We consider the Earth–Venus mass-optimal interplanetary transfer of a low-thrust spacecraft and show that the optimal guidance can be represented by deep networks in a large portion of the state space. This opens up the possibility to compute the guidance profile continuously on board the spacecraft, circumventing all convergence issues associated to optimal solvers, and thus constituting a valid alternative to solutions based on the convexification of the underlying optimal control problem. A new general methodology called “backward generation of optimal examples” is proposed to create the data necessary to train the artificial neural networks. With respect to previous works, our databases contain orders of magnitude more optimal trajectories. Several schemes to train representations of either the optimal policy (thrust profile) or the value function (optimal mass) are proposed and tested. We find that they accurately approximate the optimal thrust and that a spacecraft employing these networks would be able to reach the target conditions using only 2% more propellant than in the corresponding mathematically optimal transfer. All networks trained are tested during simulations of interplanetary transfers with respect to their ability to reach optimally the target conditions starting from nominal and off-nominal conditions.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Real-Time Guidance for Low-Thrust Transfers Using Deep Neural Networks


    Contributors:

    Published in:

    Publication date :

    2021-01-21


    Size :

    13 pages




    Type of media :

    Article (Journal)


    Type of material :

    Electronic Resource


    Language :

    English




    Multiconstrained Real-Time Entry Guidance Using Deep Neural Networks

    Cheng, Lin / Jiang, Fanghua / Wang, Zhenbo et al. | IEEE | 2021



    MULTI-CONSTRAINED REAL-TIME ENTRY GUIDANCE USING DEEP NEURAL NETWORKS

    Cheng, Lin / Wang, Zhenbo / Jiang, Fanghua et al. | TIBKAT | 2020



    Near-Optimal Feedback Guidance for Low-Thrust Earth Orbit Transfers

    Atmaca, D. / Pontani, M. | Springer Verlag | 2024

    Free access