Advertisement

Learning Visually Guided Risk-Aware Reaching on a Robot Controlled by a GPU Spiking Neural Network

  • Terence D. Sanger
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 9947)

Abstract

Risk-aware control is a new type of robust nonlinear stochastic controller in which state variables are represented by time-varying probability densities and the desired trajectory is replaced by a cost function that specifies both the goals of movement and the potential risks associated with deviations. Efficient implementation is possible using the theory of Stochastic Dynamic Operators (SDO), because for most physical systems the SDO operators are near-diagonal and can thus be implemented using distributed computation. I show such an implementation using 4.3 million spiking neurons simulated in real-time on a GPU. I demonstrate successful control of a commercial desktop robot for a visually-guided reaching task, and I show that the operators can be learned during repetitive practice using a recursive learning rule.

Keywords

Stochastic control Spiking network Reaching Optimal feedback control Stochastic Dynamic Operators Risk-aware control 

References

  1. 1.
    Bertucco, M., Bhanpuri, N.H., Sanger, T.D.: Perceived cost and intrinsic motor variability modulate the speed-accuracy trade-off. PloS one 10(10), e0139988 (2015)CrossRefGoogle Scholar
  2. 2.
    Dunning, A., Ghoreyshi, A., Bertucco, M., Sanger, T.D.: The tuning of human motor response to risk in a dynamic environment task. PloS one 10(4), e0125461 (2015)CrossRefGoogle Scholar
  3. 3.
    Sanger, T.D.: Distributed control of uncertain systems using superpositions of linear operators. Neural Comput. 23(8), 1911–1934 (2011)MathSciNetCrossRefzbMATHGoogle Scholar
  4. 4.
    Sanger, T.D.: Risk-aware control. Neural Comput. (2014)Google Scholar
  5. 5.
    Sanger, T.D.: Probability density methods for smooth function approximation and learning in populations of tuned spiking neurons. Neural Comput. 10(6), 1567–1586 (1998)CrossRefGoogle Scholar

Copyright information

© Springer International Publishing AG 2016

Authors and Affiliations

  1. 1.Department Biomedical EngineeringUniversity of Southern CaliforniaLos AngelesUSA

Personalised recommendations