scholarly journals Prototyping Smart Assistance with Bayesian Autonomous Driver Models

2011 ◽  
pp. 460-512 ◽  
Author(s):  
Claus Moebus ◽  
Mark Eilers
Keyword(s):  
Human Behavior ◽  
Autonomous Driving ◽  
Training Data ◽  
Driver Model ◽  
Critical Question ◽  
Human Driver ◽  
Markov Decision ◽  
Traffic Scenario ◽  
Assistance Systems ◽  
Game Simulation

The Human or Cognitive Centered Design (HCD) of intelligent transport systems requires digital Models of Human Behavior and Cognition (MHBC) enabling Ambient Intelligence e.g. in a smart car. Currently MBHC are developed and used as driver models in traffic scenario simulations, in proving safety assertions and in supporting risk-based design. Furthermore, it is tempting to prototype assistance systems (AS) on the basis of a human driver model cloning an expert driver. To that end we propose the Bayesian estimation of MHBCs from human behavior traces generated in new kind of learning experiments: Bayesian model learning under driver control. The models learnt are called Bayesian Autonomous Driver (BAD) models. For the purpose of smart assistance in simulated or real world scenarios the obtained BAD models can be used as Bayesian Assistance Systems (BAS). The critical question is, whether the driving competence of the BAD model is the same as the driving competence of the human driver when generating the training data for the BAD model. We believe that our approach is superior to the proposal to model the strategic and tactical skills of an AS with a Markov Decision Process (MDP). The usage of the BAD model or BAS as a prototype for a smart Partial Autonomous Driving Assistant System (PADAS) is demonstrated within a racing game simulation.

A motivational driver model for the design of a rear-end crash avoidance system

2019 ◽  
Vol 234 (1) ◽  
pp. 10-26
Author(s):  
Hamed Mozaffari ◽  
Ali Nahvi
Keyword(s):  
Driving Simulator ◽  
Human Subjects ◽  
Autonomous Driving ◽  
Driver Model ◽  
Driver Assistance Systems ◽  
Optimal Point ◽  
Time To Collision ◽  
Human Aspects ◽  
Crash Avoidance ◽  
Assistance Systems

A motivational driver model is developed to design a rear-end crash avoidance system. Current driver assistance systems use engineering methods without considering psychological human aspects, which leads to false activation of assistance systems and complicated control algorithms. The presented driver model estimates driver’s psychological motivations using the combined longitudinal and lateral time to collision, the vehicle kinematics, and the vehicle dynamics. These motivations simplify both autonomous driving algorithms and human-machine interactions. The optimal point of a motivational multi-objective cost function defines the decision for the autonomous driving. Moreover, the motivations are used as risk assessment factors for driver–machine interaction in dangerous situations. The system is evaluated on 10 human subjects in a driving simulator. The assistance system had no false activation during the tests. It avoided collisions in all the rear-end crash avoidance scenarios, while 90% of human subjects did not.

Cloning Safe Driving Behavior for Self-Driving Cars using Convolutional Neural Networks

2019 ◽  
Vol 12 (2) ◽  
pp. 120-127 ◽  
Author(s):  
Wael Farag
Keyword(s):  
Gradient Descent ◽  
Autonomous Driving ◽  
Driving Behavior ◽  
Training Data ◽  
Stochastic Gradient Descent ◽  
Data Set ◽  
Safe Driving ◽  
Processing Pipeline ◽  
Self Driving Cars ◽  
And Training

Background: In this paper, a Convolutional Neural Network (CNN) to learn safe driving behavior and smooth steering manoeuvring, is proposed as an empowerment of autonomous driving technologies. The training data is collected from a front-facing camera and the steering commands issued by an experienced driver driving in traffic as well as urban roads. Methods: This data is then used to train the proposed CNN to facilitate what it is called “Behavioral Cloning”. The proposed Behavior Cloning CNN is named as “BCNet”, and its deep seventeen-layer architecture has been selected after extensive trials. The BCNet got trained using Adam’s optimization algorithm as a variant of the Stochastic Gradient Descent (SGD) technique. Results: The paper goes through the development and training process in details and shows the image processing pipeline harnessed in the development. Conclusion: The proposed approach proved successful in cloning the driving behavior embedded in the training data set after extensive simulations.

scholarly journals CarLink: A Real-Time V2X Validation Platform for a Safe and Sustainable Mobility

2021 ◽  
Author(s):  
Babak Mafakheri ◽  
Pierpaolo gonnella ◽  
Barbara Masini ◽  
Alessandro Bazzi
Keyword(s):  
Short Range ◽  
Field Tests ◽  
Hardware In The Loop ◽  
Driver Assistance ◽  
Sustainable Mobility ◽  
Driver Assistance Systems ◽  
Laboratory Environment ◽  
Traffic Scenario ◽  
Assistance Systems ◽  
External Management

<div>In this article, we present CarLink, a new simulation platform with hardware-in-the-loop (HiL), designed and implemented to reduce the time spent on field tests through the emulation of a complex vehicular scenario in a controlled laboratory environment. Specifically, CarLink can simulate a generic traffic scenario and let each vehicle in it communicate with a vehicle under test (VUT), which is actually physically available HiL and equipped with long- and short-range wireless communication capabilities. Communication between simulated vehicles and the VUT is provided by an external management unit (EMU) that integrates the virtual word with the physical one. The architecture is also designed to allow the integration of advanced driver assistance systems (ADAS) testing for the validation of future connected and automated vehicles.</div>

