Abstract
Professional baseball players somehow manage to intercept and catch fly balls successfully. Three perceptual strategies have been proposed to explain the outfielder's ability and to predict the fielder's interception path. (1) TP: The trajectory prediction strategy (Saxberg, 1987) proposes that the fielder perceives the ball's initial conditions, computes its trajectory, and runs directly to the predicted landing point. (2) OAC: The optical acceleration cancellation strategy (Chapman, 1968) proposes that the fielder continuously controls their radial motion (toward the ball) by running to null the vertical optical acceleration of the ball. Transverse motion (lateral) is controlled independently by holding the bearing direction of the ball constant. (3) LOT: The linear optical trajectory strategy proposes that the fielder runs so as to keep the ball's rising optical trajectory straight rather than curved; this maintains a constant ratio between the tangents of the ball's elevation and azimuth angles (tana/tanb = c). These strategies are normally highly correlated, but they can be tested in VR by manipulating the ball's trajectory. Brown University players were tested in an immersive virtual environment (12 x 12 m) while wearing a head-mounted-display (60 x 40 deg, 50–70 ms latency) with trackers on the head and glove. The ball's trajectory was manipulated by increasing or decreasing the gravitational constant (g) at the peak of flight. Trials in which gravity was changed were matched with normal trials in which the ball traveled the same distance on a parabolic arc. Participants made rapid and continuous adjustments to compensate for the change in gravity, inconsistent with TP. Moreover, their radial and transverse velocities varied independently, inconsistent with LOT. Behavior was most consistent with the combination of OAC and constant bearing strategies. We describe a dynamical model that implements this control strategy.