Metaplanning Framework to Evaluate and Select Decision Making Paradigms for UAV Missions

Project: Research project

Project Details

Description

Metaplanning Framework to Evaluate and Select Decision Making Paradigms for UAV Missions Metaplanning Framework to Evaluate and Select Decision Making Paradigms for UAV Missions In recent times, the advancements in the capabilities of Unmanned Air Vehicles (UAVs), along with the focus on mixed-criticality architecture for UAVs, have enabled them to be used in complex mission scenarios. Consequently, AI planning for the UAV missions is getting increasingly multi-dimensional in nature making different planners to be applicable for different mission types. Moreover, description of the UAV domain, required for the planners to interpret the problem characteristics, can get undesirably clumsy because of the multi-dimensional nature of the missions. The overall goal of this project is, firstly, to create a meta-planning framework, which i) classifies planning domains, i.e creates a taxonomy of different planning problems and associated planners based on the problems characteristics, ii) evaluates which domain and planner are best suited for spatio-temporal decision making for UAV missions, and iii) dynamically selects the appropriate planners to generate plans such that the mission objectives are optimized; and secondly, to develop a user-friendly description language for the UAV domain. The PI has performed a preliminary study on Q-value based evaluation of planned actions for spatio-temporal decision making under critical situations; and plan to transcend beyond the action evaluation to perform planner and domain evaluation for the proposed meta-planning framework. If need be ASU has a technology plan for ITAR in place which have been used for other project.
StatusFinished
Effective start/end date2/8/1012/31/11

Funding

  • INDUSTRY: Domestic Company: $174,524.00

Fingerprint

Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.