This paper presents a method for trajectory planning, threatassessment, and semi-autonomous control of manned andunmanned ground vehicles. A model predictive controlleriteratively replans a stability-optimal trajectory through the saferegion of the environment while a threat assessor and semi-autonomous control law modulate driver and controller inputs tomaintain stability, preserve controllability, and ensure that thevehicle avoids obstacles and hazardous areas. The efficacy of thisapproach in avoiding hazards while accounting for various typesof human error, including errors caused by time delays, isdemonstrated in simulation.