Intelligent Driver Model Research Articles

Truck Platooning Merging Control Method in Merge Areas Based on the IDM Model and V2X Communication

Truck platooning technology has been widely studied for its advantages in traffic efficiency and energy savings. Although recent studies have made significant progress in specific scenarios, research on merging ramp trucks into the main road truck platoons remains insufficient. This paper proposes an innovative cooperative control method for merging ramp trucks and main road truck platoons based on vehicle-to-everything (V2X) technology. The method integrates the intelligent driver model (IDM), acceleration control logic for merging trucks before merging, lane-changing decision logic, steering control methods, and on-board unit (OBU) and roadside unit (RSU) communication technologies. This ensures that autonomous merging trucks can smoothly join the main road truck platoon under various insertion positions and acceleration lane lengths. After successfully joining the platoon, the merging trucks can maintain appropriate spacing and high speed, thereby improving overall traffic efficiency. Further analysis on different insertion positions and acceleration lane lengths reveals that when using the proposed method, merging trucks have the least impact on the overall platoon when they join at the head or tail of the main road truck platoon. Although the proposed method performs better with longer acceleration lanes, these lanes are costly to construct. The proposed method can also enable successful platoon joining with shorter acceleration lanes, optimizing the use of acceleration lanes to some extent.

Mechanism and modeling of car-following behavior under local multi-vehicle influence

Aiming at the limitations of single-dimensional car-following model to describe vehicle car-following behavior in local multi-vehicle environment, the mechanism of vehicles in adjacent lane influencing the subject vehicle car-following behavior is explored, and a vehicle car-following model more suitable for local multi-vehicle environment is attempted to be established. The driving behavior variable reflecting the influence of vehicles in adjacent lane was determined by correlation analysis. Vortisch indicator of similarity(VIS) was used to characterize the influence of vehicles in adjacent lane on car-following behavior. Chi-square independent test and kernel density curve were used to determine the VIS demarcation threshold reflecting whether the influence is significant. Recursive feature elimination was used to screen the variables related to car-following samples significantly affected by vehicles in adjacent lane. The influence mechanism of variables was determined according to the statistical analysis results. Based on the mechanism proposed, the car-following model suitable for local multi-vehicle environment was constructed and its prediction effect was evaluated. Results show that VIS between the speed of subject vehicle and following vehicle in adjacent lane can characterize the influence and the VIS threshold is 0.668. It is concluded that the attention mechanism and memory effect can explain the influence mechanism of car-following behavior in multi-vehicle environment. Considering the attention mechanism and memory effect, RMSE decrease by 65%in Full Velocity Difference Model (FVD) model and 62%in Intelligent Driver Model (IDM) model, MAE decrease by 65%and 59%and R2 increased by 180%and 288%respectively, which proved the rationality of the attention mechanism and memory effect in explaining the car following behavior of subject vehicle in local multi-vehicle environment.

A Microscopic On-Ramp Model Based on Macroscopic Network Flows

While macroscopic traffic flow models adopt a fluid dynamic description of traffic, microscopic traffic flow models describe the dynamics of individual vehicles. Capturing macroscopic traffic phenomena accurately remains a challenge for microscopic models, especially in complex road sections. Based on a macroscopic network flow model calibrated to real traffic data and new rules for the acceleration and merging behavior on the on-ramp, we propose a microscopic model for on-ramps. To evaluate the performance of the new flow-based model, we conduct traffic simulations assessing speeds, accelerations, lane change positions, and risky behavior. Our results show that, although the proposed model exhibits some limitations, its performance is superior to the Intelligent Driver Model in the evaluated aspects. While the Intelligent Driver Model simulations are almost free of conflicts, the proposed model evokes a realistic amount and severity of conflicts and therefore can be considered for safety analysis.

