Vision-Alignment Mechanism for Feedback Control in Robotic Additive Manufacturing with Limited Degrees of Freedom

One of the key existing challenges in extrusion-based additive manufacturing is the presence of geometric inaccuracies resulting from material uncertainty. Although vision-based feedback control systems have demonstrated promise to control printing parameters to improve geometric fidelity in real time, their utility thus far has been limited to unidirectional linear toolpaths. This limitation arises in the presence of robotic manipulators with limited degrees of freedom and due to the dynamics of the system, such as in Selective Compliance Assembly Robot Arm robots. In such cases, maintaining a linear trajectory over a toolpath induces an axial rotation, and therefore, a vision attachment can work only if such rotation is countered, even in a linear one-dimensional toolpath. Vision-based feedback control systems for bidirectional and orthogonal toolpaths [e.g., two-dimensional (2D)] are therefore also hindered because they require additional step rotations at the corners of the toolpath. In this study, a vision alignment mechanism is developed and extended to a 2D toolpath, which adjusts the rotation of the camera with respect to the direction of a nonlinear toolpath to enable use in an orthogonal toolpath as well as the continuous axial rotation of the camera around the z -axis along the toolpath to counter that of the tool center point (TCP). This approach removes dependency in assessment and control of the filament geometry from the designed toolpath. To probe the utility of the vision alignment mechanism in a feedback control system, the mechanism is examined by implementing a feedback control system to enhance the geometry of the extruded filaments (e.g., width and height). The feedback control system includes two independent proportional-integral-derivative (PID) controllers to adjust the nozzle’s horizontal speed and its height. The results show that the vision alignment mechanism successfully adjusts the camera’s rotation and improves geometry measurements, which impact the results of the feedback control system, respectively.