Semester Projects and Thesis Topics
Planning & Optimization | Mobile Robotics | Industry and Research | Finished Projects
We offer topics from various fields of robotics to students of bachelor’s and master’s studies. Our students solve tasks such as navigation, localization, planning for one or more robots, or processing information from rangefinder sensors and standard and RGBD cameras. We work with both simulated data and real robots, from classic manipulators, through mobile robots to drones. In the laboratory, we also have Tiago++ – a two-armed mobile manipulator. The topics are listed so that they can be followed up with a bachelor’s/master’s thesis in the following semester.
Note: The listed topics may be reserved, so always check with our staff. It is also possible to come up with your own work assignment after prior consultation with our staff.
Leaflet with Semester Projects and Thesis Topics
PLANNING & OPTIMIZATION
Contact person: Miroslav Kulich (miroslav.kulich@cvut.cz)
We focus on the fundamental and applied research on planning problems in robotics and artificial intelligence. Two main streams of our research are routing problems and multi-agent path finding. Our solutions combine knowledge from different scientific disciplines: artificial intelligence, computational geometry, graph theory, operations research, combinatorial optimization, and robotics.
Multi-Agent Path Finding
Contact person: David Zahrádka (david.zahradka@cvut.cz)
Imagine a warehouse where hundreds of robots are transporting goods from one place to another. How do they avoid collisions and reach their destinations efficiently? This is the problem of Multi-Agent Path Finding, where the goal is to find collision-free paths for a team of agents. In this project, you can design and implement algorithms for pathfinding in dense environments. Another problem is the development of robust execution methods, which ensure safe and effective execution of planned paths on real-life robotic fleets in the presence of unexpected delays.
Mission Planning in Polygonal Domains
Contact person: Miroslav Kulich (miroslav.kulich@cvut.cz)
This discipline combines tasks and path planning for mobile robots as well as 2D geometry and optimization. The tasks include inspecting a known environment or specific regions within it or searching for a hidden object of interest. The ultimate objective is constructing the most efficient feasible path that enables the robot to complete the selected task in the shortest possible time.
Metaheuristics and Optimization Challenges
Contact person: David Woller (david.woller@cvut.cz)
We design metaheuristic algorithms – custom approximation algorithms used for addressing challenging NPhard combinatorial optimization problems that are beyond the capabilities of exact solvers. Each year, our student teams successfully participate in international optimization competitions from various application sectors, such as truck fleet loading optimization, power transmission network maintenance scheduling, or electric vehicle routing.
MOBILE ROBOTICS
Robotic autonomy encompasses a diverse range of applications and problems. We aim to develop and improve advanced methods in various fields, including autonomous navigation, mapping, exploration, and movement control.
Mobile Robot Navigation
Contact persons: Václav Truhlařík (vaclav.truhlarik@cvut.cz), Tomáš Pivoňka (tomas.pivonka@cvut.cz), Libor Přeučil (libor.preucil@cvut.cz)
The research primarily focuses on vision-based teach-and-repeat navigation for long-term autonomous systems, including many topics for student projects. These topics relate to computer vision, sensor fusion, robust control, robotics, and software development. Most of the tasks also provide an opportunity to work with real mobile robots or small aerial vehicles.
The sample topics are:
- Development of a new teach-and-repeat system
- Visual place recognition using deep-learning techniques
- Robot localization and control based on local visual features
- Adaptive teaching of the trajectory and environment model
- Systems for obstacle detection and collision avoidance
- Neural circuit policies for teach-and-repeat navigation
F1Tenth Autonomous Racing Competition
Contact person: David Zahrádka (david.zahradka@cvut.cz)
Develop a 1/10 scale autonomous racing car for the international F1Tenth racing competition. At IMR, you can take part in a prestigious racing competition where CTU has repeatedly secured top placements, including victories. You can develop LiDAR and camera-based localization, mapping, and control algorithms. The goal is to drive through the racing track safely and, at the same time, as fast as possible!
Aerial Vehicles Navigation
Contact person: Jan Chudoba (jan.chudoba@cvut.cz)
Flying robots (drones) have broad potential applications, particularly in inspection, surveillance, and monitoring tasks within various industrial environments. We focus on the control and development of navigation algorithms for drones, especially in settings where no external navigation system (e.g., satellite navigation) is available. To support the development and testing of these systems, we use a laboratory equipped with a Vicon motion capture system. Achieving fully autonomous navigation without reliance on external infrastructure also requires addressing challenges in low- and mid-level drone flight control.
