GRL (Generic Robot Language) is a simple language that extends traditional functional programming techniques to behavior-based systems. It is an architecture neutral language embedded within Scheme. GRL provides a wide range of constructs for defining data flow within communicating parallel control loops and manipulating them at compile time using functional programming techniques. Most of the characteristics of popular behavior-based robot architectures can be concisely written as reusable software abstractions in GRL. This makes it easier to write clear, modular code, to “mix-and-match” arbitration mechanisms, and to experiment with variations on existing mechanisms.

Functional Programming of Behavior-Based Systems, Ian Horswill, Autonomous Robots 9, 2000

Project members : Ian Horswill

The Cerebus Project is an attempt to extend parallel-reactive architectures to higher level cognitive tasks. Cerebus is intended to be a “self-demoing” robot in the sense that it is able to give a short talk about itself and it is able to field (slightly) more detailed questions (asked by typing English on a keyboard) about itself, in addition to demonstrating specific capabilities on demand.

Cerebus received the Nils Nilsson Award for Integrating AI Technologies at the 2001 AAAI Conference.

The Cerebus Project, Ian Horswill et al., 2000, forthcoming 

Project members : Robert Zubek, Ian Horswill

Project No-Excuses is an ongoing project to enable our robots to function indefinitely in a human environment. The robots should be able to operate for long periods with little or no human assistance. Furthermore, they should be actually performing some useful function, e.g. fetching coffee, make announcements, etc, while interacting with humans as needed. That is, they should not appear to be simply wandering aimlessly in the hallways.

Project members : All AMRG undergrads, Ian Horswill

The RObot Self-Explains whY (ROSEY) project aims to enable a behavior-based robot to generate explanations of its own changing behaviors and state in response to questions such as "Why are you turning now?", "Why are you stopped?". ROSEY can be treated as an extension of Cerebus; it borrows Cerebus's simple natural language interface while extending what Cerebus can explain. 

Project members : Dac Le, Ian Horswill

SEEKers is an attempt to build a search-and-capture system using groups of robots. The robots cooperatively search a network of corridors for a human intruder. When the intruder is located, the robots will try to cover chokepoints in an attempt to prevent the human from escaping. Then, the remaining area is narrowed bit by bit until the human is completely surround by robots and is no longer able to maneuver. 

Project members : Aaron Khoo, Ian Horswill

The FlexBot project aims to create a framework for developing smart bots for first-person shooters. We're using the Half-Life engine as our foundation to build on and are providing a clean API and other tools for rapid, simple bot development. See the FlexBot page for more details. 

Project members : Greg Dunham, Nick Trienens, Sanjay Sood, Aaron Khoo, Ian Horswill