Multi-objective optimization for connected and automated truck platoon control with improved CACC model

Connected and Automated Truck Platoon (CATP) refers to a group of trucks traveling closely together with minimal spacing to improve fuel economy and safety. However, challenges arise from instability due to internal platoon factors and external traffic disturbances. This research presents an improved Cooperative Adaptive Cruise Control (CACC) model tailored for CATP to address these challenges. The model is designed to enhance safety, fuel efficiency, and traffic efficacy. The improvements of the proposed model are in two aspects: the optimizing of the time headway strategy and the dynamic parameter adjustments of controller based on multi-objectives. The Dynamic Safety Requirement Time Headway (DSRTH) strategy facilitates the timely detection of the accelerations of the leading vehicles within the platoon, enabling quick driving responses. Additionally, Model Predictive Control (MPC) enables dynamic calibration of Proportional-Derivative (PD) control parameters and issuance of velocity commands. Meanwhile, the integration of a second-order time-delay response model has been implemented to adapt to dynamic changes in commands. A transfer function has been established, and stability has been proven. To evaluate the model performance, simulation analysis was performed using real vehicle trajectories as the CATP following vehicles. The results indicate that the DSRTH strategy outperforms both the Constant Time Headway (CTH) and Variable Time Headway (VTH) strategies, allowing rear vehicles to reach the speed trough earlier, with response speeds improved by 3.1 % and 1.5 %, respectively. Compared to the Intelligent Driver Model (IDM) and CACC models, the improved CACC model achieves a steady state of constant acceleration sooner, with recovery times reduced by 17.7 % and 3.2 %. Additionally, compared to the IDM model, the improved CACC model can save 3.23 % in fuel consumption. Furthermore, sensitivity analysis indicates that as the CATP proportion and platoon size increase, there is a positive impact on traffic flow. However, when the platoon size exceeds 5 vehicles, it shows a negative impact on the stability of other vehicles in the traffic flow besides those in the CATP.

MetaFollower: Adaptable personalized autonomous car following

Car-following (CF) modeling, a fundamental component in microscopic traffic simulation, has attracted increasing research interest in recent decades. In this study, we propose an adaptable personalized car-following framework —– MetaFollower, by leveraging the power of meta-learning. Specifically, we first utilize Model-Agnostic Meta-Learning (MAML) to extract common driving knowledge from various CF events. Afterward, the pre-trained model can be fine-tuned on new drivers with only a few CF trajectories to achieve personalized CF adaptation. We additionally combine Long Short-Term Memory (LSTM) and Intelligent Driver Model (IDM) to reflect temporal heterogeneity with high interpretability. MetaFollower can accurately capture and simulate the intricate dynamics of car-following behavior while considering the unique driving styles of individual drivers. We demonstrate the versatility and adaptability of MetaFollower by showcasing its ability to adapt to new drivers with limited training data quickly. To evaluate the performance of MetaFollower, we conduct rigorous experiments comparing it with both data-driven and physics-based models. The results reveal that our proposed framework outperforms baseline models in predicting car-following behavior with higher accuracy and safety. To the best of our knowledge, this is the first car-following model aiming to achieve fast adaptation by considering both driver and temporal heterogeneity based on meta-learning.

A trajectory optimization method for CAVs at intersections considering energy-efficiency balance

With the development of vehicle-to-everything (V2X) technology, connected automated vehicles (CAVs) will be an important part of urban road transportation. However, frequent acceleration, deceleration, and rapid stopping within intersections lead to increased energy consumption, reduced driving efficiency and balanced benefits. The study proposes an optimal control method for vehicle trajectories that considers energy-efficiency balance (OCM-EEB) based on the characteristics of V2X technology. The method first develops a vehicle trajectory control model based on the Intelligent Driver Model (IDM) by accounting for the features of V2X technology, and further, introduces the idea of balanced optimization, considering energy-efficiency balance, a vehicle trajectory optimal control model (OCM) is proposed. The proposed model was solved by design algorithms based on the non-dominated sorting genetic algorithm-II (NSGA-II). The experimental results showed that the vehicle trajectory optimization control method could increase the balanced driving efficiency of the vehicle within intersection, which is conducive to improving the robustness of the vehicle’s trajectory state, and then prove the effectiveness of the method proposed in this study. The method also provides an important technical basis for the vehicle trajectory optimization experiments under a real V2X environment in the future.

