Abstract of project:
It may be observed that the current architectures and programming models do not adequately address the many fundamental and crucial challenges posed by these networked robots.
Some of the main challenges in networked robots include the need for the following -
(1) A generic and scalable framework for facilitating rapid deployment of robots and associated devices,
(2) Efficient mechanisms for servicing robots,
(3) Management of network resources and
(4) Realizing an application development environment to aid in writing and deploying adaptive programs.
It may also be observed that though mobile agents have been used in conjunction with robots no attempts at fine-tuning their mobility and population to improve performance have been made. The work reported herein is an attempt to formalize the use of mobile agents in the domain of networked robotic systems.
The main objectives of my research work were to address the challenges listed earlier and to try and evolve solutions for the same. Though it is difficult to address all of them directly, attempts were made to tackle the more pertinent ones. First we developed an architecture that uses mobile agents and their features to cater to service-requests by the robots. The use of this architecture will definitely encourage novices to come forward and join mainstream networked robotics. This aspect also stresses on the quick deployment and use of robots on the network thus addressing the first challenge cited earlier. The second challenge is addressed by proposing a bio-inspired migration mechanism for mobile agents and its performance is evaluated and compared with other migration strategies. The third challenge of effective utilization of resources has also been looked into by a proposal for a mechanism for mobile agent cloning using a stigmergy based cloning controller. The last challenge on the use of mobile agents that have an ability to adapt and/or modify the code they carry as payload is left for future research.