Abstract:
The evolution of robotics and automation technologies has spurred innovative
advancements in autonomous vehicles. The hospitals lost most of its staff during
pandemic when they needed them the most and the biggest reason for that was that the
virus was contagious and hazardous to life. We came across this idea which aims to
contribute to this evolution by developing an autonomous obstacle avoidance system
for a four-wheeled vehicle. The system integrates Ultrasonic (US) and Infrared (IR)
sensors for comprehensive obstacle detection, a Feed-Forward Neural Network for
real-time decision-making, Bluetooth communication for interaction with a dedicated
Flutter mobile application, and the potential for future path planning integration. The
project's foundation lies in the Sensor Integration Module, where US and IR sensors
are strategically positioned to detect obstacles in the vehicle's vicinity. The Neural
Network Module involves extensive back propagation training to derive accurate
weights, enabling the neural network to process sensor data and generate motor
activation commands. The Decision-Making and Control Module translates these
outputs into specific motor activations, guiding the vehicle to navigate in response to
obstacles. Seamless interaction between the vehicle and users is achieved through the
Bluetooth and Mobile App Integration Module. The Flutter mobile application
provides real-time tracking, performance metrics, and user input capabilities. The
robot utilises several sensors to gather and subsequently communicate data from its
surroundings to a central computing core. This core operates a mapping programme
that utilises the acquired data to generate a map of the environment.
