This study conducted a series of rear-impact, side-impact, barrier, and free-operation collisions using a bumper car ride at an active amusement park. Two conditions were studied: staged and free operation. Each staged test included a bullet (impacting) vehicle operated by a rider and a target (impacted) static vehicle or structure. Impact configurations of frontal collisions of the bullet vehicle into the rear and side of a target vehicle were consistent with the existing literature. The free operation condition involved collisions which were not pre-determined, and operators may not have been prepared for collision timing, magnitude, and direction. Results demonstrated high repeatability for vehicle parameters, such as impact velocity, change in velocity, and peak acceleration. Peak changes in velocity during vehicle-to-vehicle collisions were 2.2-2.5 m/s (8-8.9 km/hr; 5-5.5 mph) for the target vehicle and 1.6-1.8 m/s (5.6-6.4 km/hr; 3.5-4 mph) for the bullet vehicle, while those during vehicle-to-retaining barrier collisions were approximately 3.6 m/s (13 km/hr; 8 mph). Coefficients of restitution and overall vehicle and occupant kinematics were similar to prior bumper car studies, and collision magnitudes were similar in the free-operation test to the staged, single-axis collisions. Bumper cars present a model environment to study vehicle and occupant kinematics in vehicle collisions that are within human tolerance and include aware but possibly unprepared occupants. This is relevant to establishing occupant kinematics in and limits to autonomous vehicle emergency handling maneuvers.