Traditional robotic control suits require profound task-specific knowledge for designing, building and testing control software. The rise of Deep Learning has enabled end-to-end solutions to be learned entirely from data, requiring minimal knowledge about the application area. We design a learning scheme to train end-to-end linear dynamical systems (LDS)s by gradient descent in imitation learning robotic domains. We introduce a new regularization loss component together with a learning algorithm that improves the stability of the learned autonomous system, by forcing the eigenvalues of the internal state updates of an LDS to be negative reals. We evaluate our approach on a series of real-life and simulated robotic experiments, in comparison to linear and nonlinear Recurrent Neural Network (RNN) architectures. Our results show that our stabilizing method significantly improves test performance of LDS, enabling such linear models to match the performance of contemporary nonlinear RNN architectures. A video of the obstacle avoidance performance of our method on a mobile robot, in unseen environments, compared to other methods can be viewed at https://youtu.be/mhEsCoNao5E.


    Access

    Download


    Export, share and cite



    Title :

    Gershgorin loss stabilizes the recurrent neural network compartment of an end-to-end robot learning scheme ; ICRA


    Contributors:

    Publication date :

    2020-01-01


    Remarks:

    Lechner M, Hasani R, Rus D, Grosu R. Gershgorin loss stabilizes the recurrent neural network compartment of an end-to-end robot learning scheme. In: Proceedings - IEEE International Conference on Robotics and Automation . IEEE; 2020:5446-5452. doi: 10.1109/ICRA40945.2020.9196608


    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English


    Keywords :

    Classification :

    DDC:    000 / 629



    On Bauer's generalized Gershgorin discs

    Deutsch, Eckart / Zenger, Christoph | TIBKAT | 1973


    EELS ICRA Presentation

    Tempest, Marco / Ono, Hiro / Hovsepian, Sarah et al. | NTRS | 2020


    Designing worm-inspired neural networks for interpretable robotic control ; ICRA

    Lechner, Mathias / Hasani, Ramin / Zimmer, Manuel et al. | BASE | 2019

    Free access

    ICRA: An Intelligent Clustering Routing Approach for UAV Ad Hoc Networks

    Guo, Jingjing / Gao, Huamin / Liu, Zhiquan et al. | IEEE | 2023


    Gaming boat demand stabilizes

    Online Contents | 1995