摘要 |
PURPOSE: To immediately return from an abnormal state of shift control to a normal state by a method wherein when a shift time exceeds an allowable time and is continued for a long time due to abnormality of a value inputted to control computation, a line pressure is set to a preferable given state like, for example, an initial value during delivery from a factory. CONSTITUTION: An automatic transmission 2 is provided with a torque converter 10 and a shift gear mechanism 11 and controls a control valve 15 driven according to a command from a computer 14 for shift control, and an expected shift is realized by operating various friction elements by means of an oil pressure. In this case, measurement of a shift time is started according to a shift command and when a shift is decided to complete, measurement of a shift time is completed. When it is decided that the determined shift time is longer than a shift allowable time, it is decided that some trouble occurs to setting of an initial line pressure through learning control and number of revolutions feedback control and learning correction of a line pressure is initilized. |