A Real-Time Trajectory Optimization Method for Hypersonic Vehicles Based on a Deep Neural Network
Considering the high-efficient trajectory planning requirements for hypersonic vehicles, this paper proposes a real-time trajectory optimization method based on a deep neural network. First, the trajectory optimization model of the hypersonic vehicle reentry phase is developed. The pseudo-spectral m...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-04-01
|
| Series: | Aerospace |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2226-4310/9/4/188 |
| _version_ | 1797437451622416384 |
|---|---|
| author | Jianying Wang Yuanpei Wu Ming Liu Ming Yang Haizhao Liang |
| author_facet | Jianying Wang Yuanpei Wu Ming Liu Ming Yang Haizhao Liang |
| author_sort | Jianying Wang |
| collection | DOAJ |
| description | Considering the high-efficient trajectory planning requirements for hypersonic vehicles, this paper proposes a real-time trajectory optimization method based on a deep neural network. First, the trajectory optimization model of the hypersonic vehicle reentry phase is developed. The pseudo-spectral method is used to perform the trajectory optimization offline, and multiple optimal trajectory data are obtained. In addition, based on the inherent relationship between the state and control variables of a trajectory, a neural network is established to predict the current control outputs. The sample library of optimal trajectory data is used to train the parameters of the deep neural network to obtain an optimal neural network model. Finally, the simulation verification of the hypersonic vehicle reentry phase is performed. The simulation results show that under the condition of the initial value deviation and environmental interference, the proposed deep learning-based method can achieve a fast generation of hypersonic vehicle optimal trajectories, while achieving the advantages of high computational efficiency and reliability. Compared to traditional trajectory optimization algorithms, the proposed method has the generalization capability that satisfies the accuracy requirements and meets the demands of online real-time trajectory optimization. |
| first_indexed | 2024-03-09T11:19:28Z |
| format | Article |
| id | doaj.art-b0efac11ad084b12bcb4651d51eb6942 |
| institution | Directory Open Access Journal |
| issn | 2226-4310 |
| language | English |
| last_indexed | 2024-03-09T11:19:28Z |
| publishDate | 2022-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Aerospace |
| spelling | doaj.art-b0efac11ad084b12bcb4651d51eb69422023-12-01T00:21:50ZengMDPI AGAerospace2226-43102022-04-019418810.3390/aerospace9040188A Real-Time Trajectory Optimization Method for Hypersonic Vehicles Based on a Deep Neural NetworkJianying Wang0Yuanpei Wu1Ming Liu2Ming Yang3Haizhao Liang4School of Aeronautics and Astronautics, Sun Yat-sen University, Shenzhen 510275, ChinaSchool of Aeronautics and Astronautics, Sun Yat-sen University, Shenzhen 510275, ChinaScience and Technology on Space Physics Laboratory, Beijing 100190, ChinaScience and Technology on Space Physics Laboratory, Beijing 100190, ChinaSchool of Aeronautics and Astronautics, Sun Yat-sen University, Shenzhen 510275, ChinaConsidering the high-efficient trajectory planning requirements for hypersonic vehicles, this paper proposes a real-time trajectory optimization method based on a deep neural network. First, the trajectory optimization model of the hypersonic vehicle reentry phase is developed. The pseudo-spectral method is used to perform the trajectory optimization offline, and multiple optimal trajectory data are obtained. In addition, based on the inherent relationship between the state and control variables of a trajectory, a neural network is established to predict the current control outputs. The sample library of optimal trajectory data is used to train the parameters of the deep neural network to obtain an optimal neural network model. Finally, the simulation verification of the hypersonic vehicle reentry phase is performed. The simulation results show that under the condition of the initial value deviation and environmental interference, the proposed deep learning-based method can achieve a fast generation of hypersonic vehicle optimal trajectories, while achieving the advantages of high computational efficiency and reliability. Compared to traditional trajectory optimization algorithms, the proposed method has the generalization capability that satisfies the accuracy requirements and meets the demands of online real-time trajectory optimization.https://www.mdpi.com/2226-4310/9/4/188hypersonic vehiclepseudo-spectral methodtrajectory optimizationdeep learningreentry phase |
| spellingShingle | Jianying Wang Yuanpei Wu Ming Liu Ming Yang Haizhao Liang A Real-Time Trajectory Optimization Method for Hypersonic Vehicles Based on a Deep Neural Network Aerospace hypersonic vehicle pseudo-spectral method trajectory optimization deep learning reentry phase |
| title | A Real-Time Trajectory Optimization Method for Hypersonic Vehicles Based on a Deep Neural Network |
| title_full | A Real-Time Trajectory Optimization Method for Hypersonic Vehicles Based on a Deep Neural Network |
| title_fullStr | A Real-Time Trajectory Optimization Method for Hypersonic Vehicles Based on a Deep Neural Network |
| title_full_unstemmed | A Real-Time Trajectory Optimization Method for Hypersonic Vehicles Based on a Deep Neural Network |
| title_short | A Real-Time Trajectory Optimization Method for Hypersonic Vehicles Based on a Deep Neural Network |
| title_sort | real time trajectory optimization method for hypersonic vehicles based on a deep neural network |
| topic | hypersonic vehicle pseudo-spectral method trajectory optimization deep learning reentry phase |
| url | https://www.mdpi.com/2226-4310/9/4/188 |
| work_keys_str_mv | AT jianyingwang arealtimetrajectoryoptimizationmethodforhypersonicvehiclesbasedonadeepneuralnetwork AT yuanpeiwu arealtimetrajectoryoptimizationmethodforhypersonicvehiclesbasedonadeepneuralnetwork AT mingliu arealtimetrajectoryoptimizationmethodforhypersonicvehiclesbasedonadeepneuralnetwork AT mingyang arealtimetrajectoryoptimizationmethodforhypersonicvehiclesbasedonadeepneuralnetwork AT haizhaoliang arealtimetrajectoryoptimizationmethodforhypersonicvehiclesbasedonadeepneuralnetwork AT jianyingwang realtimetrajectoryoptimizationmethodforhypersonicvehiclesbasedonadeepneuralnetwork AT yuanpeiwu realtimetrajectoryoptimizationmethodforhypersonicvehiclesbasedonadeepneuralnetwork AT mingliu realtimetrajectoryoptimizationmethodforhypersonicvehiclesbasedonadeepneuralnetwork AT mingyang realtimetrajectoryoptimizationmethodforhypersonicvehiclesbasedonadeepneuralnetwork AT haizhaoliang realtimetrajectoryoptimizationmethodforhypersonicvehiclesbasedonadeepneuralnetwork |