Figure A1.

The recurrent neural network consisting of 2 × 6 units. The network is composed of two identical networks (The one for the x-components is shown with black lines and the other for the y-components of the vectors is shown with the dashed lines). The units represent the components of the six vectors L₁, L₂, L₃, D₁, D₂, and R of the manipulator. When an input is given, the corresponding recurrent channel is suppressed (symbolized by the open arrow heads).