Marin Simina

Creative Design: Reasoning and Understanding

This paper investigates memory issues that influence longterm creative problem solving and design activity, taking a case-based reasoning perspective. Our exploration is based on a well-documented example: the invention of the telephone by Alexander Graham Bell. We abstract Bells reasoning and understanding mechanisms that appear time and again in long-term creative design. We identify that the understanding mechanism is responsible for analogical anticipation of design constraints and analogical evaluation, beside case-based design. But an already understood design can satisfy opportunistically suspended design problems, still active in background. The new mechanisms are integrated in a computational model, ALEC1, that accounts for some creative behavior in case-based design.

Kinodynamic motion planning

This paper addresses the problem of kinodynamic motion planning for a non-holonomic bicycle system moving on a 2D plane in minimum time. As a solution to this optimization problem, we propose an algorithm for determining feed-forward control values, which in the absence of disturbances and uncertainties, provide a plan to move a system from an initial state to a goal state, in minimum-time. This algorithm provides great benefits at the actuator level, also known as the Executor. At this level, the algorithm is a planner which searches the state-space for control values that deliver minimum-time state trajectories. By finding feed-forward control values at the actuators level and by combining them with the reactive part of the system present in the feed-back control component, the control system can achieve better performance and smoother response to disturbances.

Computational Lessons from a Cognitive Study of Invention

This paper investigates both the role of fine-grained historical cases in developing computational models of techno-scientific thinking and the impact of such models for supporting information search and further inventions and discoveries. In particular, we investigate Alexander Graham Bells invention of the telephone and we propose a computational model to explain its essential aspects. We further derive lessons about how such model can be used to build human-computer interaction systems that augment the intelligence of users involved in information search. We conclude that historical data can be used to advance cognitive and computational theories of techno-scientific thinking and to build better human-information systems.