Preliminary Project Ideas

1
Preliminary Project Ideas

• John Morgan
• Autonomous object interception in an adaptive
  learning strategy.
• Use of genetic algorithm to evolve and evaluate
  generations of neural nets to solve problem.
• Use of event-driven architecture for firing the
  nodes of the neural network.
• Dan Vevang
• Perform visual reasoning by training a system
  to recognize objects in digital images and to
  categorize them by feature and function.
• Make use of existing databases containing already
  classified scene-based objects and contexts for
  training the recognition system.
• Narrow project scope to 4 or 5 scene categories.
• Arindam Khaled
• Perform autonomous robot navigation by
  interpreting surrounding context and comparing it
  to pre-stored maps.
• Use an A or genetic algorithm for adaptive
  comparison to map context.
• Provide ability to reason to whereabouts and to
  avoid obstacles.