Velocity Planning Method Base on Fuzzy Neural Network for Autonomous Vehicle

Abstract

In order to improve the comfort performance and reduce the planning algorithm complexity in autonomous vehicle, an intelligent longitudinal velocity planning method based on fuzzy neural network (FNN) is proposed. With the manual driving experience, fuzzy planning model is established. By utilizing the self-learning function of neural network, fuzzy planning model is modified, which is attempted to establish FNN planning model. The planning method is applied to velocity planning. Three kinds of driving scenes are analyzed, and velocity planning models based on FNN are established accordingly. The simulation and experiment results indicate that acceleration generated by FNN planning model has good smooth property, and it is easy to be tracked by the subsequent control module. Compared with traditional method, the proposed method has certain anti-disturbance ability and self-adaptability. Also, the proposed method is convenient for engineering application, which ensures both the real-time performance and stability of the algorithm.

Implications of As-built Highway Horizontal Curves on Vehicle Dynamic/Kinematic Characteristics Under Adaptive Cruise Control

“…Note that τde of V-ACC is calibrated within the range of 1.2-2.2 s because many ACC system configurations (e.g., exact τde value) were not elaborated in the literature. In addition, a constant τde on curves is adopted in the PID-based ACC algorithm to save computing resources and avoid determining the complicated topological relationship between the clearance and road curvature while ensuring the safety and comfort requirements, which is used by many commercial ACC systems [29,30].…”

A Comprehensive Overview of Classical and Modern Route Planning Algorithms for Self-Driving Mobile Robots

“…Finally, the least total energy path is found. However, this route is not the quickest nor most efficient [34]. The mobile robot's working environment is unpredictable and difficult to model.…”

“…[1], [2], [4], [5], [10], [11], [12], [13], [34], [37], [58], [59], [60] V. RESEARCH METHODOLOGIES This paper presents a systematized classification of path planning algorithms that are organized by environment types, date of publications, advantages, drawbacks, and principles of path search and modelling of autonomous mobile robots. The search for relevant literature was conducted in 2 main databases, IEEE Xplore and Scopus with particular attention paid to topics such as route planning, static navigation, dynamic path planning, global and local route navigation, traditional algorithms, classical path planning approaches, modern intelligence algorithms, heuristic approaches, metaheuristic approaches, and neural networks techniques.…”