Autonomous four-wheeled robots have been widely studied and developed for various purposes over several decades. We have developed an All Terrain Vehicle(ATV)-based, four-wheeled, car-like robot to support the environmental field survey of landfills. Navigating the robot toward an observation spot must have feedback control of the vehicle. We propose a path-generating regulator (PGR) for car-like robots and show the properties of its control performance. Originally, the PGR was a control method for two-wheeled mobile robots to converge at the origin of a coordinate frame, of which the heading angle is controlled so as to align the tangential angle of one of the path among the path function group. Unlike other control methods of the nonholonomic system, neither a coordinate transformation nor an input transformation is needed in this method. In this paper, the PGR is extended to car-like robots. We take care of the singular points, due to both of the mechanism of a car-like robot and the control method. First we show that the origin of the control system is stable in the sense of Lyapunov in accordance with Lyapunov's theory. Next, we discuss initial state convergence at the origin. We examine the performance of the control method via simulations, and apply the method to the ATV-based, car-like robot. As a result, we show the property of convergence to the origin in the car-like robot using the proposed control method.