To perform self bias conditioning corresponding to an external environment and to improve the accuracy of a response at all times by autonomous learning.

In a 1st-order response network 1, signals for indicating whether or not the drive of an actuator C by a motor control node M is appropriate are received from a sensor SB. Based on the signals, in an associative memory 2, weighting in a positive or negative direction is performed to the signals inputted through other sensors S1-SN and thus, the conditioning of the response is performed.