No abstract

Time-aware agents are agents capable of reasoning about their tasks duration and deadlines, and, more generally, to manage the temporal aspects of the execution of their tasks. We first focus on the case of agents in charge of long duration computations, sustaining that it is not acceptable for an autonomous agent to remain unaware of its environment for too long. We then consider deadline meetings when several time-aware agents share the same CPU. To achieve these goals, we recognize the importance of the artifact concept [16]. We introduce computational artifacts for long duration tasks and a coordination artifact for managing the CPU agenda and acting as an intermediary when agents negotiate CPU power. Control of computational artifacts is done thanks to a set of operating instructions dynamically computed by the coordination artifact.

The problem of controlling an object according to a proportional navigation guidance scheme has been widely studied. One of the major results of these studies is the preponderance of pure proportional navigation guidance laws (PPN) over true proportional navigation (TPN) ones. Unfortunately PPN is harder to study because of the difficulty to find a simple closed form solution to the differential equations defining the trajectory. The authors propose here a new approach of PPN. Simulations have then been performed in order to validate the theoretical approach. An Extended " a n Filter (EKF) was implemented in order to estimate the kinematic parameters of the trajectory. A processing using probabilistic and fuzzy modelling was then used to estimate the destination point (DP) of the object.

No abstract

No abstract