Discretized ISO-learning neural network for obstacle avoidance in reactive robot controllers

Isotropic sequence order learning (ISO-learning) and its variations, input correlation only learning (ICO-learning) and ISO three-factor learning (ISO3-learning) are unsupervised neural algorithms to learn temporal differences. As robotic software operates mainly in discrete time domain, a discretization of ISO-learning is needed to apply classical conditioning to reactive robot controllers. Discretization of ISO-learning is achieved by modifications to original rules: weights sign restriction, to adequate ISO-learning devices outputs to the usually predefined kinds of connections (excitatory/inhibitory) used in neural networks, and decay term in learning rate for weights stabilization. Discrete ISO-learning devices are included into neural networks used to learn simple obstacle avoidance in the reactive control of two real robots.

[1]  Florentin Wörgötter,et al.  ISO Learning Approximates a Solution to the Inverse-Controller Problem in an Unsupervised Behavioral Paradigm , 2003, Neural Computation.

[2]  Florentin Wörgötter,et al.  Temporal Hebbian Learning in Rate-Coded Neural Networks: A Theoretical Approach towards Classical Conditioning , 2001, ICANN.

[3]  Wulfram Gerstner,et al.  Mathematical formulations of Hebbian learning , 2002, Biological Cybernetics.

[4]  José Manuel Cuadra Troncoso,et al.  Discretization of ISO-Learning and ICO-Learning to Be Included into Reactive Neural Networks for a Robotics Simulator , 2007, IWINAC.

[5]  Florentin Wörgötter,et al.  Strongly Improved Stability and Faster Convergence of Temporal Sequence Learning by Using Input Correlations Only , 2006, Neural Computation.

[6]  V. Braitenberg.  Vehicles, Experiments in Synthetic Psychology , 1984 .

[7]  Florentin Wörgötter,et al.  Stabilising Hebbian Learning with a Third Factor in a Food Retrieval Task , 2006, SAB.

[8]  Florentin Wörgötter,et al.  Isotropic-sequence-order learning in a closed-loop behavioural system , 2003, Philosophical Transactions of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences.

[9]  Wulfram Gerstner,et al.  Intrinsic Stabilization of Output Rates by Spike-Based Hebbian Learning , 2001, Neural Computation.

[10]  Florentin Wörgötter,et al.  Isotropic Sequence Order Learning , 2003, Neural Computation.

[11]  Bernd Porr,et al.  Sequence-learning in a self-referential closed-loop behavioural system , 2003 .

[12]  Florentin Wörgötter,et al.  Improved stability and convergence with three factor learning , 2007, Neurocomputing.