Robotic Additive Manufacturing Path Planning via HPC

Most three-dimensional printing systems use a three-axis gantry configuration and build objects as a sequence of planar layers. Emerging robot based systems have the ability to work outside of this layered sequence and deposit material in more intricate patterns such as weaving or depositing on non-planar surfaces. The majority of software tools to produce paths target the layer by layer limitations of the gantry type systems and new software is required to realize the full capabilities of robotic additive manufacturing systems. This work will address this problem by applying recent advances in path planning for multi-axis subtractive manufacturing to analogous situations in robot based additive manufacturing. The technology is enabled by the advent of affordable high performance computing equipment in the form of many core graphical processing units and multi-core central processing units. Phase I will attempt to produce a part using an industrial robot arm with a material depositing end effector. A robot path for depositing the material such that it prints the desired object correctly will be computed using high performance computing. Experiments running robotic additive systems with the generated paths will be conducted. This pilot project will flush out the requirements necessary in a follow-on phase II effort to democratize this technology for non-expert high performance computing users and scale to computation on higher end high performance computing systems. Robots can drop pieces of plastic or metal in a pattern to build something but new software is needed to make the path the robot needs to follow. This work will build software that runs on high performance computers to determine the correct printing paths for the robots. Commercial Applications and Other Benefits: A successful effort through all phases of this work would result in a complete robot additive manufacturing solution encompassing complimentary hardware and software. Such a system would reduce manufacturing costs and enable more automated flow from product concept to physical product. Additionally, the software solution proposed allows non-expert users to leverage high performance computing and thereby opening potential new markets for such computing hardware.