Cooperative adaptive cruise control (CACC) systems have the potential to increase traffic throughput by allowing smaller headway between vehicles and moving vehicles safely in a platoon at a harmonized speed. CACC systems have been attracting significant attention from both academia and industry since connectivity between vehicles will become mandatory for new vehicles in the USA in the near future. In this paper, we review three basic and important aspects of CACC systems: communications, driver characteristics, and controls to identify the most challenging issues for their real-world deployment. Different routing protocols that support the data communication requirements between vehicles in the CACC platoon are reviewed. Promising and suitable protocols are identified. Driver characteristics related issues, such as how to keep drivers engaged in driving tasks during CACC operations, are discussed. To achieve mass acceptance, the control design needs to depict real-world traffic variability such as communication effects, driver behavior, and traffic composition. Thus, this paper also discusses the issues that existing CACC control modules face when considering close to ideal driving conditions.

A Review of Communication, Driver Characteristics, and Controls Aspects of Cooperative Adaptive Cruise Control (CACC)

This study proposes a framework for human-like autonomous car-following planning based on deep reinforcement learning (deep RL). Historical driving data are fed into a simulation environment where an RL agent learns from trial and error interactions based on a reward function that signals how much the agent deviates from the empirical data. Through these interactions, an optimal policy, or car-following model that maps in a human-like way from speed, relative speed between a lead and following vehicle, and inter-vehicle spacing to acceleration of a following vehicle is finally obtained. The model can be continuously updated when more data are fed in. Two thousand car-following periods extracted from the 2015 Shanghai Naturalistic Driving Study were used to train the model and compare its performance with that of traditional and recent data-driven car-following models. As shown by this study’s results, a deep deterministic policy gradient car-following model that uses disparity between simulated and observed speed as the reward function and considers a reaction delay of 1 s, denoted as DDPGvRT, can reproduce human-like car-following behavior with higher accuracy than traditional and recent data-driven car-following models. Specifically, the DDPGvRT model has a spacing validation error of 18% and speed validation error of 5%, which are less than those of other models, including the intelligent driver model, models based on locally weighted regression, and conventional neural network-based models. Moreover, the DDPGvRT demonstrates good capability of generalization to various driving situations and can adapt to different drivers by continuously learning. This study demonstrates that reinforcement learning methodology can offer insight into driver behavior and can contribute to the development of human-like autonomous driving algorithms and traffic-flow models.

Human-like autonomous car-following model with deep reinforcement learning

In machine learning, if the training data is an unbiased sample of an underlying distribution, then the learned classification function will make accurate predictions for new samples. However, if the training data is not an unbiased sample, then there will be differences between how the training data is distributed and how the test data is distributed. Standard classifiers cannot cope with changes in data distributions between training and test phases, and will not perform well. Domain adaptation and transfer learning are sub-fields within machine learning that are concerned with accounting for these types of changes. Here, we present an introduction to these fields, guided by the question: when and how can a classifier generalize from a source to a target domain? We will start with a brief introduction into risk minimization, and how transfer learning and domain adaptation expand upon this framework. Following that, we discuss three special cases of data set shift, namely prior, covariate and concept shift. For more complex domain shifts, there are a wide variety of approaches. These are categorized into: importance-weighting, subspace mapping, domain-invariant spaces, feature augmentation, minimax estimators and robust algorithms. A number of points will arise, which we will discuss in the last section. We conclude with the remark that many open questions will have to be addressed before transfer learners and domain-adaptive classifiers become practical.

An introduction to domain adaptation and transfer learning.

Vehicle driving safety is influenced by many factors, including drivers, vehicles, and road environments. The interactions among them are quite complex. Consequently, existing methods that evaluate driving safety perform inadequately because they only consider limited factors and their interactions. As such, it is difficult for kinematics-based and dynamics-based vehicle driving safety assistant systems to adapt to increasingly complex traffic environments. In this paper, we propose a new concept, i.e., the driving safety field . The concept makes use of field theory to represent risk factors owing to drivers, vehicles, road conditions, and other traffic factors. A unified model of the driving safety field is constructed, which includes the following three parts: 1) a potential field, which is determined by nonmoving objects on the roads, such as a stopped vehicle; 2) a kinetic field, which is determined by the moving objects on roads, such as vehicles and pedestrians; and 3) a behavior field, which is determined by the individual characteristics of drivers. Moreover, the applications of the model are proposed, and its application to a typical car-following scenario is illustrated, which evaluates the risks caused by multiple traffic factors. The driving safety field can reveal driver–vehicle–road interactions and their influences on driving safety, as well as predict driving safety trends owing to dynamic changes. In addition, the model can provide a new foundation for establishing driving safety measures and active vehicle control under complex traffic environments.

The Driving Safety Field Based on Driver–Vehicle–Road Interactions

In this paper the various driving style analysis solutions are investigated. An in-depth investigation is performed to identify the relevant machine learning and artificial intelligence algorithms utilised in current driver behaviour and driving style analysis systems. This review therefore serves as a trove of information, and will inform the specialist and the student regarding the current state of the art in driver style analysis systems, the application of these systems and the underlying artificial intelligence algorithms applied to these applications. The aim of the investigation is to evaluate the possibilities for unique driver identification utilizing the approaches identified in other driver behaviour studies. It was found that Fuzzy Logic inference systems, Hidden Markov Models and Support Vector Machines consist of promising capabilities to address unique driver identification algorithms if model complexity can be reduced.

