Scene understand and recognition for unmanned vehicle

• Main Author: Or, Joel Hao En
• Other Authors: Wang Dan Wei
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Engineering; Robotics
• Online Access: https://hdl.handle.net/10356/176051
• Institution: Nanyang Technological University

This study investigates the challenges and advancements in scene understanding and recognition for unmanned vehicles operating in rainy weather conditions. This is especially true for Last Mile Delivery (LMD) of food delivery robots since there are safety challenges coupled with reduced resolution and accuracy of sensors from rainy weather. It poses a significant obstacle to visual perception systems, impairing image quality and hindering accurate scene interpretation. Through a comprehensive review of recent research and developments, this paper explores different techniques and characteristics of different sensors and algorithms to best suit the capabilities of unmanned vehicles in rainy environments. We will first source for sensor and shelter for the sensor for data collection during rainy weather. After data collection, we will do data extraction of the recordings to evaluate the images. The findings presented here contribute to the ongoing efforts in advancing the reliability and autonomy of unmanned vehicles and robots for application such as transportation, in challenging weather conditions.