An Expert Fuzzy Controller for Vehicle Lateral Control

Abstract

Vehicle lateral control is a challenging problem in areas of Intelligent Transportation Systems (ITS) and automatic control for autonomous vehicles. Most of solutions for lateral control proposed, such as PID, simple fuzzy logic control and Hinfinity, are proved accurate and effective only in simple driving task, such as lane changing, lane following and small curvature crossing turning action. For big curvature crossing turning action, the solutions mentioned above haven't dealt with. In this paper, we propose an expert fuzzy controller, which is based on a 2 Degree of Freedom (2DOF) vehicle lateral dynamic model, to solve the problem both in simple driving task and big curvature crossing turning action. The traditional 2DOF analytic model is used in lane changing, lane following and small curvature crossing turning actions, while an experimental model is used in big curvature crossing turning. That is because the traditional 2DOF analytic model is set up based on the hypothesis that the steering angle of front wheel is small. The experimental model is proved accurate and effective in the realistic driving experiments.In addition, it is proved that the expert fuzzy controller is effective to solve the lateral control problem in complicated driving tasks, such as lane changing, lane following, small and big curvature crossing turning actions, in both simulations and experiments.

Fuzzy-Model-Based Output Feedback Steering Control in Autonomous Driving Subject to Actuator Constraints

“…Hence, with the structure of output feedback controller in (14), the control signal δ f (k) also follows the rate constraint inherently.…”

“…For a given scalar λ > 0, design a dynamic output feedback compensator in ( 14) satisfying: 1) the path tracking system in (6) with uncertainties and actuator constraints can asymptotically follow the given path with zero offset under the output feedback controller (14) when the road curvature is 0, i.e., w(k) = 0; 2) under zero initial condition, the following inequality holds ∥z(k)∥ l2 ≤ λ ∥w(k)∥ l2 (19) for any different types of path with w(k) belonging to l 2 [0, ∞).…”

Multi-model fuzzy controller for vehicle lane tracking

“…However, the actual vehicle often present strong coupling, nonlinearity, uncertainty, and a wide operating range. The normal fuzzy algorithm is difficult to resolve these problems, which are caused by the fact that the structure of normal fuzzy controller is too simple, the subjective factor is too strong in the controller design and fuzzy rules can not be modified once they are established [19].…”

Hierarchic Fuzzy Approach Applied in the Development of an Autonomous Architecture for Mobile Agents

“…According to Khodayari [10] there are two main strategies to develop autonomous agents. The first one is related to heuristic knowledge and is represented in this paper by fuzzy logic as presented in [11,12,13], and adapted to the model used. The second one is based in the control theory, subdivided in linear control theory and nonlinear control theory.…”