Master thesis proposal
Infotiv ABNy2 dagar kvar
Beskrivning
Advancing Human-Robot Collaboration through Simulation
PURPOSE OF THE STUDY
Infotiv, as part of the research project ArtWork (The Smart and Connected Worker), is continuing its investigation into simulation and digital twins for human-robot collaborative manufacturing environments. Building directly on the foundational work carried out in RITA-SIM[1] and SIMLAN[2] simulation platforms, this thesis aims to advance the platform from a proof-of-concept demonstrator towards a robust human-robot collaboration system.
The current RITA-SIM implementation relies on ground-truth simulator data for collision avoidance and task execution, and its pick-and-place scenarios are limited in scope and resilience. This thesis addresses those gaps by introducing real perception, predictive human-awareness, richer humanoid simulation, and scenarios. The goal is to evaluate and improve the effectiveness of simulation for one or several of the items below:
Replacing ground-truth collision data with a camera-based perception pipeline for robust, sensor-realistic obstacle avoidance
Predicting human movement trajectories in the shared workspace so the robot can plan proactively rather than reactively.
Implementing end-to-end kitting task sequences (advance scenarios) that handle continuous operation, part variation, and real-world failure modes such as misaligned parts
Expanding humanoid simulation scenarios and investigating live interactive control of a virtual human agent inside Simulator
Improving multi-camera localisation and agent tracking accuracy, and evaluating machine learning models for human pose estimation and activity recognition in the factory floor simulation
The study will produce open-source contributions to the RITA-SIM and SIMLAN repository and validate findings both in simulation and, where feasible, on the physical UR10e platform available at Volvo Group.
POTENTIAL RESEARCH QUESTIONS
How accurately can a vision-based perception pipeline detect and localise obstacles compared to ground-truth data?
Is real-time interactive control of a humanoid avatar in simulation feasible, and how does it improve the realism and coverage of collaborative scenarios?
How well can ML-based human pose estimation and multi-camera tracking generalise from simulated to real factory environments when trained in the siomulator?
Please read more about these projects in the open-source repositories:
[1] GitHub - infotiv-research/RITA-SIM: RITA Simulation
[2] GitHub - infotiv-research/SIMLAN: SIMLAN, Simulation for Multi-Camera Robotics
WHO ARE WE LOOKING FOR?
We are looking for 2 master's students with a background in mechatronics, electrical engineering, computer science/engineering or equivalent program (e.g., MPSYS, MPCAS, MPDSC), who wish to conduct their thesis during the spring of 2027. Applicants should be interested in software development and machine learning and have experience using Python, ROS2, Git, Linux and Docker.
ABOUT US
TechDev is a department at Infotiv who focuses on SW & HW development and test solutions. We currently consist of 70 technical consultants with diverse backgrounds and experience from many technological fields. Our employees use their expertise to provide tailored solutions to all kinds of challenges, ranging from SW development, machine learning and simulations to project & test management and way of working. One of our key strengths is the friendly atmosphere in our technical community, which provides access to TechDev's collective knowledge through internal collaboration tools and competence leader programs, continuously providing updates in the latest tech.
HOW TO APPLY
Apply for this thesis no later than 2026-12-31. Assure to attach your resumé and a short summary of why you want to partake in this thesis.
For further information, contact: Maria Alemyr [maria.alemyr@infotiv.se] +46(0)-76 890 78 72
PURPOSE OF THE STUDY
Infotiv, as part of the research project ArtWork (The Smart and Connected Worker), is continuing its investigation into simulation and digital twins for human-robot collaborative manufacturing environments. Building directly on the foundational work carried out in RITA-SIM[1] and SIMLAN[2] simulation platforms, this thesis aims to advance the platform from a proof-of-concept demonstrator towards a robust human-robot collaboration system.
The current RITA-SIM implementation relies on ground-truth simulator data for collision avoidance and task execution, and its pick-and-place scenarios are limited in scope and resilience. This thesis addresses those gaps by introducing real perception, predictive human-awareness, richer humanoid simulation, and scenarios. The goal is to evaluate and improve the effectiveness of simulation for one or several of the items below:
Replacing ground-truth collision data with a camera-based perception pipeline for robust, sensor-realistic obstacle avoidance
Predicting human movement trajectories in the shared workspace so the robot can plan proactively rather than reactively.
Implementing end-to-end kitting task sequences (advance scenarios) that handle continuous operation, part variation, and real-world failure modes such as misaligned parts
Expanding humanoid simulation scenarios and investigating live interactive control of a virtual human agent inside Simulator
Improving multi-camera localisation and agent tracking accuracy, and evaluating machine learning models for human pose estimation and activity recognition in the factory floor simulation
The study will produce open-source contributions to the RITA-SIM and SIMLAN repository and validate findings both in simulation and, where feasible, on the physical UR10e platform available at Volvo Group.
POTENTIAL RESEARCH QUESTIONS
How accurately can a vision-based perception pipeline detect and localise obstacles compared to ground-truth data?
Is real-time interactive control of a humanoid avatar in simulation feasible, and how does it improve the realism and coverage of collaborative scenarios?
How well can ML-based human pose estimation and multi-camera tracking generalise from simulated to real factory environments when trained in the siomulator?
Please read more about these projects in the open-source repositories:
[1] GitHub - infotiv-research/RITA-SIM: RITA Simulation
[2] GitHub - infotiv-research/SIMLAN: SIMLAN, Simulation for Multi-Camera Robotics
WHO ARE WE LOOKING FOR?
We are looking for 2 master's students with a background in mechatronics, electrical engineering, computer science/engineering or equivalent program (e.g., MPSYS, MPCAS, MPDSC), who wish to conduct their thesis during the spring of 2027. Applicants should be interested in software development and machine learning and have experience using Python, ROS2, Git, Linux and Docker.
ABOUT US
TechDev is a department at Infotiv who focuses on SW & HW development and test solutions. We currently consist of 70 technical consultants with diverse backgrounds and experience from many technological fields. Our employees use their expertise to provide tailored solutions to all kinds of challenges, ranging from SW development, machine learning and simulations to project & test management and way of working. One of our key strengths is the friendly atmosphere in our technical community, which provides access to TechDev's collective knowledge through internal collaboration tools and competence leader programs, continuously providing updates in the latest tech.
HOW TO APPLY
Apply for this thesis no later than 2026-12-31. Assure to attach your resumé and a short summary of why you want to partake in this thesis.
For further information, contact: Maria Alemyr [maria.alemyr@infotiv.se] +46(0)-76 890 78 72
Ansök till tjänsten
Master thesis proposal
Ny2 dagar kvar
OM FÖRETAGET
Infotiv AB
