This is an exciting and developing field.
Technology doesn’t wait for permission to transform.

Title	Author, year	Description (The thesis aims to)	Programming languages or data formats uesed	Hardware used	Prototype testing videos
System design for palletizing bags (or sacks) onto a pallet	Vladislav Magilnitski, BSc 2024	The thesis aims to develop a system for Palletizing sacks onto a EPAL3 pallets, to develop a preliminary palletizing algorithm and analyze the system’s technical and economic aspects such as return on investment (ROI), risk assessment and performance evaluation.	ABB RAPID, ABB Robotstudio add-on tool Palletizing PowerPac generated visualization in GLB format	ABB IRB 460 with ABB ClawGripper, EN13698-2 pallets (simulation with in ABB Robotstudio Palletizing PowerPac), some robots from Kuka, Yaskawa and Fanuc in comparison
Machine-Vision-Based Manipulator Positioning System for Sorting	Mulat Tigabu Girmay, MSc 2022	The thesis aims to develop a machine vision-guided robotic system for automated sorting tasks. The system integrates a vision sensor such as Cognex camera with a robotic manipulator to detect, locate, and sort objects based on visual features. Cognex camera is applied for object detection and pattern recognition. Extracts object coordinates and orientation from captured images.	ABB RAPID	An ABB YuMi dual-arm robot is used for object handling. it receives position data from the vision system to perform pick-and-place operations.
Reconstruction and conversion of a manual milling machine to computer numerical control	Hardi Tambets, BSc 2021	The thesis aims to convert a manual milling machine into a CNC (Computer Numerical Control) machine, while retaining the option for manual operation. This topic was motivated from the need for a affordable, flexible machining solutions in student-led projects at Tallinn University of Technology’s Robotics Club. Popular CNC controllers (e.g., Centroid Acorn, Mach3) were compared to select the most suitable one. Custom mechanical parts were designed and manufactured for the conversion. Detailed schematics were created to integrate the selected controller with the mechanical and electrical systems of the milling machine.	G-code, CAD/CAM formats	The system was demonstrated using a CNC machine (Optimum BF20L) in TalTech Robotics club.
Integration of Industrial Manipulators and Mobile Robots with Open-Source Software Platform ROS	Andre Talvoja, BSc 2021	The thesis aims to create a unified control system that enables robotic tasks in industrial automation independent of manufacturer specific languages. ROS is used to manage communication between different robotic components. Middleware handles sensor data, actuator commands, and coordination logic.	Python	The system was demonstrated using a Mitsubishi MELFA industrial robot and a mobile robot platform available in TalTech.
Electromagnetic Near-field Scanner Based on Industrial Robot	Vladimir Šulžik, BSc 2020	The thesis aims to design and implement a system that uses an industrial robot to perform electromagnetic near-field (EMNF) scanning. The goal is to automate the scanning process for analyzing electromagnetic emissions from electronic components or systems. The spatial coordinates from the robot were mapped to the measurement data to generate near-field emission maps.	MATLAB was used to process and visualize the collected electromagnetic data, C (Arduino IDE) was used for magnetic sensor interfacing, CSV (comma-separated values) was used to store data	Honeywell magnetic sensor SS496A1, Mitsubishi RV-1A robot with FESTO DGE linear axis
Guides for Working with Industrial Automation 3D Modelling Package	Siim Kallisaar, BSc 2020	The thesis aims to create step-by-step beginner guides for using a 3D modeling package (Ciros Studio) in the context of industrial automation. It focuses on helping users model and simulate automation systems that include hardware components and programmable logic controllers (PLCs).	FBD (Function Block Diagram, CoDeSys), SFC (Sequential Function Chart, CoDeSys), MELFA Basic, KRL (KUKA Robot Language), IRL (DIN 66312)	Mitsubishi RV-1A robot with FESTO DGE linear axis, Festo MPS Distributing station

If you plan to develop a real physical prototype, you are welcome to my guidance (and supervision).
Don't forget that you do everything for yourself, not for me.

