This two-part study analyzed occupant kinematics in simulated collisions of future automated driving vehicles in terms of seating configuration. In part one, a frontal collision was simulated with four occupants with the front seats reversed. The left front seat occupant was unbelted while the others were belted. In part two of the study, occupant restraint was examined in various seating configurations using a single seat model with a three-point seatbelt. The seat direction with respect to impact was considered as forward, rearward, and lateral facing in 45 degree increments. The effect of seat recline was also studied in the forward-facing and rear-facing cases by assuming three positions:​ driving position, resting position and relaxed position. Occupants were represented by human body finite element models. The results of part one showed that the front seat (rear-facing) occupants were restrained by the seatback, resulting in T1 forward displacement less than 100 mm;​ the rear seat occupants were restrained by the seatbelt resulting larger T1 forward displacement more than 500 mm. The results of the part two showed the directional dependence of occupant restraint. Greater T1 displacements were observed when the occupant faced lateral or front oblique. However, the seatbelt provided some restraint in all directions considered. The seatback generated contact force to the occupant when it was in the impact direction, including the lateral directions. The relaxed position allowed increased excursion compared to the driving position when the occupant faced rearward, but the magnitude of this increase was lower with lower impact speed.