FMU-Based Multi Criteria Trajectory Validation of Industrial Robots

This paper presents a modular, multi-criteria trajectory validation framework for industrial robot digital twins, implemented as Functional Mock-up Units (FMUs) compliant with the FMI 2.0 Co- Simulation standard and integrated within a ROS 2 and Gazebo Ignition runtime pipeline. The framework evaluates planned trajectories generated by MoveIt 2 against three complementary criteria: dynamics feasibility via a Recursive Newton–Euler inverse dynamics solver, safety compliance through a two-tier architecture combining a physically-grounded hard gate at the critical joint margin (0.02 rad, derived from the UR10e’s 125 Hz servo control characteristics) with a weighted geometric mean soft scoring layer, and motion smoothness via a length-normalized Gaussian jerk decay function.Aratio-based test-level aggregation mechanism inspired by the IEC 61508 distinction between systematic and random failures classifies hard-failure patterns and prevents weighted score averaging from masking genuine safety deficiencies. Experimental evaluation across 450 tests—spanning two industrially-representative scenarios (chassis inspection and rapid workspace traversal), three OMPL planners (RRTConnect, RRT*, EST), three velocity scales, and five weight configurations—demonstrates the framework’s discriminant power: the same planner and velocity configurations yield 96.4% PASS for rapid traversal yet 0% PASS for chassis inspection, where all tests are classified as systematic failures due to trajectory-level joint margin violations that are invisible to goal-position verification. The score paradox—where 96% of inspection tests achieve soft scores above the PASS threshold yet all receive FAIL verdicts—validates the necessity of the two-tier architecture for safety-critical trajectory assessment in digital twin environments.