Steering micro-objects using acoustic radiation forces is challenging for several reasons: resonators tend to create fixed force distributions that depend primarily on device geometry, and even when using switching schemes, the forces are hard to predict a priori. In this paper an active approach is developed that measures forces from a range of acoustic resonances during manipulation using a computer controlled feedback loop based in matlab, with a microscope camera for particle imaging. The arrangement uses a planar resonator where the axial radiation force is used to hold particles within a levitation plane. Manipulation is achieved by summing the levitation frequency with an algorithmically chosen second resonance frequency, which creates lateral forces derived from gradients in the kinetic energy density of the acoustic field. Apart from identifying likely resonances, the system does not require a priori knowledge of the structure of the acoustic force field created by each resonance. Manipulation of 10 μm microbeads is demonstrated over 100 s μm. Manipulation times are of order 10 s for paths of 200 μm length. The microfluidic device used in this work is a rectangular glass capillary with a 6 mm wide and 300 μm high fluid chamber.