This paper is concerned with the problem of the two-dimensional optimal online trajectory generation for a flying robot for terrain following/terrain avoidance purposes using neural network. To this end, the terrain is modeled as a terrain following constraint. In the next step, the inverse dynamics method is utilized for offline trajectory generation. Since the main objective of this research is online trajectory generation, the neural network method is employed for this purpose. To train the neural network, various approaches such as steepest descent, conjugate gradient, resilient back propagation, and Levenberg–Marquardt are explored in detail, and finally, the Levenberg–Marquardt method is selected on account of some merits. The efficacy of the neural network method is demonstrated by extensive simulations, and in particular, it is verified that this method is able to produce a solution satisfying all hard constraints of the underlying problem.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Optimal online trajectory generation for a flying robot for terrain following purposes using neural network


    Contributors:


    Publication date :

    2015-05-01


    Size :

    18 pages




    Type of media :

    Article (Journal)


    Type of material :

    Electronic Resource


    Language :

    English







    Optimal Control Approach to Terrain Following Trajectory Generation

    Stephens, Shawn S. / Hess, Joshuah / Kunz, Donald L. | AIAA | 2020


    OPTIMAL CONTROL APPROACH TO TERRAIN FOLLOWING TRAJECTORY GENERATION

    Stephens, Shawn S. / Hess, Joshuah / Kunz, Donald L. | TIBKAT | 2020