6 March 2024 in Chinese Journal of Aeronautics
13 December 2022 in Aerospace Science and Technology
13 August 2024 in IEEE Transactions on Aerospace and Electronic Systems
This article researches synergistic planning for multiple hypersonic vehicles. Taking significant differences in vehicle parameters and synergistic demands into account, it is significantly challenging for computing a series of trajectories simultaneously in an efficient way. This challenge will become severer when there exists a high degree of nonlinear coupling between dynamics, thermal and load factors, constraints, and objective functions. To tackle with this problem, our proposal begins with a profile, which features natural satisfaction of terminal states, nofly zones, and waypoints. Meanwhile, we research and conduct analytic dynamics using this profile for relieving the coupling. Then, we formulate a straightforward parameter-searching model to tackle the synergistic problem. In typical synergistic scenarios, this model can comprise more than thirty design variables, which further leads to challenges in initial guess, global searching, and convergence. Therefore, we propose a hybrid algorithm that combines particle swarm optimization (PSO) with reinforcement learning (RL). PSO is independent of initial guess and suitable for searching in high dimensional spaces, but is typically lack of efficiency. While a RL strategy is featured with immediate network computation and PSO global searching after training of PSO-based offline synergistic trajectories. In the end, we carry out simulation in different synergistic scenarios for validating the effectiveness and robustness of our algorithm.
20 July 2024 in Aerospace Science and Technology
3 October 2025 in IEEE Transactions on Intelligent Transportation Systems
Rapid trajectory optimization for reusable launch vehicle (RLV) has become a focus in aerospace research; however, its inherent nonlinearities and multiple constraints make it highly challenging. To address this issue, our proposal begins with a glide profile by considering strict terminal and steady-glide constraints, from which analytic expressions were derived to reduce nonlinear coupling and facilitate optimization. Nofly zones were modeled as polygons, and a cubic spline-based lateral avoidance strategy was implemented. Then, a hybrid optimization framework integrating analytic solutions, particle swarm optimization (PSO), and deep reinforcement learning (DRL) was developed. PSO’s exploration capability was retained, while DRL dynamically adjusted swarm hyperparameters based on swarm behavior to improve efficiency and solution quality by balancing exploration and exploitation. In the end, numerical simulations under multiple scenarios validated full constraint satisfaction, with a mean terminal state error below 0.92%. The DRL-PSO framework demonstrated adaptability and efficiency in rapid trajectory computation, achieving an average runtime of approximately 1.0 s. Furthermore, its effectiveness was validated through comparative analysis with alternative methods.