UAV 4D Trajectory Planning for Dense Urban Air Traffic
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional air traffic control systems are inadequate for managing dense drone traffic due to the need for rapid decision-making and lack of human expertise, especially in urban environments where drones operate at low altitudes and near obstacles, requiring automated solutions for optimizing trajectories and avoiding conflicts.
Innovation Solution
A computer-implemented method using deep reinforcement learning and autoencoders to compute optimized 4D trajectories for drones, considering manoeuvrability, obstacles, and safety corridors, with a flight simulator to ensure flyability, and a centralized system for managing multiple drones.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional air traffic control systems are used to manage drone traffic, then human controllers can supervise and interact with aircraft, but the system cannot process the high density of drones fast enough and lacks automated optimization capability
Solution Approach 1:
The patent replaces the mechanical human controller system with an automated computer-based system that uses deep reinforcement learning algorithms to compute optimized trajectories. The system substitutes human cognitive processing with AI models that can rapidly analyze airspace conditions and generate safe trajectories for multiple drones simultaneously, addressing the speed bottleneck while managing complexity through algorithmic automation.
Solution Approach 2:
The system enables drones to autonomously request and receive optimized trajectories through automated computation. The deep reinforcement learning model independently analyzes the airspace state, constraints, and safety requirements to generate trajectories without continuous human intervention, allowing the system to serve itself in processing high-density drone traffic at rapid speeds.
2Reliability
If human air traffic controllers manually analyze and propose trajectories for each drone, then safety can be maintained through human judgment, but the decision-making process becomes too slow for high-density drone traffic
Solution Approach 1:
The deep reinforcement learning system incorporates feedback mechanisms where the model continuously monitors the generated trajectory against safety constraints, obstacle locations, and drone maneuverability parameters. The algorithm iteratively adjusts trajectory parameters based on feedback from the environment state, ensuring safety while reducing decision time through automated validation rather than manual review.
Solution Approach 2:
The system performs preliminary computation of multiple candidate trajectories and their associated risk assessments before final selection. By pre-computing and filtering trajectories based on safety constraints and maneuverability parameters, the system reduces real-time decision-making time while maintaining reliability through proactive safety verification rather than reactive human judgment.
3Productivity
If automated trajectory optimization is implemented using deep reinforcement learning and autoencoders, then trajectory computation speed increases, but the system complexity and computational requirements increase
Solution Approach 1:
The patent segments the complex trajectory optimization problem into multiple manageable components: a 3D airspace representation module, a recurrent autoencoder for spatial encoding, a deep reinforcement learning agent for trajectory generation, and a flight simulator for validation. This segmentation allows each component to handle specific aspects of the problem independently, reducing overall system complexity while maintaining high computational productivity.
Solution Approach 2:
The recurrent autoencoder acts as an intermediary between the raw 3D airspace representation and the trajectory generation algorithm. It transforms complex spatial data into a compressed latent space representation that the reinforcement learning model can process more efficiently, reducing computational complexity while preserving essential spatial information for accurate trajectory optimization.
4Reliability
If a flight simulator is used to validate trajectories, then trajectory flyability can be ensured, but additional computational time is required for simulation
Solution Approach 1:
The flight simulator performs preliminary validation of trajectory feasibility before the trajectory is finalized and sent to the drone. By pre-simulating the trajectory in a virtual environment that replicates actual flight conditions, the system identifies and corrects flyability issues beforehand, reducing the need for iterative re-computation and actually reducing total processing time while ensuring reliability.
Solution Approach 2:
The system creates a virtual copy of the physical airspace and flight dynamics model in the flight simulator. This digital twin allows trajectory validation without requiring physical test flights or complex real-time hardware-in-the-loop simulations, reducing computational time while maintaining accurate assessment of trajectory flyability through replicated environmental conditions.
Data Source
AI summary
A computer-implemented method includes receiving a trajectory request from an unmanned aerial vehicle, the request comprising: an initial point; a final point; at least one manoeuvrability parameter of the unmanned aerial vehicle; computing a plurality of optimized 4D trajectories between the initial point and the final point, complying with the at least one manoeuvrability parameter, and avoiding obstacles in an airspace, each 4D trajectory being associated with a performance score; a flight simulator simulating the plurality of 4D trajectories in order of decreasing performance score, until a 4D trajectory is considered to be flyable by the flight simulator; sending the trajectory considered to be flyable by the flight simulator to the unmanned aerial vehicle.