Increasing air traffic in the U.S.A has led to runway capacity limitations at the airports. Increasing the capacity of the existing runways involves reducing the runway occupancy time of an aircraft landing on a runway. The location of runway exits plays an important role in minimizing the runway occupancy time. Locating an optimal location for an exit is getting complex with a rapid increase in the number of aircraft types. So, the Air Transportation and Systems Laboratory at Virginia Tech developed the Runway Exit Interactive Design Model (abbreviated as REDIM). This model finds the optimal exit location considering multiple aircraft and a variety of environmental conditions. To find the optimal exit location, REDIM simulates the landing aircraft behavior. The kinematic model simulating the aircraft landing behavior in REDIM using pseudo-nonlinear deceleration heuristic algorithm. REDIM models the aircraft landing behavior into five phases. The five phases are: 1) a flare phase, 2) a free roll period occurring between the aircraft touchdown and the brakes initiation 3) the braking phase, 4) a second free roll phase starting after the braking phase and ending before the turnoff maneuver and 5) a turnoff maneuver phase. The major contributors to the runway occupancy time (ROT) are the braking phase (60% of ROT) and the turnoff phase (25% of ROT). Calculating the turnoff time requires few input variables such as deceleration rate along the turnoff and the speed at which an aircraft takes an exit (exit speed at the point of curvature). The deceleration rate along the turnoff is specific to every aircraft. This study involves predicting the exit speed at the point of curvature based on the type of exit taken. It begins with collecting the exit geometry parameters of 37 airports in the U.S.A. The exit geometry parameters define the type of exit. The ASDE-X data provides the observed exit speeds at the point of curvature for these exits. This study examines a few models with observed exit speeds as the response variable and exit geometry as the predictor variables. / MS / Increasing air traffic in the U.S.A has led to runway capacity limitations at the airports. Increasing the capacity of the existing runways involves reducing the runway occupancy time of an aircraft landing on a runway. The location of runway exits plays an important role in minimizing the runway occupancy time. Locating an optimal location for an exit is getting complex with a rapid increase in the number of aircraft types. So, the Air Transportation and Systems Laboratory at Virginia Tech developed the Runway Exit Interactive Design Model (abbreviated as REDIM). This model finds the optimal exit location considering multiple aircraft and a variety of environmental conditions. To find the optimal exit location, REDIM simulates the landing aircraft behavior. The kinematic model simulating the aircraft landing behavior in REDIM using pseudo-nonlinear deceleration heuristic algorithm. REDIM models the aircraft landing behavior into five phases. The five phases are: 1) a flare phase, 2) a free roll period occurring between the aircraft touchdown and the brakes initiation 3) the braking phase, 4) a second free roll phase starting after the braking phase and ending before the turnoff maneuver and 5) a turnoff maneuver phase. The major contributors to the runway occupancy time (ROT) are the braking phase (60% of ROT) and the turnoff phase (25% of ROT). Calculating the turnoff time requires few input variables such as deceleration rate along the turnoff and the speed at which an aircraft takes an exit (exit speed at the point of curvature). The deceleration rate along the turnoff is specific to every aircraft. This study involves predicting the exit speed at the point of curvature based on the type of exit taken. It begins with collecting the exit geometry parameters of 37 airports in the U.S.A. The exit geometry parameters define the type of exit. The ASDE-X data provides the observed exit speeds at the point of v curvature for these exits. This study examines a few models with observed exit speeds as the response variable and exit geometry as the predictor variables.