Influence of an Automated Vehicle with Predictive Longitudinal Control on Mixed Urban Traffic Using SUMO

In this paper, an approach to quantify the area of influence of an intelligent longitudinally controlled autonomous vehicle in an urban, mixed-traffic environment is proposed. The intelligent vehicle is executed with a predictive longitudinal control, which anticipates the future traffic scenario in order to reduce unnecessary acceleration. The shown investigations are conducted within a simulated traffic environment of the city center of Darmstadt, Germany, which is carried out in the traffic simulation software “Simulation of Urban Mobility” (SUMO). The longitudinal dynamics of the not automated vehicles are considered with the Extended Intelligent Driver Model, which is an approach to simulate real human driver behavior. The results show that, in addition to the energy saving caused by a predictive longitudinal control of the ego vehicle, this system can also reduce the consumption of surrounding traffic participants significantly. The area of influence can be quantified to four vehicles and up to 250 m behind.

Cooperative energy optimal control involving optimization of longitudinal motion, powertrain, and air conditioning systems

This paper is concerned with the optimal control strategies for the longitudinal control, powertrain, and air conditioning (A/C) system of connected four-wheel hub-drive electric vehicles (EVs). A hierarchical control framework is developed to enhance the energy economy of the vehicle. Real-time connected information is utilized in the upper layer to determine the travel mode. Then, a multi-objective motion planning system (MOMPS) is proposed to plan the optimal acceleration trajectory. In the lower layer, an offline global optimization approach is employed to find the torque combinations that minimize the total power loss. The proposed A/C controller operates based on the bi-level model predictive control (Bi-level MPC) algorithm. A novel prediction model is developed to enable the A/C system to decrease energy consumption by leveraging the speed of the vehicle. The performance of the MOMPS is evaluated using urban test road data, demonstrating that the MOMPS can balance multiple objectives compared to global dynamic programing (Global DP) and the intelligent driver model (IDM). In addition, the proposed torque distribution strategy results in a 4.98% energy-savings rate through comparison with the even torque distribution strategy. Moreover, the A/C controller proposed in this paper can optimize energy consumption by 13.57% compared to a baseline strategy that maintains a constant setting.

Car-following model based on artificial potential field with consideration of horizontal curvature in connected vehicles environment

Connected vehicles (CVs) will gradually replace traditional vehicles to become the main components of traffic flow. Studying the car-following behavior characteristics is crucial for improving traffic flow stability and safety in CVs environment. Additionally, the radius of road curvature significantly impacts vehicle driving behavior, making it necessary to consider it for the car-following models of CVs. The artificial potential field (APF) theory can more accurately and comprehensively depict various microscopic driving behaviors, offering a new approach for modeling vehicle microscopic behavior. Firstly, this paper constructs the attractive and repulsive potential fields considering horizontal curve curvature based on a road coordinate transformation model. Secondly, an Artificial Potential Field-Based Car-Following Model Considering Curvature (APFCCM) in connected vehicles environment is proposed. Finally, the model is calibrated and validated using the Hangzhou - Xifu Freeway dataset from the Tongji Road Trajectory Sharing (TJRD TS) platform, and compared with the full velocity difference model(FVDM), the Intelligent Driver Model (IDM) and the Driving Risk Potential Field Model (DRPFM). The results show that the APFCCM performs well in trajectory simulation, model accuracy, and scenario adaptability, and it has the lowest mean absolute error(MAE) and root mean square error(RMSE) in position, speed, and acceleration metrics.

An improved hierarchical deep reinforcement learning algorithm for multi-intelligent vehicle lane change

Lane change is a pivotal technology in autonomous driving, playing a significant role in improving traffic conditions. The control task of lane change becomes exceptionally complex for coordination of multiple intelligent vehicles, especially in unpredictable situations. To address this issue, the paper proposes a novel multi-intelligent vehicle lane change method (MVLC) based on Deep Reinforcement Learning (DRL). The MVLC features a hierarchical framework that includes a lower-layer lane change controller and an upper-layer reinforcement learning decision-making method. In the lower layer, a lane change controller is designed to execute motion control for a single vehicle, consisting of a lateral controller utilizing dueling double deep Q-network to make lane change decision and a longitudinal controller that employs Intelligent Driver Model (IDM) to follow the preceding vehicle. In the upper layer, a model-free DRL method is used for simultaneous control of multiple intelligent vehicles. To learn the optimal policy, a comprehensive reward function is designed. Finally, experiments are conducted in mixed traffic to valid the effectiveness of the proposed method. Results show that MVLC achieves secure and timely lane changes for multiple vehicles. Therefore, the method is expected to have significant potential for improving overall traffic efficiency and mitigating traffic congestion.

