An apparatus and method for training a landmark detector receives training data which includes a plurality of positive training bags, each including a plurality of positively annotated instances, and a plurality of negative training bags, each including at least one negatively annotated instance. Resource consumption on a hardware device is monitored. The differential pressure may be compared with a known benchmark value to determine whether the differential pressure has changed. Not only can the problem that inter-cell interference is introduced during sub-frame switching in the TDD system be solved, but also the normal uplink and downlink measurement can be guaranteed, without affecting ACKNACK feedback and HARQ process, and meantime the frequency spectrum utilization ratio is increased.