My Detailed Projects
Apple Counting Using Machine Learning
We have employed two machine learning techniques to accurately count apples in an orchard, streamlining agricultural processes. The first method used YoloV5s, a state-of-the-art object detection algorithm, and the second method was a conventional approach using a blob detection algorithm on top of a mask created using HSV colour space. The project aimed to compare the accuracy and efficiency of these two approaches. The results showed that YoloV5s performed better in accuracy and speed than the conventional image processing method.
Socially and Physically Assistive Robots in Human-Robot Interaction
Conducted an in-depth analysis of socially and physically assistive robots, examining their impact on human-robot interaction and potential applications in various settings. The experiments demonstrated the potential of these robots in improving the quality of life for individuals with disabilities. Notably, the socially assistive robot, Pepper, was able to assist with daily tasks, and a TurtleBot3 robot successfully delivered medication at specific times, showcasing effective human-robot cooperation.
Path Traversing
Our mobile robot was programmed to autonomously survey the length of a pipeline in the ocean, find a breakage, and return to its drop-off point. This was achieved using the Pololu 3Pi+ mobile robot, demonstrating advanced autonomous navigation and problem-solving capabilities in challenging environments.
Investigating Odometry Errors
This study focuses on the errors in odometry caused by collisions, analyzed through experiments with our improved robot incorporating accelerometer and encoder-based kinematics. The findings contribute to better understanding and mitigating collision-induced inaccuracies in mobile robotics, enhancing reliability and performance in autonomous navigation.
Nao as a Mediataion Assistant
A user study was conducted involving two experimental conditions with NAO robot as the meditation assistant and another with a smartphone app guiding a video. The study sought to assess the impact of the robot on participants’ emotional state and stress levels.
View the Report