An Improved Longitudinal Driving Car-Following System Considering the Safe Time Domain Strategy.

Car-following models are crucial in adaptive cruise control systems, making them essential for developing intelligent transportation systems. This study investigates the characteristics of high-speed traffic flow by analyzing the relationship between headway distance and dynamic desired distance. Building upon the optimal velocity model theory, this paper proposes a novel traffic car-following computing system in the time domain by incorporating an absolutely safe time headway strategy and a relatively safe time headway strategy to adapt to the dynamic changes in high-speed traffic flow. The interpretable physical law of motion is used to compute and analyze the car-following behavior of the vehicle. Three different types of car-following behaviors are modeled, and the calculation relationship is optimized to reduce the number of parameters required in the model's adjustment. Furthermore, we improved the calculation of dynamic expected distance in the Intelligent Driver Model (IDM) to better suit actual road traffic conditions. The improved model was then calibrated through simulations that replicated changes in traffic flow. The calibration results demonstrate significant advantages of our new model in improving average traffic flow speed and vehicle speed stability. Compared to the classic car-following model IDM, our proposed model increases road capacity by 8.9%. These findings highlight its potential for widespread application within future intelligent transportation systems. This study optimizes the theoretical framework of car-following models and provides robust technical support for enhancing efficiency within high-speed transportation systems.

Bayesian Calibration of the Intelligent Driver Model

Bayesian Calibration of the Intelligent Driver Model

A novel car-following model for adaptive cruise control vehicles using enhanced intelligent driver model

ABSTRACT This paper proposes Enhanced Intelligent Driver Model for Adaptive Cruise Control (EIDM-ACC) vehicles, a novel car-following model that dynamically adjusts desired speed and considers acceleration inertia. The EIDM-ACC model is compared with two widely used models for simulating ACC vehicles – the ACC model developed by the PATH Project (PATH-ACC) at the University of California Transportation Institute and Continuous Asymmetric Optimal Velocity Relative Velocity (CAOVRV) model. Three models are calibrated and cross-validated using real vehicle trajectory data from the OpenACC dataset. Results show that the EIDM-ACC outperforms the other two models in small and large fluctuation stages. In addition, EIDM-ACC has better performance in capturing the instability and energy consumption of ACC vehicles, and also has advantages over the other two models in terms of safety.

Application of dynamic desired headway based adaptive backstepping sliding mode control design to mixed traffic system

To optimize the mixed car-following behavior of connected autonomous vehicles (CAVs) and human-driven vehicles (HDVs), an adaptive backstepping sliding mode control (ABSMC) strategy for longitudinal velocity and distance control model (namely, MCF-ABSMCM) is put forth. First, the variable desired headway model (VDHM) is designed based on the vehicle-to-vehicle and vehicle-to-infrastructure information. Also, an asymmetric stochastic car-following model (ASCM) is designed to realistically describe the stochastic factor as well as vehicle acceleration and braking behaviors in mixed car-following. Second, an adaptive cruise sliding mode control law with a backstepping approach is devised to precisely track the desired following distance based on the VDHM. After creating the basic diagram of the two-model traffic flow, the intelligent driver model (IDM) is used as a baseline to compare and analyze the traffic capacity. Lastly, the effectiveness of the MCF-ABSMCM approach is confirmed in terms of both the mixed traffic flow density wave evolution and the vehicle control strategy, and the MCF-ABSMCM is validated under various adversary inputs. The simulation results demonstrate that the mixed-flow queue can finish headway tracking and keep a safe distance when using the MCF-ABSMCM described. This control strategy has been validated in VISSIM to significantly improve the efficiency of traffic operations compared to IDM and ASCM.

