Below are some ideas for topics that you can work on as part of a project in a course, as thesis or as an internship.
Graduate students working with me should take my lecture "Mobile Robotics". Students should have good programming skills in C++ or be willing to learn programming quickly. Unfortunately, I do not speak Chinese (yet), so you must be comfortable speaking English with me - do not worry if you think your English it not good enough - you will get better through practice :) !
Almost all projects will require you to work with the Robot Operating System (ROS) - you can program using Python and C++.
Excellent students will present their results in publications at international conferences or participate at international RoboCup Rescue competitions.
If you are interested in any of the topics, or have an idea for your own topic, please come by the Robotics Lab or write me an e-mail: soerensch (at) shanghaitech.edu.cn.
- 3D Map Evaluation:
Collect 3D scans and create ground truth maps. Develop algorithms to compare robot generated maps with the ground truth map. Basically continue the work outlined here.
- Hierachical Topological Maps for map matching, map merging, path planning and map evaluation:
This project will continue the work presented in this paper as well as the work by Yu Tianyan.
- Small Robotics: Program Android phones to control our small robots. Build and modify small tracked robots. Implement autonomous behaviours for the robots. Implement visual SLAM on the phone (using GPU programming).
- Autonomous Mapping and Navigation with the Jackal Robots:
Mount our 2D LRF sensors and IMUs on the Jackal robots and use available open source software to have them mapping and autonomously navigating and exploring.
- Programming of the Shunk 6-axis arm:
Use ROS Moveit! to program the arm to plan and execute trajectories while avoiding obstacles.
- Car Robotics:
Connect to the OBDII connection of a car to gather all kinds of status data, including wheel odometry, and feed that data back to ROS. Also install GPS receivers on the car and record their data in ROS.
- Car Mapping:
Install our 3D Velodyne scanners on a car and - in conjunction with the above project - map ShanghaiTech (including the awesome underground structures) and its surroundings.
- 3D Mapping:
Review, test, improve and develop datastructures and algorithms that can hande huge amounts of 3D data for online mapping, visualization and storage.
- Offline 3D Mapping:
Word with our extensive dataset (more than 20 billion colored points of the SIST building) on problems like registration, feature extraction, semantic annotation, map evaluation, planning or rendering.
- Localization using 3D Features:
Implement components for a generalized framework for robot localization using visual features that are localized in 3D space.
- Control of Micro Aerial Vehicles (MAV) using the tracking system.
- Localization of a Micro Aerial Vehicle (MAV) using (Stereo) Vision:
Use cameras to find a micro aerial vehicle from the ground. Determine the 3D pose of the MAV (localization) and track it in consecutive frames.
- 3D Mapping using Structure from Motion on Localized MAVs:
Use the video feed from a micro aerial vehicle to create a 3D map. This makes use of the localization from the previous topic. A structure from motion algorithm is applied to create dense point clouds.
- Multi-Robot Coordination:
Develop algorithms to coordinate a group of robots (e.g. 5 TurtleBots and/ or 5 Clearpath Jackals) to complete exploration or search tasks.
Mount a laser projector on a mobile robot and explore how users can interact or play games with the robot this way.
- Choose your own Topic!
If you have an idea for a work topic that is related to robotics, just contact me via e-mail or visit me in my office to find out if we can make a project out of it!