INDUSTRIAL AND RESEARCH PROJECTS
Our laboratory collaborates with numerous industrial partners whom we help with challenging real-world robotic applications. Thus, students not interested in purely research topics can gain practical experience by working on these.
Generalized Pick and Place Solutions
Contact person: Viktor Kozák (viktor.kozak@cvut.cz)
Our laboratory collaborates with numerous industrial partners whom we help with challenging real-world robotic applications. Thus, students not interested in purely research topics can gain practical experience by working on these. This project aims to design and implement a generalized framework for solving pick-and-place tasks using a robotic arm, with an emphasis on adaptability to various object types and target placements. The student will investigate state-of-the-art methods for object localization, grasp planning, and accurate placement, using sensory input from an RGB-D camera. The work will involve exploring relevant software tools and frameworks, including ROS, MoveIt, and other perception libraries. The solution will be implemented and validated on a robotic arm in laboratory settings. The final outcome should be a modular and reusable system.
Intralogistics with Autonomous Mobile Robots
Contact person: Karel Košnar (karel.kosnar@cvut.cz)
This project involves the use of self-driving technology within industrial environments to optimize the transportation of goods. Autonomous vehicles navigate through the facility, fulfilling tasks such as picking up, transporting, delivering items, loading, and unloading goods onto shelves while reacting dynamically to the environment. Research involves advanced sensor processing, mapping, localization, collision avoidance, path-planning, and coordination with other machines or systems in real-time.
Mobile Robotic Platforms
Contact person: Jan Chudoba (jan.chudoba@cvut.cz)
We focus on controlling various mobile robots with special kinematics, which enable better maneuverability or passage in confined spaces or are structurally simpler. This is advantageous in numerous industrial and research applications. Different chassis designs with omnidirectional movement capabilities require the development of new motion control methods that effectively utilize expanded kinematic possibilities.
EXAMPLES OF FINISHED STUDENT PROJECTS
The Intelligent and Mobile Robotics Group has a history of more than 25 years conducting both basic and applied cutting-edge research, and numerous bachelor and master thesis projects.
Navigation for a Two-Handed Mobile Manipulator TIAGo++
Navigation of the TIAGo++ robot in multi-floor environments, including elevator usage for inter-floor movement. Implementation of existing methods for singlefloor navigation and development of elevator usage strategy based on object detection and visual navigation. Successful achievement of floor-to-floor movement and dynamic map adaptation.
TIAGo++: A Robotic Archer
Development of a robotic archer using TIAGo++ robot equipped with a camera for target detection. Mathematical model simulation, trajectory design for low motor torque, target detection, aiming, and shooting processes implemented. Experimental evaluation of trajectory, target detection, and shooting components.
Local Outdoor Navigation of a Mobile Robot by Terrain Traversability Mapping
Autonomous navigation of mobile robots in unknown environments, including simultaneous localization and mapping (SLAM) techniques such as the iterative closest point (ICP) algorithm. This project addressed traversability evaluation based on robot properties and geometric features. Implementation of ICP SLAM for navigation, including traversability analysis, pathfinding, and path following. Testing was performed in Gazebo simulator and real city environment.
Search for a Static Object in a Known Environment
Locating a static object in a known 2D polygonal environment using a mobile robot with a 360° sensor. Design and implementation of an efficient search strategy using innovative discretization and metaheuristic techniques. It has been confirmed that the proposed method surpasses existing approaches in terms of solution quality. Additionally, it remains competitive in terms of computational time.
Close Enough Travelling Salesman Problem in a Polygonal Domain
Solving the Close-Enough Traveling Salesman Problem (CETSP) in circular-shaped areas. Introduction of a novel General Large Neighborhood Search (GLNS-CETSP) method combining shortest path and point-circle-point algorithms with GLNS and Touring Circle Problem algorithms. Comparative experiments demonstrated improved results in shorter computation time.
Trajectory Planning for a Heterogeneous Team in an Automated Warehouse
Trajectory planning for a cooperative team of robots in an automated warehouse environment. Adaptation of planning algorithms to specific warehouse requirements, including modifications of existing plans. Development of a fleet management system with human-worker integration. Implementation and comparative evaluation.