A novel dynamic radiographic imaging system, one which tracks the patient during an activity, can provide measurements of natural skeletal motion. It is believed that accurate measurements of skeletal kinematics can improve both the treatment and diagnosis of musculoskeletal conditions. However, the measurement and control obstacles of such a proposed system have not been overcome. Using heavy radiographic equipment in combination with lightweight manipulators leads to joint deflections and degraded tracking performance. Furthermore, relative motion between imaging components leads to motion blur and degraded measurement performance. Novel static and dynamic identification methods are developed and executed in order to ascertain kinematic, inertial, frictional, and stiffness properties for model-based observer and controller methods. An extended Kalman filter and adaptive reduced parameterization Lyapunov-based controller are developed in order to control the robotic imaging system. The effect of motion blur on 3-D to 2-D image registration is also explored, given the anticipated relative motion between components.