A system is provided that includes a machine assembly, a first imaging sensor, an encoder, and one or more processors. The machine assembly is movable to actuate a brake lever of a vehicle in order to open a valve of an air brake system. The first imaging sensor is positioned to acquire two-dimensional perception information of a working environment that includes the brake lever during movement of the machine assembly towards the brake lever. The encoder detects a displacement of the machine assembly relative to a reference position of the machine assembly. The one or more processors estimate a target position of the brake lever relative to the machine assembly during movement of the machine assembly based on the two-dimensional perception information and the displacement. The one or more processors drive the movement of the machine assembly towards the target position of the brake lever.