07-10, 17:20–17:30 (Europe/Amsterdam), While Loop (4.2)
This presentation demonstrates the use of ModelingToolkit based AerialVehicles.jl model libraries, JuliaSimControl and InfiniteOpt for optimal rendezvous/docking of a quadrotor with multiple ground vehicles focusing on the efficient cooperation of ground vehicles with a landing quadrotor. The central theme is the use of a collocation-based approach to optimize control efforts, ensuring a seamless rendezvous between multiple ground vehicles and the quadrotor using a shared time-to-go estimate.
This talk presents an in-depth exploration of a novel solution for the cooperative, optimal multi-vehicle rendezvous problem, focusing on rendezvous of a quadrotor with multiple ground vehicles. The solution, pioneering in its approach, combines a gain-free, time-dependent optimal control problem with a consensus protocol based on shared time-to-go information. This dual-stage method not only streamlines vehicle convergence but also ensures adaptability and efficiency in trajectory refinement.
The presentation will be structured into several key sections, each delving into different aspects of the proposed solution:
Introduction to Multi-Vehicle Rendezvous Problems:
We will begin by providing an overview of the challenges and current state of multi-vehicle rendezvous solutions, particularly emphasizing the complexities of coordinating the rendezvous. This will set the stage for understanding the novelty and necessity of the proposed method.
Overview of the Proposed Methodology:
The first stage involves solving the optimal control problem to generate initial rendezvous trajectories. Here, the focus will be on the gain-free, time-dependent nature of the control problem and its implications for efficient vehicle convergence. The second stage of our approach refines these trajectories in real-time, employing a consensus-based adaptation protocol.
Consensus Protocol and Trajectory Refinement:
The consensus based adaption would be such that the time-to-go estimates of the ground vehicles will be set to the time-to-go of the approaching quadrotor and our finite-time control framework would make the approach optimal.
Simulation Demonstrations:
A significant portion of the presentation will be dedicated to showcasing numerical simulations. These simulations will demonstrate the method's effectiveness, its superiority in terms of convergence and adaptability, and the impact of communication topology on performance. The simulations will utilize JuliaSimControl for optimal control generation and the AerialVehicles model library to model the quadrotor and ground vehicles, providing a practical and visual representation of the method's capabilities.
Practical Implications and Applications:
The talk will explore the practical applications of this methodology, discussing its potential in various fields such as autonomous transport, logistics, emergency response, and military operations. We will delve into how this solution can help the efficiency and safety of autonomous vehicle operations in complex, dynamic environments.
This presentation is designed to provide an understanding of our innovative solution to the multi-vehicle rendezvous problem. Attendees will gain valuable insights into advancements in autonomous systems, optimal control, and consensus protocols, leaving with a deeper appreciation of the complexities and potentials in this rapidly evolving field.
