Motion Planning for Highly Automated Road Vehicles with a Hybrid Approach Using Nonlinear Optimization and Artificial Neural Networks

Abstract

Over the last decade, many different algorithms were developed for the motion planning of road vehicles due to the increasing interest in the automation of road transportation. To be able to ensure dynamical feasibility of the planned trajectories, nonholonomic dynamics of wheeled vehicles must be considered. Nonlinear optimization based trajectory planners are proven to satisfy this need, however this happens at the expense of increasing computational effort, which jeopardizes the real-time applicability of these methods. This paper presents an algorithm which offers a solution to this problematic with a hybrid approach using artificial neural networks (ANNs). First, a nonlinear optimization based trajectory planner is presented which ensures the dynamical feasibility with the model-based prediction of the vehicle's motion. Next, an artificial neural network is trained to reproduce the behavior of the optimization based planning algorithm with the method of supervised learning. The generation of training data happens off-line, which eliminates the concerns about the computational requirements of the optimization-based method. The trained neural network then replaces the original motion planner in on-line planning tasks which significantly reduces computational effort and thus run-time. Furthermore, the output of the network is supervised by the model based motion prediction layer of the original optimization-based algorithm and can thus always be trusted. Finally, the performance of the hybrid method is benchmarked with computer simulations in terms of dynamical feasibility and run-time and the results are investigated. Examinations show that the computation time can be significantly reduced while maintaining the feasibility of resulting vehicle motions.

Next Generation X-in-the-Loop Validation Methodology for Automated Vehicle Systems

“…1, SciL testing is about to artificially create a complete traffic scenario with all the participants and the model of the surrounding environment, which continuously provides the necessary data for the testing [91]. The generated scenario involves the motion planning and control of the tested vehicle system [92], including potentially critical situations [93] and the other interacting components of the whole transportation process (such as vulnerable road users, other vehicles, or the road traffic in general). Furthermore, the SciL model may also reflect the relevant external influencing factors like weather and lighting conditions, infrastructure characterisation, or road environment properties.…”

Optimization of the Internal Roller Burnishing Process for Energy Reduction and Surface Properties

“…The designed ANFIS model using fivelayer feed-forward neural networks for the burnishing responses is expressed as follows (Fig. 3) [20] and [21]:…”

Hierarchical Evasive Path Planning Using Reinforcement Learning and Model Predictive Control

“…The mentioned NLP solver can generate many feasible paths for different constraints and goals and use these as a training dataset of a neural network, which can generalize the problem and create trajectories in real-time. Such a solution can be found in [16]. However, it is not always possible to generate datasets that are large enough for training.…”