A fast and accurate hybrid simulation model for the large-scale testing of automated driving functions

Abstract

The upcoming market introduction of highly automated driving functions and associated requirements on reliability and safety require new tools for the virtual test coverage to lower development expenses. In this contribution, a computationally efficient and accurate simulation environment for the vehicle’s lateral dynamics is introduced. Therefore, an analytic single track model is coupled with a long-short-term-memory neural network to compensate modelling inaccuracies of the single track model. This ‘Hybrid Vehicle Model’ is parameterized with selected training batches obtained from a complex simulation model serving as a reference to simplify the data acquisition. The single track model is parameterized using given catalogue data. Thereafter, the long-short-term-memory network is trained to cover for the single track model’s shortcomings compared to the ground truth in a closed-loop setup. The evaluation with measurements from the real vehicle shows that the hybrid model can provide accurate long-term predictions with low computational effort that outperform results achieved when using the models isolated.

Efficient Simulation Based Calibration of Automated Driving Functions Based on Sensitivity Based Optimization

“…To further rank these solutions, a robustness analysis with respect to vehicle model parameters is applied. The vehicle model used for the simulation based optimization is a so called 'hybrid vehicle model' that consists of a single track model and a neural network (details see [44]). Since a variation of parameters related to the neural network would require a new training every time, the robustness analysis is performed based on single track model parameters.…”

Lateral Trajectory Tracking of Self-Driving Vehicles Based on Sliding Mode and Fractional-Order Proportional-Integral-Derivative Control

“…With increasing attention to the research of autonomous driving control technology, certain results have been achieved for research into the lateral control of self-driving vehicles. Fraikin et al proposed a hybrid vehicle model that combines a vehicle monorail model with a long short-term memory neural network, which not only reduces model computation time but also enables more accurate long-term prediction of vehicle lateral dynamics [9]. Guo et al proposed a double envelope path tracking method using a model prediction algorithm with variable sampling time and variable prediction step, which can effectively deal with the modeling error and improve the path tracking accuracy [10].…”

Hybrid physics and neural network model for lateral vehicle dynamic state prediction

“…Pracny et al 25 coupled neural networks and spline function to study the influence of oil temperature change on the operating characteristics of the shock absorber. Fraikin et al 26 established an efficient and accurate vehicle lateral dynamics simulation by coupling the long short-term memory (LSTM) neural network with the single-track model. Graeber et al 27 combined neural networks with a vehicle kinematics model in the side-slip angle estimation and increased the number of input features of a neural network by using the kinematics model to improve the estimation quality of the side-slip angle.…”