Optimal performance of adaptive restraint systems in the vehicle requires an accurate assessment of occupant characteristics including physical properties and pre-collision response of the occupant. To provide a feasible framework for incorporating occupant characteristics into adaptive restraint schemes, this study evaluates the sensitivity of injury risk in frontal collisions to four occupant parameters:​ mass, stature, posture and bracing level. The numerical approach includes using commercial multi-body software to develop occupant models that span a range of occupant parameters representative of the real-world driver population. Coupled with a multi-body model of the vehicle interior and standard restraint system, risk of occupant injuries within specific body regions are predicted through numerical simulations in conjunction with established injury risk functions. The results show occupant posture to be the most significant parameter affecting the overall risk of injury in frontal collisions. The causal relationship as predicted using the numerical model has been compared to the traffic injury epidemiology findings, and the feasibility of an analytical methodology to provide real-time estimates of injury severity has been discussed. Preliminary estimates from the study indicate that the proposed methodology will provide a framework to optimize restraint performance and potentially reduce the risk of injuries up to 35% (based on parameter-specific optimization), using accurate information regarding the pre-collision occupant characteristics.