Future complex systems are becoming increasingly modular and reconfigurable. In aerospace, robotics, industrial automation, and cyber-physical infrastructure, system configurations may change as components are added, removed, exchanged, or fail. Today, such changes often require manual integration, configuration, and validation. This creates a major challenge for engineering systems that must remain flexible while maintaining a clear understanding of their structure, capabilities, and operational state.

This project presents a self-describing modular robotic platform for model-based adaptive complex systems. The platform combines physical modularity, automatic component discovery, domain-specific runtime modeling, and bidirectional synchronization between hardware and model. Its purpose is to explore how technical systems can recognize their configuration, describe capabilities, and make adaptation understandable through model-based representations.

The physical design of the robotic platform is simple, modular, and readable. A central core module acts as the structural and electronic hub. Functional modules are attached around this core using standardized D-Sub connectors, creating a robust and visible plug-in architecture. The box-like geometry is not only a construction choice, but also part of the interaction design: users can immediately understand where modules belong, how the system is extended, and how physical configuration changes relate to system behavior. This makes the platform accessible for demonstrations while supporting engineering experimentation.

The interchangeable modules can represent different capabilities, such as mobility, sensing, actuation, energy supply, communication, or mission-specific functions. When a module is connected or removed, the system detects the change. Each recognized module is associated with a machine-readable description of its type, properties, and capabilities. Based on this information, the system updates its runtime representation and makes the changed configuration available to higher-level software components.

A central innovation is the connection between the physical modular design and Domain-Specific Modeling. Instead of exposing users to low-level code, wiring, or raw data structures, the system represents its architecture through a domain-specific model. Components, ports, capabilities, and relationships become visible in a structured form. Commands and configuration changes can be propagated from the modeling environment back to the platform, enabling bidirectional interaction between model and hardware.

This model-based approach makes the system understandable beyond expert users. Visitors do not need to know embedded programming or system architecture to understand what is happening. They can attach a physical module, observe the platform recognize it, see the model update, and understand how available capabilities change. The model acts as a bridge between the physical object and the underlying software behavior.

The platform provides a foundation for adaptive behavior in complex modular systems. Since the system maintains knowledge about its current components and capabilities, it can support runtime monitoring, configuration management, capability-based control, and dynamic task allocation. This makes the concept relevant beyond robotics, including Plug-and-Fly Avionics, modular spacecraft, satellite constellations, reconfigurable industrial systems, and adaptive cyber-physical infrastructures.

By combining readable modular hardware design with self-description, Domain-Specific Modeling, and runtime synchronization, the platform turns invisible engineering concepts into an interactive physical experience. It demonstrates how complex systems may configure, describe, and reason about themselves while remaining understandable to engineers and non-expert audiences.