UAV Path Planning With Frontier Perception and Dynamic Landing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Unmanned aerial vehicles face challenges in unknown large scenarios with inaccurate environmental perception, inefficient path planning, and unsafe landing due to reliance on local path planning methods and traditional control strategies that require extensive expert experience and repetitive learning.
Innovation Solution
An autonomous environmental perception and path planning method for unmanned aerial vehicles that includes real-time three-dimensional reconstruction, dynamic path optimization using local and frontier-perceived path optimization algorithms, and deep reinforcement learning for accurate dynamic landing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If local path planning methods are used for real-time operation, then the responsiveness and computational efficiency are improved, but the path optimization efficiency deteriorates due to path redundancy and continuous meaningless replanning at the same space
Solution Approach 1:
The patent performs preliminary action by constructing a global path in advance using global environmental information before local execution. This global path serves as a pre-planned route that guides local path planning, reducing redundant replanning and improving overall path optimization efficiency while maintaining real-time responsiveness through local adjustments.
2Device complexity
If traditional control strategies are used for landing control, then the system simplicity is maintained, but the landing accuracy and adaptability deteriorate due to reliance on expert experience and repetitive learning requirements
Solution Approach 1:
The patent replaces traditional mechanical control strategies with a deep reinforcement learning-based neural network controller. This substitution enables the system to learn optimal landing policies through experience without requiring manual expert tuning, significantly improving landing accuracy and adaptability to different environments while maintaining reasonable system complexity through the use of standardized deep learning frameworks.
3Loss of time
If local path planning is used with limited perception space, then the real-time processing capability is improved, but the environmental perception accuracy deteriorates due to lack of memory and continuous meaningless replanning
Solution Approach 1:
The patent introduces an intermediary mechanism in the form of a global path planner that processes global environmental information and generates a reference path. This intermediary guides the local path planner, reducing redundant replanning operations and allowing the system to maintain accurate environmental perception over larger spaces without sacrificing real-time processing capability, as the global planner filters out meaningless replanning scenarios.
Data Source
AI summary
An autonomous environmental perception, path planning and dynamic landing method includes: obtaining three-dimensional environment information in real time; determining a global starting point and a global end point, and generating an initial path; optimizing the initial path based on a local path optimization algorithm to obtain a first optimized path; when a perception threshold of the current position of the unmanned aerial vehicle is greater than a preset threshold, optimizing the initial path based on a frontier-perceived path optimization method to obtain a second optimized path and a local end point; when the unmanned aerial vehicle advances to the local end point, switching to optimizing the initial path in real time based on the local path optimization algorithm; and when the unmanned aerial vehicle arrives at the global end point, carrying out dynamic landing based on a deep reinforcement learning algorithm.