Theory-data dual driven car following model in traffic flow mixed of AVs and HDVs

To model the car following (CF) behavior of mixed traffic flow composed of autonomous vehicles (AVs) and human-driving vehicles (HDVs) in the future, this paper calibrated the lower control algorithm of AVs and proposed a model named theory-data dual driven stochastically generative adversarial networks (TDS-GAN) to describe the CF behavior of HDVs based on experimental data from mixed traffic flow. Firstly, the experimental scenario and collected data were introduced. Secondly, two transfer functions for the lower control algorithm of AVs were compared. Then, to account for the stochasticity of HDVs, the idea of physics-informed deep learning (PIDL) was used to improve generative adversarial networks (GAN) and integrate it with two-dimensional intelligent driver model (2D-IDM). Finally, the effectiveness of the model was verified from both a micro prediction and a macro simulation perspective. The influence of AVs on the stability of mixed traffic flow at different Market Penetration Rate (MPR) was also observed. The results show that TDS-GAN can effectively describe the following behavior and stochasticity of HDVs. When combined with CF model of AVs, it can depict the evolution of mixed traffic flow more accurately. Additionally, AVs can improve traffic flow stability under different penetration rates, and the effect is more significant with higher MPR. However, the introduction of AVs may not necessarily be positive in terms of road capacity.

Identifying drivers’ perception-reaction time (PRT) in car-following processes via two different methods using vehicle trajectory data

To judge whether the connected and automated vehicles (CAVs) could perform better than human, this study proposes two methods to identify perception-reaction time (PRT) of human drivers using vehicle trajectory data - the calibrated-based and duration-based methods. The calibrated-based method utilises an intelligent driver model (IDM) to capture the longitudinal driving behaviours and delineate the PRT as a model parameter. While the duration-based approach collects PRT by manually extracting the PRT value based on the trajectory data. To investigate the factors influencing these types of PRTs, two regression models are proposed. Results reveal important contributing factors affecting the PRT values of human drivers, including the mean and standard deviation of the leading vehicle’s speed, and the speed and gap difference between leading and following vehicles, as well as the mean speed of the following vehicle. Besides, the same factor has different effects on the calibrated and selected PRTs. For instance, the std of the gap between leading and following vehicles manifests negative and positive impacts on the calibrated and selected PRTs, respectively. In summary, findings of this study can provide valuable insights into the identification of the PRT values of human drivers, potentially improving future design of CAVs.

The calculation and distribution of CAV carbon emissions on urban transportation systems: A comparative analysis of renewable and non-renewable energy sources

When powered by electricity, Connected and Autonomous Vehicles (CAVs), an emerging mode of transportation, possess the capacity to reduce exhaust emissions greatly. However, accurately measuring carbon emissions in urban transportation remains a challenge, especially considering emissions from electricity generation and gasoline consumption. This paper proposes an innovative method for calculating CAVs' carbon emissions distribution, utilizing both renewable and non-renewable energy. The study employs SUMO, an agent-based simulation platform, to develop an intelligent driver model and cooperative adaptive cruise control modules, tracking vehicle movement behavior across various vehicle types, including Gasoline Vehicles (GVs), Electric Vehicles (EVs), Human-Driven Vehicles (HDVs), and CAVs. Subsequently, a lifecycle electric carbon emission model is constructed, integrating the energy consumption model of EVs with carbon emission factors of renewable and non-renewable energy. Visualization models are then developed to clarify the carbon emission distribution within the traffic network. A case study conducted in Suzhou, China validates the model, analyzing the spatiotemporal distribution of carbon emissions. Results show EVs can reduce carbon emissions by 70%–90 % compared to GVs on urban roads during rush hour, while CAVs can further reduce emissions by 35%–50 % compared to HDVs. Additionally, carbon emissions from non-renewable energy sources were found to exceed those from renewable sources.

