Computer games have gained attention as a testbed for research in AI and related disciplines early already. A recent addition to the set of games that is being used in this regard is Angry Birds. It offers several interesting aspects in terms of research such as reasoning about physical effects and qualitative spatial reasoning, to name just two. The KBSG has recently engaged in this new testbed, participating as Team Akbaba in the Angry Birds AI Competition founded in 2012.
Projects conducted at KBSG
In a project funded by the ICT project house Foundations of a Digitized Industry, Economy, and Society of RWTH Aachen University we are investigating how to optimize the performance of robot fleets in production logistics scenarios.
For more information visit the POPL-SMT project page at our partners from the Theory of Hybrid Systems Group
Mobile robots that are deployed, for instance, for domestic tasks, need to fulfill them even if they do not have complete information about the domain beforehand. Consider, for example, the task of fetching an object from a specific room. Consider further that the robot has a map of the environment and that it can plan a path to the living room. However, only when reaching the door of the living room, it will know whether or not the door is open. Moreover, it might not be known where in the living room the object is located such that the robot needs to make a plan how to find it. Next, a plan for grabbing the object in a safe way is required. This includes taking camera shots from different angles and finding an appropriate location to grab the object. Basically, the action control of the robot continuously needs to combine active perception and high-level planning with action execution. The robot has to deal with qualitative action and world descriptions as well as with uncertainty and quantitative data from sensors and actuators. The objective of this project is to develop methods for solving tasks like the above in an intelligent way by combining perception, high-level planning and action execution. An evaluation of the developed methods will be conducted on existing mobile robotic platforms.
The action language GOLOG has been used, among other things, for the specification of the behaviour of mobile robots. Since the task of such autonomous systems is typically open-ended, their GOLOG programs are usually non-terminating. To ensure that the program will let the robot exhibit the intended behaviour, it is often desirable to be able to formally specify and then verify the desired properties, which are often of a temporal nature. This task has been studied within our preliminary work from two perspectives: On the one hand, the problem was tackled for very expressive specification and action program formalisms, but without the goal of achieving decidability, i.e. the developed verification methods were not guaranteed to terminate. On the other hand, the verification problem was studied for action formalisms based on decidable description logics and very limited means of specifying admissible sequences of actions, which allowed us to show decidability and complexity results for the verification problem. The purpose of this project is to combine the advantages of both approaches by, on one hand, developing verification methods for GOLOG programs that are effective and practically feasible and, on the other hand, going beyond the formalisms with very limited expressiveness to enhance their usefulness. Among other things, both qualitative and quantitative temporal program properties will be addressed.
Biological algorithms of sound localization may be useful in studying acoustic orientation of robots. One of us (Lakemeyer) collaborated in the software design of a mobile robot that can give guided tours through exhibitions, among other things. We now want to equip this robot with a sound localization and a speech recognition system. The Jeffress model of binaural interaction contains delay lines and coincidence detection. It is realized in the barn owl. We have implemented a Jeffress-like model on a computer and have tested it with noise and speech signals (Calmes et al., in preparation). This model works pretty well even in cluttered surroundings. Currently we are implementing the software on the robot.
Homepage: HeRBiE @ Bio II
Welcome to the Working Group Cognitive Robotics. We have regular meetings to discuss (current) topics in Cognitive Robotics. This includes but is not limited to reviewing and discussing recent papers, books, and events.
Target of this project is developing a deliberative component supporting precise coordinated acting of multirobotic systems under real-time constraints. The integration of a total system and validation of results for simulated as well for real robots take place within the framework of ROBOCUP. As the general basis for the deliberative component which we will develop, the logic based controller language GOLOG, developed at the University of Toronto, will be used. This language has been advanced within our workgroup over the recent years and has been utilized successfully for controlling mobile robots.
Strategic Networks (e.g. in enterprises or research-groups) are the most ambitious models for cooperative perception and problemsolving at the moment. With these one tries to combine the advantages of the flexibility of market mechansims with the stability of hierarchical organizations. The Aachen TROPOS-projects' aim is the development of an agent oriented environment for the lifecycle-accompaniing requirements engineering in strategic networks. This analyzing environment should step by step be enlarged to a simulation- and supporting-tool. Every step concerning development and implementation will be evaluated by empirical analysis.
The aim of this project is the tight integration of the latest developments in the areas of planning languages and action logics in order to arrive at systems that are both very epressive and efficient. For comparing action languages like Golog and Flux with planning formalisms like PDDL -- which is being viewed as the standard in the area of planning -- a common semantics based on the situation calculus shall be developed. The expressivity of the different approaches shall be assessed via compilation techniques. A special focus will be on extensions of Golog, such as temporal and concurrent Golog. The resulting planning problems shall be approached with heuristic search, as implemented in the FF system, and the aid of compilation techniques. In return, language constructs from the Golog-family shall be integrated into PDDL and existing planners. The developed methods shall be evaluated in the context of autonomous robots.
The Knowledge-Based Systems Group covers research in the areas of Artificial Intelligence (AI), Knowledge Representation and Reasoning (KR+R), and Cognitive Robotics.
An application domain in this regard is domestic service robotics. The robot platform that we use for research in this domain is Caesar.