Kaa: A Python Implementation of Reachable Set Computation Using Bernstein Polynomials

Abstract

Reachable set computation is one of the many widely-used techniques for the verification of safety properties of dynamical systems. One of the simplest algorithms for computing reachable sets for discrete nonlinear systems uses parallelotope bundles and Bernstein polynomials. In this paper, we describe Kaa, a terse Python implementation of reachable set computation which leverages the widely used symbolic package sympy. Additionally, we simplify the user interface and provide easy-to-use plotting utilities. We believe that our tool has pedagogical value given the simplicity of the implementation and its userfriendliness.