Over recent decades, advanced driver assistance systems have been developed to aid human drivers in tasks such as lane keeping, cruise control, collision avoidance, and blind spot recognition. To enhance the efficacy of these assistance systems, cooperative or shared control approaches involving the driver have been extensively investigated. This study introduces a cooperative lane-keeping assist architecture that adapts to driver torque and validates it experimentally using a prototype vehicle. Employing the kinematic model of the vehicle, a proportional-integrative (PI) autonomous controller is devised, utilizing lane deviation error and heading error as output variables. For cooperative control implementation, an assistance factor is determined based on the correlation between driver torque and the required assistance level. Consequently, the overall control exerted on the vehicle is a blend of the driver and autonomous controller steer inputs. The effectiveness of the proposed cooperative control approach is experimentally verified on a test track, with illustrative results presented.