Optimal Modified Feedback Strategies in LQ Games under Control Imperfections