[Paper Review] Modeling and Validating Hybrid Systems Using VDM and Mathematica

Bernhard K. Aichernig, Reinhold Kainhofer|arXiv (Cornell University)|Jul 20, 2004

Real-Time Systems Scheduling5 references3 citations

TL;DR

This paper presents a hybrid modeling and validation approach for safety-critical control systems by integrating VDM-SL formal specifications with Mathematica’s symbolic and numerical computation capabilities. It enables simultaneous simulation and animation of discrete control logic and continuous physical dynamics—demonstrated on NASA’s SAFER life-support backpack—showcasing improved validation, error detection, and intuitive visualization for both engineers and stakeholders.

ABSTRACT

Hybrid systems are characterized by the hybrid evolution of their state: A part of the state changes discretely, the other part changes continuously over time. Typically, modern control applications belong to this class of systems, where a digital controller interacts with a physical environment. In this article we illustrate how a combination of the formal method VDM and the computer algebra system Mathematica can be used to model and simulate both aspects: the control logic and the physics involved. A new Mathematica package emulating VDM-SL has been developed that allows the integration of differential equation systems into formal specifications. The SAFER example from Kelly (1997) serves to demonstrate the new simulation capabilities Mathematica adds: After the thruster selection process, the astronaut's actual position and velocity is calculated by numerically solving Euler's and Newton's equations for rotation and translation. Furthermore, interactive validation is supported by a graphical user interface and data animation.

Motivation & Objective

• Address the gap in formal methods for validating hybrid systems by integrating continuous physical models into formal specifications.
• Overcome limitations of traditional formal tools that lack support for continuous mathematics and differential equations.
• Enhance confidence in formal models through executable testing and visual animation prior to formal proof.
• Provide a practical, interactive validation environment for control engineers and system designers using a familiar computational platform.
• Demonstrate the feasibility and benefits of combining formal specification (VDM-SL) with computer algebra systems (Mathematica) for hybrid system development.

Proposed method

• Developed a custom Mathematica package to emulate VDM-SL syntax and semantics within the Mathematica environment.
• Integrated systems of ordinary differential equations (ODEs) modeling Newton’s and Euler’s laws for 3D translational and rotational motion.
• Enabled on-the-fly numerical solution of ODEs during VDM-SL model execution for real-time simulation of physical behavior.
• Extended the VDM-SL specification of the SAFER control logic with physical dynamics, including thruster firing and angular velocity evolution.
• Implemented a graphical user interface (GUI) in Mathematica for interactive control, real-time data inspection, and animation of system trajectories.
• Supported pre- and post-condition checking within the VDM-SL environment to validate logical correctness of control decisions.

Experimental results

Research questions

• RQ1 Can a formal specification language like VDM-SL be effectively embedded within a computer algebra system to support hybrid system modeling?
• RQ2 How can continuous physical dynamics (e.g., Newtonian and Eulerian motion) be integrated into formal models for simulation and validation?
• RQ3 To what extent does combining discrete logic with continuous physics improve the detection of flawed assumptions in control system design?
• RQ4 Can interactive visualization and animation in Mathematica enhance communication between formal method experts and domain engineers?
• RQ5 What are the practical benefits of hybrid validation—combining testing and simulation of both discrete and continuous components—over traditional formal verification alone?

Key findings

• The integration of VDM-SL and Mathematica enables executable simulation of both discrete control logic and continuous physical dynamics in a single environment.
• The approach successfully simulated the SAFER system’s behavior under normal and failure conditions, including a broken thruster (6-F2), revealing that AAH could not compensate for torque unless translational commands were disabled.
• Visual animation of control cycles allowed clear observation of angular velocity drift and recovery, supporting intuitive understanding of system behavior.
• The method detected critical design flaws—such as suboptimal thruster selection logic—before formal proof, reducing the risk of proving the wrong model.
• Interactive GUI and real-time data inspection facilitated communication with domain experts and improved stakeholder confidence in the model.
• The hybrid validation approach demonstrated that continuous dynamics must be included in early validation to avoid incorrect assumptions that could invalidate formal proofs.