Friction drag is reduced in a two-dimensional (2D) boundary-layer flow by delaying the laminar-to-turbulence transition. A localised forcing in the wall region is used to attenuate the growing 3D disturbances that eventually trigger the turbulent regime farther downstream. An adaptive filtered-X least-mean-squared (FXLMS) algorithm is used to process the information of the flow gathered from two rows of surface hot-wires sensors and compute the forcing, performed by a row of plasma actuators. LES simulations are used to evaluate and analyze the performance of the described control strategy: in particular, a study on the streamwise position of the sensor and an actual transition delay scenario are presented.