-
State Derivative Normalization for Continuous-Time Deep Neural Networks
Authors:
Jonas Weigand,
Gerben I. Beintema,
Jonas Ulmen,
Daniel Görges,
Roland Tóth,
Maarten Schoukens,
Martin Ruskowski
Abstract:
The importance of proper data normalization for deep neural networks is well known. However, in continuous-time state-space model estimation, it has been observed that improper normalization of either the hidden state or hidden state derivative of the model estimate, or even of the time interval can lead to numerical and optimization challenges with deep learning based methods. This results in a r…
▽ More
The importance of proper data normalization for deep neural networks is well known. However, in continuous-time state-space model estimation, it has been observed that improper normalization of either the hidden state or hidden state derivative of the model estimate, or even of the time interval can lead to numerical and optimization challenges with deep learning based methods. This results in a reduced model quality. In this contribution, we show that these three normalization tasks are inherently coupled. Due to the existence of this coupling, we propose a solution to all three normalization challenges by introducing a normalization constant at the state derivative level. We show that the appropriate choice of the normalization constant is related to the dynamics of the to-be-identified system and we derive multiple methods of obtaining an effective normalization constant. We compare and discuss all the normalization strategies on a benchmark problem based on experimental data from a cascaded tanks system and compare our results with other methods of the identification literature.
△ Less
Submitted 14 May, 2024; v1 submitted 5 January, 2024;
originally announced January 2024.
-
Deep Learning-Based Vehicle Speed Prediction for Ecological Adaptive Cruise Control in Urban and Highway Scenarios
Authors:
Sai Krishna Chada,
Daniel Görges,
Achim Ebert,
Roman Teutsch
Abstract:
In a typical car-following scenario, target vehicle speed fluctuations act as an external disturbance to the host vehicle and in turn affect its energy consumption. To control a host vehicle in an energy-efficient manner using model predictive control (MPC), and moreover, enhance the performance of an ecological adaptive cruise control (EACC) strategy, forecasting the future velocities of a target…
▽ More
In a typical car-following scenario, target vehicle speed fluctuations act as an external disturbance to the host vehicle and in turn affect its energy consumption. To control a host vehicle in an energy-efficient manner using model predictive control (MPC), and moreover, enhance the performance of an ecological adaptive cruise control (EACC) strategy, forecasting the future velocities of a target vehicle is essential. For this purpose, a deep recurrent neural network-based vehicle speed prediction using long-short term memory (LSTM) and gated recurrent units (GRU) is studied in this work. Besides these, the physics-based constant velocity (CV) and constant acceleration (CA) models are discussed. The sequential time series data for training (e.g. speed trajectories of the target and its preceding vehicles obtained through vehicle-to-vehicle (V2V) communication, road speed limits, traffic light current and future phases collected using vehicle-to-infrastructure (V2I) communication) is gathered from both urban and highway networks created in the microscopic traffic simulator SUMO. The proposed speed prediction models are evaluated for long-term predictions (up to 10 s) of target vehicle future velocities. Moreover, the results revealed that the LSTM-based speed predictor outperformed other models in terms of achieving better prediction accuracy on unseen test datasets, and thereby showcasing better generalization ability. Furthermore, the performance of EACC-equipped host car on the predicted velocities is evaluated, and its energy-saving benefits for different prediction horizons are presented.
△ Less
Submitted 30 November, 2022;
originally announced December 2022.
-
Cyber-Automotive Simulation and Evaluation Platform for Vehicular Value Added Services
Authors:
Raja Sattiraju,
Pratip Chakraborty,
Hans D. Schotten,
Xiaohai Lin,
Daniel Görges
Abstract:
An easily moving and safe transportation is an indicator in any country in the world of economic growth and well-being. For the past 100 years, innovation within the automotive sector has brought major technological advances leading to safer, cleaner and more affordable vehicles. But for the most time since the inception of the moving assembly line for vehicle production by Henry Ford, the changes…
▽ More
An easily moving and safe transportation is an indicator in any country in the world of economic growth and well-being. For the past 100 years, innovation within the automotive sector has brought major technological advances leading to safer, cleaner and more affordable vehicles. But for the most time since the inception of the moving assembly line for vehicle production by Henry Ford, the changes have been incremental / evolutionary. Thanks to the new possibilities due to the IT / wireless revolution, the automotive industry appears to be on the cusp of revolutionary change with potential to dramatically reshape not just the competitive landscape but also the way we interact with vehicles, and indeed the future design of our roads and cities. Apart from connected personal mobility, vehicles are also envisioned to provide Value Added Services (VAS) such as autonomous driving via Vehicle-to-Vehicle (V2V) and Vehicle-to-infrastructure (V2I), electric load balancing via Vehicle-to-Grid (V2G) solutions, communication solutions using Visible Light Communications (VLC) etc. The development and evaluation of vehicular VAS requires a modular and scalable multidisciplinary simulation platform. In this paper we propose a novel simulation platform named Cyber-Automotive Simulation \& Evaluation Platform (CASEP). The purpose of CASEP is to evaluate and visualize the gains of various vehicular VAS with special emphasis on commercial vehicle VAS. The use cases are evaluated with respect to the mission-specific performance indicators, thereby providing usable metrics for optimization. The visualization platform is being developed using the UNITY 3D engine, thereby enabling intuitive interaction as in real physics-based games
△ Less
Submitted 30 October, 2017;
originally announced October 2017.
