CLSE: Closed-Loop Symbolic Execution

We present CLSE, a closed-loop symbolic execution engine for control system implementations. CLSE takes as input the description of a physical plant represented by a system of linear ordinary differential equations, the software implementation and execution frequency for a discrete-time controller that senses and actuates the plant, and a time horizon, and symbolically executes the closed-loop system —the combination of the plant and the controller— up to the time horizon. The execution helps capture the bounded-time dynamics of the system in terms of the finite sequences of the plant’s sampled state-sets and symbolic control inputs. We show the use of CLSE in symbolic execution of a set of control systems benchmarks. Using the symbolic execution engine, we also build a robustness analysis tool which computes the maximum deviation of the states of the plant due to measurement uncertainties in the controller up to the time horizon.