Todor Stoyanov

I am a post doctoral researcher at the MR&O Lab in AASS since June 2012. In April 2012 I completed my PhD thesis in Computer Science, also at the MR&O Lab at AASS. Before coming to Örebro University, I completed a Bachelors degree in Electrical Engineering and Computer Science and a Masters degree in Smart Systems, both at Jacobs University Bremen.

Research Interests

My main research interests are in the field of probabilistic robotics and intelligent systems. During my PhD studies, I worked primarily on 3D perception, with a focus on registration and mapping using 3D range sensors. I have previously engaged in work on Simultaneous Localization and Mapping (SLAM), extraction of semantic information, and design of systems for safety security and rescue robotics (SSRR). My current work focuses on applications of 3D mapping and perception techniques for mobile manipulation in the context of industrial automation.

Current Projects

My current work focuses on mobile manipulation for intra-logistics applications. In April, AASS participated (in cooperation with University of Pisa) in the KUKA innovation award. We mounted the new KUKA iiwa robot on one of our autonomous forklift trucks. Coupled with the Velvet Fingers II gripper, the system was able to perform a simplified commissioning operation at the Hannover fair, running reliably over five days.
The next video shows a brief system description we submitted for the preliminary evaluation. Based on this video, we were selected to participate in the final round at Hannover fair.

The next video shows some more details about the grasp acquisition process we used for the Velvet Fingers gripper.

Software and Data Sets

Several of my recent publications and projects have also produced open source software, bundled together in a stack of ROS packages. Most of the software is available for download through our lab software webpage or directly from the oru-ros-pkg project webpage.

Supervision

I have been involved in several courses and Masters thesis supervision in the course of our International Masters program in Robotics and Intelligent Systems. I am also teaching the Sensors and Sensing Masters level course. For Lecture notes and further information, check the course web page

International Master's Thesis: 3D Scan-based Navigation using Multi-Level Surface Maps, Andreas Persson (2009).

International Master's Thesis: Evaluation of a Bisensor System for 3D Modeling of Indoor Environments, Athanasia Louloudi (2010).

International Master's Thesis: Scene Representation, Registration and Object Detection in a Truncated Signed Distance Function Representation of 3D Space, Daniel Ricão Canelhas (2012).

If you are interested in a Master's Thesis with me, please check the project ideas that we have posted, or alternatively contact me with your suggestions.

Previous Projects

Some of my previous research projects are shortly described in this section. For further details, please check the project webpages and respective publications.

Cognitive Robot for Automation of Logistic Processes (RobLog)

RobLog was a European Union FP 7 Project, dealing with the automatic unloading of standard shipping containers. At the moment, containers are unloaded manually since they are nearly always packed chaotically, the variety of transported goods is high, and time requirements are strict. Unloading of containers is a strenuous task as goods have a weight of up to 70 kg, thus motivating an automated unloading system. A robot for unloading generic, chaotically stacked goods presents interesting challenges in several research areas and in particular in 3D perception.

RobLog final demonstrations
This video shows the RobLog advanced demonstrator - a robot for unloading of heterogenouse loosely packed goods from shipping containers.

This video shows the final demonstration of the RobLog industrial demonstrator - a robot for unloading of heavy sacks of coffee from shipping containers.

RobLog year 2 demonstrations
In this video we show a trial run of the RobLog system, when used in the industrial use case scenario for unloading of coffee sacks.

This video demonstrates the RobLog project's scientific scenario platform. Goods are identified automatically and offloaded with a simple suction cup gripper.

Safe Autonomous Navigation (SAUNA)

SAUNA is a Swedish KKS project aimed at building and showcasing an autonomous navigation system in a large scale industrial scenario. The SAUNA project is one of the major unifying projects at AASS and advances several research fronts. One of the challenges for the SAUNA system is 3D perception in a dynamic environment. Building and maintaining consistent 3D world models over a life-long longterm operation is thus one of the main challenges in this project. In addition, the SAUNA approach also makes use of advanced traversability analysis of the constructed 3D models, as well as several layers of control, path and task planning, collision avoidance and coordination.

Demonstration of components
Several of the compnents that were developped for SAUNA have already been published in articles and conference proceedings. Some short videos outlining some of the developed functionalitites are also available. For more details, check the associated publications.
This video accompanies the article Normal Distribution Transform Occupancy Maps: Application to Large-Scale Online 3D Mapping

This video accompanies the article Fast 3D Mapping in Highly Dynamic Environments using Normal Distributions Transform Occupancy Maps

This video accompanies the article Normal Distributions Transform Monte-Carlo Localization (NDT-MCL)

This video accompanies the article Normal Distributions Transform Occupancy Map Fusion: Simultaneous Mapping and Tracking in Large Scale Dynamic Environments

DustBot European Project (FP6)

The Networked and Cooperating Robots for the Urban Hygiene project (DustBot) are aimed at developing a robotic system for urban hygiene. The project is now at a final phase, with three robots already developed and demonstrated at sites across Europe. The official website of the project provides more detailed information on the goals and achievements. Two types of robots were built for this project - DustClean and DustCart. DustClean is a four-wheeled platform equipped with brushes and is primarily designed for cleaning of pedestrian areas. DustCart is a two-wheeled robot that provides door-to-door garbage collection on demand. Both robots are also equipped with environmental monitoring modules.

Involvement in the project: safe robot navigation.

Note: The DustBot project was completed in January 2010.

Finally, here is a video sumarizing the DustBot project and showing some of the robot demonstrations:

SOHOT Project

The SOHOT Project aims at developing an interactive tour guide robot for combined indoor and outdoor environments.

Involvement in the project: localization and navigation in 3-dimensional environments.

ALL-4-eHAM Project

The ALL-4-eHAM Project aims at developing fully autonomous wheel loader machines for operation in asphalt production sites.

Involvement in the project: scanning while moving, navigation in 3-dimensional environments.

Publications

Link to my Google Scholar Profile.

Journal Articles

Refereed Conference and Workshop Papers

Theses

Contact

Todor Stoyanov
AASS Research Center, Dept. of Computer Science, Írebro University,
SE - 70182 Írebro, Sweden
Room: T1225,
Phone: (+46/0) 19 - 30 3358
Fax: (+46/0) 19 - 30 3463
Email (masked to avoid spamming): todor {dot} stoyanov {at} oru {dot} se