IDDMBSE Framework Merges Data-Driven and Model-Based Engineering for Autonomous CPS

Autonomous cyber-physical systems (CPS) currently lack a unified systems engineering methodology that spans both Model-Based Systems Engineering (MBSE) and data-driven ML/AI. To address this, John Baras and colleagues from the University of Maryland propose IDDMBSE (Integrated Data-Driven and Model-Based Systems Engineering). This methodology extends the rigorous MBSE V-process with a data-driven loop at every stage, anchored in SysML, the autonomy stack, and a hybrid model-based plus data-driven trade-off architecture. The work is detailed in a 9-page paper with 11 figures, submitted to IEEE.

The team instantiates IDDMBSE as three interoperable open-source tools: PERFECT maps SysML system architectures to executable ROS 2 autonomy stacks for scalable performance evaluation; TRADES-X decomposes design-space exploration into model-based optimization followed by data-driven evaluation; VERITAS combines formal, data-driven, and runtime verification into a single assurance workflow. The methodology was validated on a Trusted Autonomous Ground Robot across its full lifecycle, including sensor-suite selection, risk-sensitive path planning, behavior-tree task verification, conformal-prediction-based perception, and multi-robot coordination. Experiments were conducted in a contested-terrain Isaac Sim test range released with the tool chain. The authors also sketch a reformulation on SysML v2/KerML for tighter ML/AI integration.