In order to more accurately simulate the load distributions and histories experienced by automotive seats in field use, more biofidelic motion and loading devices are needed. Lear and KUKA have developed a system capable of controlling the coordinated motions of a pelvis, thighs and torso dummy in order to mimic human motions. The system takes kinematic data collected from human trials and converts them directly to a robot program. Additionally, simultaneous measures of human loading using pressure distribution mats can be obtained, and these measures are used as the basis for teaching the robot to correct the kinematic data using a neural net learning algorithm. The robot has direct and indirect load feedback integration that allows the load to be precisely maintained throughout the duration of a cycle test.