In this paper, we propose a novel system architecture and control scheme, a human-machine perceptual feedback system and control, to enhance a user's control performance while he is teleoperating a mobile robot with a joystick. First, we model the user, robot, and human-machine perceptual feedback controller. The two key roles of the controller are: a) displaying a preferred road width to the user with respect to various road types, and b) constraining the maximum linear and angular velocity of the robot for safer maneuver as well. Second, we define a perception to behavior stability and perform a stability analysis. Finally, we present simulation results showing that a novice user can teleoperate the mobile robot successfully along sidewalks by our approach.