Tristan Barnett has worked on several research projects in the field of artificial intelligence, particularly intelligent agents. Problems addressed in Tristan’s research include:

• Robotic learning
• Reinforcement learning in real-world applications
• Multi-agent learning and simulation
• Information security and information warfare.

CAEDA Cognitive Architecture CAEDA is a cognitive architecture developed by Tristan Barnett which allows agents to learn together more effectively in complex environments. CAEDA has been tested in a multi-robot scenario in which several robots must learn to cooperatively navigate through a collapsed building to find injured or lost persons.
HIVE Multi-agent Simulator HIVE is a set of simulation tools developed by Tristan Barnett for use in embodied multi-agent environments. It includes 2D and 3D graphical environments, an intelligent agent API and environmental tools such as weather simulation and random maze generation. It has been used to experimentally test the CAEDA cognitive architecture in tank-tracked robot search and rescue scenarios.
StegaSpider Surveillance System The StegaSpider Surveillance System is an agent-based system developed by Tristan Barnett which searches for covert communication channels on the Internet. StegaSpider uses a logic-based agent that searches for hidden messages embedded into documents on the Internet. A Graphical User Interface (GUI) allows the user to intuitively program and monitor the agent’s behaviour as well as view reports on documents found to contain hidden messages (steganography).