Self-organization in a Simple Task of Motor Control Based on Spatial Encoding  

Oleh:

Linder, Christian R.

Jenis: Article from Journal - e-Journal
Dalam koleksi: Adaptive Behavior vol. 13 no. 3 (Sep. 2005), page 189–209.
Topik: self-organization; spatial encoding; DSOM; self-organizing map; motor control; neural mechanism
Fulltext: 189.pdf (599.09KB)

Isi artikel

This paper elaborates on the possibilities for self-adjustment of a biological neural network used as feedback controller in the motor control system of a six-legged walker. As biological systems, in contrast to technical systems, show an impressive capability of self-adaptation, this is meant as a proof of principle. Complementing an intensity encoded system (Linder, 2002), where scalar values are represented as the activity of a given neuron, this mechanism is based on spatial encoding, where a scalar value is represented as the location of the most active neuron in a chain of neurons. This encoding scheme can often be observed in biological systems. While the intensity encoded system requires linear input characteristics and symmetrical distribution of the input values over the whole range for both target angles and actual angles, the spatially encoded system presented here is completely self organizing for evenly distributed target angles and actual angles. By employing an internal teaching signal, it can even adjust for arbitrary (i.e., biologically relevant) distributions of the input. This internal signal is provided through body geometry. Instead of error back-propagation, the system exploits local neuronal mechanisms implicated by a biologically plausible realization of self-organizing maps.

Opini Anda

Klik untuk menuliskan opini Anda tentang koleksi ini!

Kembali