive chassis, powered by two DC BO motors and controlled using an Arduino UNO. The system includes ultrasonic sensors for mid-range obstacle detection, infrared sensors for precise close-range detection, and an L293D motor driver for motor control. A dedicated power system ensures isolation between the motor and microcontroller supply.
The software is written in Embedded C using the Arduino IDE, supported by the NewPing library for non-blocking sensor operations.
The system continuously reads sensor data, processes it using edge logic, and makes real-time decisions on whether to move forward, stop, reverse, or turn.
The decision-making algorithm operates in a continuous loop. If no obstacle is detected, the vehicle moves forward. If an obstacle is within fifteen centimeters, the vehicle stops, reverses, and takes a turn. The logic ensures that sensor data is filtered, and the actions taken are based on distance thresholds.
This mimics the behavior of intelligent systems that react like human reflexes — fast and reliable, with minimal delay.
The implementation involved assembling the chassis and mounting the sensors, connecting the ultrasonic sensor to analog pins A1 and A2, mapping motor control to digital pins four to seven, coding the logic using the NewPing library to avoid blocking operations, and testing in various indoor obstacle courses.
We calibrated turn delays and distance thresholds to ensure optimal response and stability.
We conducted a series of tests, including unit testing for individual sensors and motors, integration testing for the complete control system, and functional testing for real-time obstacle detection and avoidance.
The robot successfully navigated environments with up to ninety-five percent accuracy in avoiding collisions. The response time was under five hundred milliseconds, and the system ran for about thirty minutes on a nine-volt battery.
Despite its success, the prototype has certain limitations. These include a lack of side-mounted sensors, fixed turn angles, and no predictive behavior — the system reacts but does not anticipate.
Power management and occasional sensor noise were also challenges, which we mitigated through calibration and filtering.
Potential improvements include adding infrared sensors on the sides, implementing pulse-width modulation motor control for smoother navigation, integrating Bluetooth or Wi-Fi for remote monitoring, and using machine learning for predictive obstacle detection.
This would take the system closer to semi-autonomous or autonomous vehicle capabilities.
In conclusion, this project demonstrates a practical, scalable, and affordable approach to vehicle safety using IoT and edge computing. By leveraging simple sensors and microcontrollers, we've created a prototype that autonomously detects and avoids obstacles in real-time.
Our solution is independent of cloud services, GPS, or expensive components, making it highly suitable for both academic research and practical deployment in low-cost vehicles.
Thank you for watching this presentation. We hope this project inspires further innovation in intelligent transportation systems. If you have any questions, feel free to reach out.