A collection of easily deployable adversarial traffic sign stickers

2021 ◽  
Vol 69 (6) ◽  
pp. 511-523
Author(s):  
Henrietta Lengyel ◽  
Viktor Remeli ◽  
Zsolt Szalay
Keyword(s):  
Traffic Safety ◽  
Autonomous Driving ◽  
Driver Assistance ◽  
Driver Assistance Systems ◽  
Traffic Sign ◽  
Practical Possibility ◽  
Special Knowledge ◽  
Environment Perception ◽  
Assistance Systems ◽  
Traffic Sign Classification

Abstract The emergence of new autonomous driving systems and functions – in particular, systems that base their decisions on the output of machine learning subsystems responsible for environment perception – brings a significant change in the risks to the safety and security of transportation. These kinds of Advanced Driver Assistance Systems are vulnerable to new types of malicious attacks, and their properties are often not well understood. This paper demonstrates the theoretical and practical possibility of deliberate physical adversarial attacks against deep learning perception systems in general, with a focus on safety-critical driver assistance applications such as traffic sign classification in particular. Our newly developed traffic sign stickers are different from other similar methods insofar that they require no special knowledge or precision in their creation and deployment, thus they present a realistic and severe threat to traffic safety and security. In this paper we preemptively point out the dangers and easily exploitable weaknesses that current and future systems are bound to face.

Fuse -Replacement Noise-Filter Development for Automotive Audio–Video Navigation Systems Based on Loudness Isotropy Curves

2020 ◽  
Vol 25 (3) ◽  
pp. 83-92
Author(s):  
Bong-Seo Park ◽  
Hyun-cheol Park ◽  
Jung-jun Her
Keyword(s):  
Autonomous Driving ◽  
Navigation Systems ◽  
Driver Assistance ◽  
Driver Assistance Systems ◽  
Audio Quality ◽  
Noise Filter ◽  
Supply Noise ◽  
Low Pass ◽  
Assistance Systems ◽  
Audio Video

With the development of advanced driver assistance systems, the more reliable the autonomous driving technology is, the more the rest and entertainment times of the driver of the car increases. Hence, the importance of the entertainment function of automotive audio-video navigation (AVN) systems is increasing. Currently, the AVN system of automobiles has a monitoring function for fault diagnosis and a combination of functions. Applying these technologies is challenging for drivers who want to tune the audio quality to their musical taste. In this study, a method for upgrading the sound quality using a power supply noise filter without deforming the AVN system was developed. The low-pass attenuation that appeared as a side effect was solved by applying a filter using the loudness isotropic curve. In the installation method of the filter, the method of using a fuse holder minimized the inconvenience of AVN detachment and wiring. Based on the results obtained in this study, further research and improvement of the filter are required for audio tuning of various models.

scholarly journals Model predictive control–based steering control algorithm for steering efficiency of a human driver in all-terrain cranes

2019 ◽  
Vol 11 (6) ◽  
pp. 168781401985978
Author(s):  
Ja-Ho Seo ◽  
Kwang-Seok Oh ◽  
Hong-Jun Noh
Keyword(s):  
Model Predictive Control ◽  
Predictive Control ◽  
Control Algorithm ◽  
Steering System ◽  
Driver Model ◽  
Control Technique ◽  
Change Scenario ◽  
Steering Control ◽  
Human Driver ◽  
Steering Effort

All-terrain cranes with multi-axles have large inertia and long distances between the axles that lead to a slower dynamic response than normal vehicles. This has a significant effect on the dynamic behavior and steering performance of the crane. Therefore, the purpose of this study is to develop an optimal steering control algorithm with a reduced driver steering effort for an all-terrain crane and to evaluate the performance of the algorithm. For this, a model predictive control technique was applied to an all-terrain crane, and a steering control algorithm for the crane was proposed that could reduce the driver’s steering effort. The steering performances of the existing steering system and the steering system applied with the newly developed algorithm were compared using MATLAB/Simulink and ADAMS with a human driver model for reasonable performance evaluation. The simulation was performed with both a double lane change scenario and a curved-path scenario that are expected to happen in road-steering mode.

scholarly journals Development of Deep Learning Based Human-Centered Threat Assessment for Application to Automated Driving Vehicle

2019 ◽  
Vol 10 (1) ◽  
pp. 253 ◽  
Author(s):  
Donghoon Shin ◽  
Hyun-geun Kim ◽  
Kang-moon Park ◽  
Kyongsu Yi
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
Threat Assessment ◽  
Collision Probability ◽  
Driver Model ◽  
Sequential Data ◽  
Automated Driving ◽  
Human Driver ◽  
Acceleration And Deceleration ◽  
The Individual

This paper describes the development of deep learning based human-centered threat assessment for application to automated driving vehicle. To achieve naturalistic driver model that would feel natural while safe to a human driver, manual driving characteristics are investigated through real-world driving test data. A probabilistic threat assessment with predicted collision time and collision probability is conducted to evaluate driving situations. On the basis of collision risk analysis, two kinds of deep learning have been implemented to reflect human driving characteristics for automated driving. A deep neural network (DNN) and recurrent neural network (RNN) are designed by neural architecture search (NAS), and by learning from the sequential data, respectively. The NAS is used to automatically design the individual driver’s neural network for efficient and effortless design process while ensuring training performance. Sequential trends in the host vehicle’s state can be incorporated through hand-made RNN. It has been shown from human-centered risk assessment simulations that two successfully designed deep learning driver models can provide conservative and progressive driving behavior similar to a manual human driver in both acceleration and deceleration situations by preventing collision.

Sign in / Sign up

Export Citation Format

Share Document