Probabilistic Mission Evaluation for Transparent UAV Route Validation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current mission design systems for unmanned aerial vehicles and other mobile entities lack transparency, adaptability, and the ability to handle complex mission conditions, failing to provide intuitive and dynamic solutions that can be adjusted by non-experts to reflect changing rules and preferences.
Innovation Solution
A method and system that utilize probabilistic logic programming to transform information about planned trajectories, mobile entities, environments, and public body rules into a symbolic representation, enabling intuitive and adaptive mission evaluation through probabilistic logic queries, allowing for transparent decision-making and classification based on knowledge and rules.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If rule-based methods with pre-defined graphs and sub-routes are used, then mission paths can be divided into manageable sub-routes with sub-goals, but specific paths or missions cannot be validated and the system requires behavior trees to be defined beforehand
Solution Approach 1:
The system dynamically validates missions by checking against probabilistic logic programs that represent public body rules, rather than relying on pre-defined behavior trees. The validation process adapts to different mission scenarios by interpreting rules probabilistically, allowing the system to handle unforeseen situations and non-expert defined rules flexibly.
2Productivity
If constrained optimization problems are used to optimize missions, then mathematical constraints can represent rules and limitations, but the optimization problem becomes non-transparent to users and impossible to adjust for non-experts
Solution Approach 1:
The system introduces probabilistic logic programs as an intermediary layer between mathematical optimization and user interaction. These programs translate complex optimization constraints into interpretable logical rules that non-experts can understand and modify, while still enabling rigorous optimization through automated solvers.
Solution Approach 2:
The system changes the representation parameters from purely mathematical constraints to probabilistic logic programs, which maintain the rigor needed for optimization while improving interpretability. The probabilistic nature allows rules to be expressed in natural language terms that non-experts can comprehend and adjust.
3Reliability
If manual clearance requests are used for new routes, then operators can evaluate newly proposed routes based on available information, but manual feedback is required and quick adaptations to the decision-making process are not possible
Solution Approach 1:
The system implements automated feedback loops where probabilistic logic programs continuously evaluate mission proposals against public body rules. This automated feedback mechanism replaces manual operator evaluation for routine clearances, enabling quick adaptations while maintaining reliable rule-based assessment. Operators remain in the loop for complex or uncertain decisions.
4Extent of automation
If automated air traffic control systems are used, then autonomous planning can be performed with clearance granted or denied in a separate step, but the system is not transparent in decision making and not able to handle complex boundary conditions
Solution Approach 1:
The system uses probabilistic logic programs as an intermediary that bridges autonomous planning and transparent decision-making. The logic programs encode public body rules in an interpretable format, allowing automated clearance decisions to be made while maintaining transparency through logical rule evaluation that can be inspected and understood by operators.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present invention relates to a system for evaluating a mission of a mobile entity. The system comprises a means for obtaining information on a planned trajectory, the mobile entity, mission requirements, environment and public body rules, a means (3..5) for transforming the information into a symbolic, probabilistic representation to generate a probabilistic logic program, wherein facts are extracted from the information and systematically designated with names within the probabilistic logic program, and a means (6) for generating an answer to the a probabilistic logic query based on the probabilistic logic program.