Projects related to industrial robots (in project subject MHK0035 spring 2019)

Title	Author, year	Description	Programming languages or data formats uesed	Hardware used	Prototype testing videos
Alternative tools for industrial robot IRB1600: Minifrees (rotary tool) Dremel	2019		ABB RAPID (Robotstudio Machining PowerPac, etc.)	IRB1600-1.45 with IRBP positioner, 3D printed tool holder, Dremel rotary tool
Alternative tools for industrial robot IRB1600: UFM500	2019		ABB RAPID (Robotstudio Machining PowerPac, etc.)	IRB1600-1.45 with IRBP positioner, tool holder, UFM500

Scientific publications related to industrial robots

Title	Author, year	Description	Programming languages or data formats uesed	Hardware used	Prototype testing videos
A Review of Programming and Human–Robot Interaction Methods in Industrial Robotics	Madis Lehtla, 2025	Industrial robot programming is rapidly evolving as classical methods are increasingly augmented by AI‑driven tools. Teach pendants, offline programming, and manual coding remain essential, but they are now complemented by generative code assistants, natural‑language interfaces, and automated visual programming systems that lower the skill barrier. These advances also introduce new HRI challenges, where flexibility, clarity, and usability are critical in dynamic, collaborative industrial settings. Within this shift, ROS serves as a unifying framework capable of integrating both traditional techniques and emerging AI‑based tools. The future of industrial robot programming is therefore hybrid—combining established practices with intelligent, adaptive systems to create better robotic workflows.	KRL, RAPID, MELFA Basic, UR Script, Python, C++, Rust, IEC 61131-3, Node-RED, Scratch, Blockly, mBlock, Stateflow, SCXML, DiNeROS, FLC-ROS, SysML, UML, URDF, MATLAB, GeafCet, Petri Net Plans for ROS	Mitsubishi MELFA RV-2AJ industrial manipulator
Pick-and-Place Operation of Linear Delta Robots with Low Energy Consumption	Valery Vodovozov; Madis Lehtla; Zoja Raud; Eduard Petlenkov, 2024	The paper shows that smooth, jerk‑limited trajectories and adaptive control strategies can significantly cut the energy consumption of linear delta robots during pick‑and‑place tasks. By optimizing motion planning and reducing unnecessary actuator effort, the approach maintains high speed and precision while improving efficiency. The work provides a useful framework for energy‑aware robot design and supports further research into lightweight, sustainable automation systems.	MATLAB, Festo Teach Language (aka KeMotion KAIRO)	FESTO EXPT-45-E1
Managing Energy Consumption of Linear Delta Robots Using Neural Network Models	Valery Vodovozov; Madis Lehtla; Zoja Raud; Natalia Semjonova; Eduard Petlenkov, 2024	This study presents a neural‑network‑based framework for reducing the energy consumption of linear Delta robots operating in high‑speed pick‑and‑place environments. Two optimization objectives are addressed: minimizing power at a single tool position and minimizing total energy across complete pick‑and‑place cycles by selecting energy‑optimal joint‑space trajectories. The approach employs two multilayer feedforward neural networks—a forward‑kinematics‑plus‑power (FKP) model and an inverse‑kinematics‑plus‑power (IKP) model—both trained using Bayesian Regularization and achieving high predictive accuracy (R ≈ 0.998). These models enable rapid estimation of robot kinematics and power demand without requiring detailed physical modeling. Results show that multiple joint configurations can reach the same Cartesian position, and selecting the configuration predicted to require the least power yields static energy savings of approximately 1–3%. For dynamic multi‑target tasks, integrating neural‑network power predictions with a graph‑based shortest‑path algorithm reduces total energy consumption by up to 20%. Experimental validation on a physical robot confirms these findings, demonstrating ~2.5% variation in static power across configurations and ~9.5% energy savings from optimized motion paths, with average operating power around 250 W. Overall, the method demonstrates that neural networks can effectively model and optimize the energy behavior of linear Delta robots, enabling 1–20% energy reduction without hardware modifications.	MATLAB, Festo Teach Language (aka KeMotion KAIRO)	FESTO EXPT-45-E1