/pdf/a-review-of-intelligent-driving-style-analysis-systems-and-34nd0lpvlu.pdf

A Review of Intelligent Driving Style Analysis Systems and Related Artificial Intelligence Algorithms.

A prototype of a longitudinal driving-assistance system, which is adaptive to driver behavior, is developed. Its functions include adaptive cruise control and forward collision warning/avoidance. The research data came from driver car-following tests in real traffic environments. Based on the data analysis, a driver model imitating the driver's operation is established to generate the desired throttle depression and braking pressure. Algorithms for collision warning and automatic braking activation are designed based on the driver's pedal deflection timing during approach (gap closing). A self-learning algorithm for driver characteristics is proposed based on the recursive least-square method with a forgetting factor. Using this algorithm, the parameters of the driver model can be identified from the data in the manual operation phase, and the identification result is applied during the automatic control phase in real time. A test bed with an electronic throttle and an electrohydraulic brake actuator is developed for system validation. The experimental results show that the self-learning algorithm is effective and that the system can, to some extent, adapt to individual characteristics.

An Adaptive Longitudinal Driving Assistance System Based on Driver Characteristics

The paper presents a curving adaptive cruise control (ACC) system that is coordinated with a direct yaw-moment control (DYC) system and gives consideration to both longitudinal car-following capability and lateral stability on curved roads. A model including vehicle longitudinal and lateral dynamics is built first, which is as discrete as the predictive model of the system controller. Then, a cost function is determined to reflect the contradictions between vehicle longitudinal and lateral dynamics. Meanwhile, some I/O constraints are formulated with a driver permissible longitudinal car-following range and the road adhesion condition. After that, desired longitudinal acceleration and desired yaw moment are obtained by a linear matrix inequality based robust constrained state feedback method. Finally, driver-in-the-loop tests on a driving simulator are conducted and the results show that the developed control system provides significant benefits in weakening the impact of DYC on ACC longitudinal car-following capability while also improving lateral stability.

A curving ACC system with coordination control of longitudinal car-following and lateral stability

The invention relates to an early warning method for preventing rear-end collision with a rearward vehicle and an early warning system thereof, wherein the method comprises the following steps: 1) arranging an early warning system for preventing rear-end collision with the rearward vehicle, which comprises an information acquisition unit, a system control unit comprising a singlechip as well as an early warning execution unit; wherein, an early warning module for preventing rear-end collision with the rearward vehicle is prearranged inside the singlechip, a collision warning range and a formula calculating the collision time of the vehicles are prearranged inside the early warning module for preventing rear-end collision with the rearward vehicle; 2) acquiring the distance between the rearward vehicle and own vehicle and relative speed information and acquiring own speed information; 3) calculating time of collision TTC and a warning time threshold of collision TTC_W of the rearward vehicle and own vehicle; 4) generating a control instruction in accordance with the difference value between the calculated time threshold of collision TTC and the warning time threshold of collision TTC_W of the rearward vehicle and own vehicle as well as the prearranged collision warning range; 5) sending corresponding warning information according to the control instruction The invention can real-timely and accurately monitor the relative motion state between own vehicle and the rear vehicle that travels in the same traffic lane, thus effectually preventing the rear-end collision of the vehicles that traveling at high speed

Pre-alarm method for preventing rear-end collision by following car

The invention relates to an automobile lane changing early-warning method and a lane changing early-warning system. The method comprises the steps of (1) arranging the lane changing early-warning system which comprises an information collecting unit, a system control unit and an early-warning execution unit, wherein a critical safety distance computing module and an early-warning decision module are arranged in the system control unit, and a computing formula for critical safety distance which is right suitable for lane changing and an early-warning decision condition are contained; (2) collecting information of distance between the closest target automobiles in a primary lane where a user automobile is located and a lane-changing target lane in a given front-back range and the user automobile and relative speed information and speed information of the user automobile; (3) computing the critical safety distance suitable for lane changing and comparing actual distance and computed critical safety distance and generating an early-warning control instruction; and (4) sending corresponding alarming information according to the early-warning control instruction. By means of the automobile lane changing early-warning method and the lane changing early-warning system, surrounding driving environment of the user automobile can be detected comprehensively, lane-changing safety under different working conditions can be judged, and collision accidents in the lane changing process can be effectively prevented.

Automobile lane changing early-warning method and lane changing early-warning system

The invention relates to an automobile tyre burst safe control method and a system thereof. The method comprises the steps as follows: 1) an automobile tyre burst safe control system is established and comprises a signal collecting unit, a main control unit, a lengthways safe executing unit and a transverse safe executing unit, and all units are communicated with each other by a vehicle CAN bus; 2) when an automobile tyre bursts, the main control unit sends a control instruction to the lengthways safe executing unit and the transverse safe executing unit; 3) after receiving the control instruction, the lengthways safe executing unit closes a throttle valve by a throttle valve automatic control apparatus and drives a lengthways braking executing device to lead the automobile to reduce the speed fast; and 4) simultaneously, after receiving the control instruction, the transverse safe executing unit controls a steering auxiliary motor to lead the automobile to keep the former running direction. The invention has less difficulty to be realized, high reliability and low cost and has important practical significance and wide application prospect.

Dezhao Zhang

Papers

An Adaptive Longitudinal Driving Assistance System Based on Driver Characteristics

A curving ACC system with coordination control of longitudinal car-following and lateral stability

Pre-alarm method for preventing rear-end collision by following car

Automobile lane changing early-warning method and lane changing early-warning system

Vehicle tyre-bursting security control method and system