We use modeling and simulation technologies to streamline the design, implementation, and test phases of our projects. System models give us the capacity to predict and trade off performance parameters to get the design right before hardware implementation. Simulators provide a "robot on every desktop" of our developers so they can evaluate their software as it moves the robot over terrain, senses the environment, and avoids obstacles.

Unlike many organizations, we invest heavily in getting our models right. Our intelligent systems must use accurate internal models of their surroundings just to be intelligent as it performs in the actual environment. We model both the systematic and stochastic behavior of: