If you see this message, JavaScript is not activated in your browser. However, you need to switch it on to see this page properly. Using JavaScript basically allows us to present a nicer Web page and to keep the maintainance effort low.
In case it is not possible to activate JavaScript, you can find an alternative (though less comprehensive) version of the AASS Learning Systems lab web pages as part of the Örebro University Web pages.
Welcome to the AASS Learning Systems Lab!
People
The Learning Systems Lab is one of three research groups within the Centre of Applied Sensor Systems (AASS). It currently comprises 20 highly motivated researchers from 10 different countries (8 PhD students, 6 postdocs, 3 professors and 3 associated researchers; further details can be found under the item "People" in the menu on the left side of this page).
The lab was founded in 1999 by Tom Duckett. Since February 2006 it is lead by Achim J. Lilienthal.
Research
The research focus of the AASS Learning Systems lab is on the development of algorithms and robotic systems for real-world tasks. In order to achieve a high level of autonomy under different and varying environmental conditions the approaches developed are characterized by learning and fusion of information from different sensor modalities. We consider "traditional" environments as well as environments populated with networked and distributed intelligent artifacts. The long-term aim is to better understand perceptual, biological and physical processes through the help of robots, using them indirectly as a model or directly as a tool for experimentation. Six major directions of our research can be identified:
- Mobile Robot Olfaction, concerning all aspects of airborne chemical sensing with mobile robots and stationary gas sensors in natural environments. This includes in particular gas source localization and Bayesian gas distribution modelling (addressed in the DHRS-CIM project); statistical gas distribution modelling of pollution levels in urban environments (DustBot project); and time-dependent gas distribution modelling and sensor planning for large-scale environmental monitoring (Diadem project).
- Dexterous Manipulation and Motion Learning, concerning learning of skills and tasks for robotic manipulators from demonstrations given by a human teacher; in particular learning and development of grasping and manipulation skills for dexterous robotic hands (HANDLE project) and Programming-by-Demonstration of robot manipulators.
- Safe Operation in Dynamic Environments, concerning navigation of mobile robots in dynamic, structured and semi-structured environments. This includes in particular safe operation in semi-structured ambient environments such as pedestrian areas (addressed in the DustBot project); and autonomous transportation applications (MALTA project and ALL-4-eHAM project).
- 3D Perception, concerning efficient representations of 3D data, 6DOF scan registration, scanning-while-moving, semantic mapping, and fusion of visual and range information in the context of different applications. Specific application domains considered are autonomous mining vehicles, forklift trucks (MALTA project) and wheel loaders (ALL-4-eHAM project).
- Robotic Map Learning and SLAM, concerning learning of spatial and appearance-based maps by mobile robots. A variety of aspects of robotic map learning are addressed: from indoor to outdoor, from static to dynamic environments, from 2D to 3D maps, from topological and geometric maps to hybrid and semantic maps.
- Robot Vision, concerning mapping of visual data (eventually fused with further sensor modalities) to consistent internal models for change detection, visual SLAM, appearance-based localization and mapping and people tracking.
Collaborative Projects
Currently we are involved in six collaborative projects (ordered by starting date):
- EU 6FP STREP Dustbot: Dec 1, 2006 – Jan 30, 2010; funding: 270'000€
- KKS project MALTA: Apr 1, 2008 – Mar 31, 2011; funding [KKS, Robotdalen, EU Structural Fund]: 5.275Mkr, approx. 527'500€
- EU 7FP STREP Diadem: Sep 1, 2008 – May 31, 2011; funding: 270'200€
- EU 7FP IAPP DHRS-CIM: Sep 1, 2008 – Aug 31, 2012; funding: 276'500€
- EU 7FP IP HANDLE: Feb 1, 2009 – Jan 31, 2013; funding: 602'000€
- KKS project ALL-4-eHAM: Apr 1, 2009 – Mar 31, 2012; funding [KKS, Vinnova, Robotdalen]: 5.7Mkr, approx. 570'000€
Good Use Declaration
Regarding the intended applications that we target, we feel indebted to the Uppsala Code of Ethics for Scientists. Our aim is ultimately to free humans from dull and dangerous tasks (as phrased by Norbert Wiener: "the human use of human beings") and to understand perceptual, biological and physical processes through the help of robots. We are aware that our results may have also other, less beneficial, applications and therefore declare that it is strictly prohibited to use or to develop, in a direct or indirect way, any of our scientific contributions by any army or armed group in the world, for military purposes and for any other use which is against human rights or the environment.