Lazy linear hybrid automata (LLHA) model the discrete time behavior of control systems containing finite-precision sensors
and actuators interacting with their environment under bounded inertial delays. In this paper, we present a symbolic technique
for reachability analysis of lazy linear hybrid automata. The model permits invariants and guards to be nonlinear predicates
but requires flow values to be constants. Assuming finite precision, flows represented by uniform linear predicates can be
reduced to those containing values from a finite set of constants. We present an abstraction hierarchy for LLHA. Our verification
technique is based on bounded model checking and k-induction for reachability analysis at different levels of the abstraction
hierarchy within an abstraction-refinement framework. The counterexamples obtained during BMC are used to construct refinements
in each iteration. Our technique is practical and compares favorably with state-of-the-art tools, as demonstrated on examples
that include the Air Traffic Alert and Collision Avoidance System (TCAS).
Supported in part by SRC contract 1355.001, NSF grants CNS-0644436 & CNS-0627734, and Microsoft Research. The first author
was also supported by the Berkeley Fellowship for Graduate Studies from UC Berkeley.