Session: DTM 08-01: Theory Development in Design Research

Paper Number: 194826

194826 - METRIX: A Metric-Driven Transformation Design Framework for Reconfigurable Kabaddi Robots 

Transformation design principles provide a systematic methodology for reconfigurable products and robots, yet existing applications assume cooperative contexts and rely on qualitative heuristics for principle selection. This paper introduces METRIX (Metric-driven Transformation Design), a pipeline of seven stages that maps competitive Kabaddi game actions to reconfigurable mechanism designs on the Smorphi modular robot platform. Kabaddi is a contact team sport in which robots must rapidly alternate between raiding and defending roles under adversarial conditions, demanding morphological reconfiguration capabilities that no existing robotic competition addresses. The pipeline employs Strategy Specification Cards organized into four facets (action geometry, dynamic requirements, interaction context, and system constraints), together with six quantitative evaluation metrics that drive principle selection through diagnostic questions and the established facilitator and generator hierarchy for mechanism realization. Five of the six metrics, namely mobility efficiency, transformation speed, contact stability, collision resilience, and adaptability, route through a deterministic diagnostic procedure with formally defined dominance thresholds to select among the three transformation principles (expand/collapse, expose/cover, fuse/divide), while the sixth metric, scoring effectiveness, validates the pipeline output. Applied to five representative Kabaddi actions, the pipeline produces 25 mechanism concepts (five per action) spanning all three transformation principles, engaging 16 of 20 facilitators, and utilizing all five Smorphi mounting zones. The central contribution is replacing intuitive principle selection with a structured procedure driven by quantitative metrics that ensures repeatability and supports systematic design space exploration. The framework is agnostic to the choice of sport, offering a generalizable pipeline for any competitive robotic domain that requires reconfigurable mechanisms.

Presenting Author: Manojkumar Devarassu Singapore University of Technology and Design

Presenting Author Biography: Manoj is an academic researcher specializing in Robotics & Automation, with over 8 years of experience in mechanical and smart product design. He holds a Master of Science in Smart Product Design from Nanyang Technological University (NTU) and a Bachelor of Engineering in Mechanical Engineering. His research contributions span multiple peer-reviewed publications in the field of robotics, and he has led seven novel robotics projects from conceptualization to full deployment. His work has been further recognized through several prestigious design awards.