Simulation-Oriented Analysis and Modeling of Distracted Driving

Distracted driving significantly affects the efficiency and safety of traffic flow. Modeling distracted driving behavior in microscopic traffic flow simulation is essential for understanding its critical impacts on traffic flow. However, due to the influence of various external factors and the considerable uncertainties in behavior characteristics, modeling distracted driving behavior remains a challenge. This study proposed a model which incorporates distraction features into the microscopic traffic flow model to simulate distracted driving behavior. Specifically, the study first examines the characteristics of distracted driving, including the intervals and durations of distraction events, as well as the patterns and environments of distraction. It then introduces distraction parameters into the Intelligent Driver Model (IDM), including reaction time delays and perception deviations in both speed difference and following distance. These parameters are quantified by probabilistic distributions to reflect the uncertainty and individual differences in driving behavior. The model is calibrated and validated using 772 distracted following events from the Shanghai Naturalistic Driving Study (SH-NDS) data. Three patterns of distraction (excessive, moderate, mild) are distinguished and modeled separately. The results show that the model’s accuracy surpasses that of the IDM under various road types and traffic volumes, with an average improvement in model accuracy of about 11.30% on expressways with high traffic volume, 4.54% on expressways with low traffic volume, and 4.46% on surface roads. Meanwhile, the model can effectively simulate the variations in reaction times and perceptual deviations in both speed and following distance for different distraction modes at the individual level, maintaining consistency with reality. Finally, the study simulates distracted driving behavior under different road environments and traffic volumes to explore the impact of distracted driving on traffic flow. The simulation results indicate that an increase in the proportion of distraction reduces the efficiency and safety of traffic flow, which is consistent with real-world observations. Since the model considers human distraction factors, it can generate more dangerous driving scenarios in simulations, which holds significant importance for safety-related research. The findings from this study are expected to be helpful for understanding distracted driving behavior and mitigate its negative influence on the efficiency and safety of traffic flow.

Towards Robust Decision-Making for Autonomous Highway Driving Based on Safe Reinforcement Learning.

Reinforcement Learning (RL) methods are regarded as effective for designing autonomous driving policies. However, even when RL policies are trained to convergence, ensuring their robust safety remains a challenge, particularly in long-tail data. Therefore, decision-making based on RL must adequately consider potential variations in data distribution. This paper presents a framework for highway autonomous driving decisions that prioritizes both safety and robustness. Utilizing the proposed Replay Buffer Constrained Policy Optimization (RECPO) method, this framework updates RL strategies to maximize rewards while ensuring that the policies always remain within safety constraints. We incorporate importance sampling techniques to collect and store data in a Replay buffer during agent operation, allowing the reutilization of data from old policies for training new policy models, thus mitigating potential catastrophic forgetting. Additionally, we transform the highway autonomous driving decision problem into a Constrained Markov Decision Process (CMDP) and apply our proposed RECPO for training, optimizing highway driving policies. Finally, we deploy our method in the CARLA simulation environment and compare its performance in typical highway scenarios against traditional CPO, current advanced strategies based on Deep Deterministic Policy Gradient (DDPG), and IDM + MOBIL (Intelligent Driver Model and the model for minimizing overall braking induced by lane changes). The results show that our framework significantly enhances model convergence speed, safety, and decision-making stability, achieving a zero-collision rate in highway autonomous driving.

A new car-following model with incorporation of Markkula's framework of sensorimotor control in sustained motion tasks

This study develops a car-following model called the “Markkula Intelligent Driver Model (MIDM)” based on Markkula's Framework of Sensorimotor Control. The MIDM predicts the start time of driver's reaction based on the evidence accumulation within Markkula's framework unlike the existing car-following models that use a constant reaction time parameter. The MIDM also accurately represents the actual shape and duration of acceleration maneuvers. Fifty drivers’ car-following behavior was observed in 2 different scenarios using a driving simulator – reaction to a decelerating lead vehicle and reaction to a stopped lead vehicle. Trajectory data from the NGSIM project were also used for the evaluation. Compared to the Gipps Model, the Wiedemann Model and the IDM, the MIDM realistically reproduced trajectories of speed, acceleration, jerk and spacing for both simulator and NGSIM data. The MIDM can also incorporate the effects of lead vehicle brake lights for more accurate estimation of the driver reaction time.