Parametric Generative Design Framework for Serial Robots (September 2026).

This paper was presented by Caitlin Sands at the Robot Design, Dynamics and Control – 26th CISM-IFToMM Symposium (RoManSys 2026) – Home – RoManSy 2026

Abstract: Robot design has traditionally relied on manual, experience-driven processes that can be time-consuming and limited in exploring complex design spaces. In response, generative design (GD) has emerged as a powerful tool for exploring robotic architectures, yet existing frameworks often rely on optimisation or AI-driven methods that are computationally intensive and offer limited transparency. This paper presents a system-level parametric GD framework for serial robots, including link parametric representation, link library creation, type synthesis of serial robots, automatic kinematics modelling and workspace generation within a single procedural workflow. A digital tool has been developed in MATLAB to demonstrate the feasibility of the proposed framework. This approach provides a balance between manual design and fully autonomous generative methods, delivering a transparent and computationally efficient solution for rapid design of robots.