- Main
Vehicle Lateral Control Under Fault in Front and/or Rear Sensors
Abstract
This report documents the findings of research performed under TO4204, "Vehicle Lateral Control under Fault in Front and/or Rear Sensors" during the year 2000- 2001. The research goal of TO4204 is to develop vehicle lateral control strategies under faulty operation of the magnetometers. The main objectives of the project are: (1) to design controllers that use the output from only one set of magnetometers, and (2) to develop an autonomous lateral control scheme that uses no magnetometers. New controllers that use just the front magnetometers have been designed based on H1 optimal design techniques in the previous research. In this year, the controller design has been continued by applying feedback linearization to eliminate the time varying terms caused by the variation of longitudinal velocity in the vehicle lateral dynamics. A mismatched observer has been designed to stabilize and tune the internal dynamics, and also to provide accurate state estimates for feedback linearization. Stability of the control system has been proved. Experimental results are presented to validate the controller design. An autonomous vehicle following control scheme is developed as a back-up system to replace the current lateral controller when all magnetometers fail. The new control scheme uses a laser scanning radar sensor, which measures the relative position of the controlled vehicle with respect to its preceding vehicle, to control the vehicle to follow its preceding vehicle. Experimental results of the autonomous vehicle following control are presented in this report. As shown in previous research, when the front magnetometers fail, the lanekeeping performance deteriorates substantially. A new back-up control system has been developed by combining the use of rear magnetometers with LIDAR. The design procedure guarantees to achieve the optimal combined use of the two sensors. Simulation results for this controller are presented in this report.
Main Content
Enter the password to open this PDF file:
-
-
-
-
-
-
-
-
-
-
-
-
-
-