I am advancing a mobile robot so it can reliably detect objects and manipulate them by combining the power of ROS (Ubuntu 22.04, ROS 2 Humble) with Matlab R2023b. The robot already moves; what I need now is the full perception-to-action loop: • Sensor fusion that brings together the on-board LIDAR, RGB-D camera, and ultrasonic array • Matlab-based training and evaluation of the detection network (Deep Learning, Computer Vision, Robotics System Toolbox), with real-time inference streamed back into ROS • Grasp or push planning through MoveIt! or an equivalent manipulation stack, closed in hardware tests Because the ultimate goal is a peer-reviewed, Scopus-indexed journal article, every line of code must be reproducible and every experiment logged. Alongside the working software I will need: 1. A clean ROS package and Matlab scripts that build with no warnings and demonstrate <50 ms inference latency 2. A rosbag plus short video showing ≥90 % detection accuracy and ≥80 % successful grasps under lab conditions 3. A LaTeX manuscript draft that follows the target journal’s format, contains comparative analysis with at least two recent papers, and incorporates the generated figures and datasets If you are comfortable moving fluidly between ROS nodes and Matlab algorithms—and enjoy writing up solid experimental results—this project should be a perfect fit.