[Show abstract][Hide abstract] ABSTRACT: As autonomous robots become more complex in their behavior, more sophisticated software architectures are required to support the ever more sophisticated robotics software. These software architectures must support complex behaviors involving adaptation and learning, implemented, in particular, by neural networks. We present in this paper a neural based schema  software architecture for the development and execution of autonomous robots in both simulated and real worlds. This architecture has been developed in the context of adaptive robotic agents, ecological robots , cooperating and competing with each other in adapting to their environment. The architecture is the result of integrating a number of development and execution systems: NSL, a neural simulation language; ASL, an abstract schema language; and MissionLab, a schema-based mission-oriented simulation and robot system. This work contributes to modeling in Brain Theory (BT) and Cognitive Psychology, with applications in Distributed Artificial Intelligence (DAI), Autonomous Agents and Robotics.
[Show abstract][Hide abstract] ABSTRACT: We review seven neurocomputational models covering key physiological, topographical, and information-processing features of striate cortex. Topics include simple and complex cells (Heeger, 1996; Dan, 2001), neurobiological learning rules (Jordan, 1991), ocu- lar dominance columns (Blasdel, 1995), binocular dispar- ity (Sejnowski, 1993), orientation columns (Shapely, 2000), and focal attention (Kohonen, 2002). A short over- view of the physiological and anatomical features of stri- ate cortex is provided, leading to a review of each neuro- computational model, its methods, and its main results. We finish with a discussion of how these models fit into the larger neurocomputational effort aimed at under- standing V1.
[Show abstract][Hide abstract] ABSTRACT: Humanoid and mobile robots are just a few exam-ples of intelligent machines that require sophisticated software to transform sensory information into pur-poseful actions. When writing this type of large-scale, complex software, developers benefit from domain-specific guidelines that promote code reuse and inte-gration. The Intelligent Machine Architecture (IMA) was designed to provide these guidelines and is cur-rently used to control the ISAC [Peters et al., 2000], Helpmate [Kawamura et al., 2000], and Scooter [Wilkes et al., 2002] robots at the Vanderbilt Univer-sity Intelligent Robotics Laboratory. IMA was also developed to address software integration and scal-ability issues in intelligent machines [Bagchi et al., 1992]. Software integration is the process of combin-ing software to extend the functionality of the system. Software scalability is a measure of how well a particular piece of software allows for future integration [Schach, 2002].
[Show abstract][Hide abstract] ABSTRACT: Thesis under the direction of Professor Don Mitchell Wilkes A revised version of the Intelligent Machine Architecture (IMA) software architecture and development environment has been implemented. The IMA 2.5 software architecture provides support for distributed concur-rent programming using integrative robotics pproaches. These approaches combine agent-based, behavior-based, and reactive algorithms to provide control for intelligent machines. In IMA 2.5, control algorithms are encapsulated within intelligent agents, collections of objects distributed across a network. Intelligent agents interact with each other according to guidelines specified in the IMA 2.5 programming paradigm. The software architecture and programming paradigm are accompanied by the IMA 2.5 development environment, a suite of tools for constructing, debugging, and managing intelligent agents.