Image-guided robots have been widely used for bone shaping and percutaneous access to interventional sites. However, due to high-accuracy requirements and proximity to sensitive nerves and brain tissues, the adoption of robots in inner-ear surgery has been slower. In this paper the authors present their recent work towards developing two image-guided industrial robot systems for accessing challenging inner-ear targets. Features of the systems include optical tracking of the robot base and tool relative to the patient and Kalman filter-based data fusion of redundant sensory information (from encoders and optical tracking systems) for enhanced patient safety. The approach enables control of differential robot positions rather than absolute positions, permitting simplified calibration procedures and reducing the reliance of the system on robot calibration in order to ensure overall accuracy. Lastly, the authors present the results of two phantom validation experiments simulating the use of image-guided robots in inner-ear surgeries such as cochlear implantation and petrous apex access.