Abstract
In order to minimize the operational difficulty of parallel parking, we proposed a novel curve based parking path and following control method, taking into account the requirements of vehicle kinematic and collision constraints in the parking process and computational real-time. Firstly, based on the combination of arctangent method to derive the boundary of feasible parking starting area for parallel parking, a parking path curve with continuous curvature and satisfying the path constraints is generated. Secondly, a constraint function for safe vehicle parking was established with the objective of vehicle collision and eventual parallel parking. Then, according to the theory of fractional order calculus (FOC), a sliding mode controller (SMC) is constructed and use radial basis function (RBF) networks to offset unknown interference to complete parking path following control. Upon this basis, a gradient descent algorithm based on FOC theory is introduced to adaptively adjust the parameter updating rules of RBF neurons in order to increase the convergence rate of network, to improve the dynamic response characteristics and robustness of the parking path following control. In the end, the effectiveness of the proposed parking path planning algorithm is evaluated using simulation software. The results indicate that the proposed parking path planning algorithm based on new curve can obtain continuous curvature, and satisfy the parking avoidance constraints and vehicle parking requirements, and the vehicle can strictly follow the target path and finally complete the parking task.
Get full access to this article
View all access options for this article.
