摘要 |
PROBLEM TO BE SOLVED: To prevent longitudinal-slide shift control from being executed in a state in which an input torque is not less than a predetermined value, thereby improving the durability of a belt and a pulley and improving responsiveness of accelerator operation after execution of the longitudinal-slide shift control.SOLUTION: An initial irregularity learning control is executed whenever learning conditions are satisfied with views of absorbing initial irregularities such as manufacturing errors and irregularities in constituent elements of a second clutch CL2. A total number of times of learning is counted, and if the total number of times of learning for the initial irregularity learning control reaches a predetermined number (five, for example), it is determined that the initial irregularity learning control converges. On condition of this learning control convergence, a so-called longitudinal-slide shift control is executed for returning a belt 50 of a variator V to a position of lowest gear ratio. As a result, the longitudinal-slide shift control is not executed in a state in which an excessive torque acts, so that durability of the belt and a pulley improves. |