When a child unexpectedly darts out in front of an automobile, the driver must know immediately whether or not it is possible to slow down in time.  If the child is too close or the automobile is moving too quickly, then the driver may need to swerve out of the way.  This example illustrates, in a very intuitive sense, how we in the PandA Labs think about the visual control of braking.  Our research is guided by the principle that successful performance on tasks such as braking depends on the ability to perceive whether or not an action is possible.  Likewise, braking is controlled by making adjustments so that it is always possible to stop within the limits of the brake.  The greatest challenge is understanding how tasks like braking are successfully performed across changing conditions.  In the real world, the braking task is performed across a wide range of speeds, in the presence of external disturbances (e.g., wind, gravity, surface friction), across variations in controller and plant dynamics, and with a small tolerance for error.  Such factors introduce a variety of critical constraints on the success of braking, and alter the effectiveness of different sources of information and control strategies. The aim of this project is to formulate a principled account of how people adapt to changes in these constraints by learning to detect, calibrate, and use more effective sources of information with practice and experience.