Merging on highways poses challenges due to limited visibility, fluctuating speeds, and possibly erratic human behavior. Since human drivers and autonomous vehicles might cohabit in mixed traffic, researchers investigated ways to alleviate the risk in these situations. They designed connected autonomous vehicles that embed empathy into their decision-making, enabling them to maneuver altruistically through coordination to assist merging vehicles. However, the percentage of connected autonomous vehicles will be too small in the early phases of mixed traffic to carry out such coordination and yield the expected safety impact. Therefore, we propose Archicool, an autonomous vehicle model performing altruistic maneuvers without any communication device. Based on multi-agent reinforcement learning and social value orientation, Archicool improves safety for merging vehicles while minimizing disturbances to traffic flow. In medium-traffic density, selective empathy reduces the risk perceived by merging vehicles by 16%.

This paper presents the hardware implementation of a hybrid control framework that combines fuzzy logic and fuzzy Q-learning for intelligent energy management in battery electric vehicles (BEVs). The proposed architecture integrates two fuzzy controllers: one dedicated to traction motor power regulation and another designed for HVAC energy optimization. Initially modeled and validated in MATLAB, the controllers were subsequently ported to C and CUDA to enable real-time deployment on embedded platforms, specifically Jetson Nano and Raspberry Pi. To improve adaptability, a fuzzy Q-learning layer was embedded into the motor controller, allowing dynamic adjustment in challenging scenarios such as zero-speed transitions, abrupt accelerations, and varying load conditions. Experimental validation was carried out on three distinct real-world driving datasets, covering diverse operational profiles. Each system version (MATLAB, C on PC, CUDA on Jetson, and ARM implementation on Raspberry Pi) was benchmarked in terms of accuracy and computational efficiency. The results demonstrate that the hybrid fuzzy–Q learning controller consistently achieved high precision (MAE < 0.3) across all datasets, while maintaining low execution times compatible with real-time automotive requirements. These findings confirm both the effectiveness and the portability of the proposed system across heterogeneous hardware architectures, highlighting its suitability for embedded energy management applications in BEVs.

The global emphasis on sustainable transportation is driving the increasing adoption of Battery Electric Vehicles (BEVs), which offer independence from fossil fuels and zero emissions during operation. However, optimizing energy efficiency and vehicle performance in BEVs remains a significant challenge due to the dynamic nature of driving conditions. Current power control methods often struggle to adapt to these varying conditions, leading to suboptimal energy distribution and reduced performance. This paper presents a novel approach to power control in BEVs using a Fuzzy Q-learning Controller (FQLC), which dynamically adjusts the motor power coefficient based on real-time driving conditions. The FQLC optimizes energy distribution to the electric motor by adapting to factors such as vehicle speed, road slope, and battery state of charge (SOC). A comparative analysis between the Fuzzy Logic Controller (FLC) and the proposed FQLC demonstrates the advantages of the new system. The Modified Mean Absolute Error (MMAE) is used to quantitatively evaluate performance across various driving scenarios. The results show that the FQLC significantly outperforms the FLC, achieving MMAE values as low as 0.01, indicating substantial reductions in error rates. In the performed tests, the FQLC’s ability to manage energy use contributed to range extensions in certain cases, achieving an increase of up to 11 km. These findings highlight the FQLC potential as an innovative solution for BEV power control.

Energy Management Systems (EMS) are pivotal in optimizing the performance and efficiency of battery electric vehicles (BEVs) by intelligently distributing power within the vehicle's components. This work aims to design an EMS approach utilizing a dynamic coefficient controlled by a Fuzzy Logic Controller (FLC) to manage the energy distribution of the propulsion systems based on driving conditions and vehicle demands. The FLC operates on six critical input variables-vehicle velocity, state of charge (SOC), road slope, ambient temperature, vehicle weight, and driving mode-to adaptively modulate the energy supplied to the electric motor. Validation of this EMS through real-world data demonstrates the FLC's capability to significantly enhance electric vehicle efficiency and responsiveness across various operational conditions.

The dynamics of a two-wheeled vehicle represent a higher level of complexity, with a high variety of rider practices, peculiarly during the bend-taking maneuver. This complexity is one reason why research on motorcyclists’ practiced trajectories is limited compared to cars. Another reason is the low representation of motorcycles in the land motor vehicle fleet.The dataset was collected within the framework of the French ANR research project VIROLO++. The objective was twofold. One concerns the analysis of the interactions between a rider and the motorcycle during a bend-taking maneuver. The second aims at studying the precise trajectory of a motorcycle.Multi-sensor instrumentation of a motorcycle (Honda CBF1000) has been achieved for precise trajectory reconstruction as well as the analysis of the interactions between the rider and the motorcycle during a ride. Real experiments were conducted on the “La Ferté Gaucher'' racetrack. The collected data include the precise location and attitude of the motorcycle and rider. Several sensors providing redundant data have been embedded to validate measurements. Two RTK GPS receivers were installed at the rear of the motorcycle to provide ground truth. A preliminary analysis was performed to validate the consistency of the dataset and the relevance of the sensors according to the aimed study.

Electric vehicles (EVs) are rising in the automotive industry, replacing combustion engines and increasing their global market presence. These vehicles offer zero emissions during operation and more straightforward maintenance. However, for such systems that rely heavily on battery capacity, precisely determining the battery’s state of charge (SOC) presents a significant challenge due to its essential role in lithium-ion batteries. This research introduces a dual-phase testing approach, considering factors like HVAC use and road topography, and evaluating machine learning models such as linear regression, support vector regression, random forest regression, and neural networks using datasets from real-world driving conditions in European (Germany) and African (Morocco) contexts. The results validate that the proposed neural networks model does not overfit when evaluated using the dual-phase test method compared to previous studies. The neural networks consistently show high predictive precision across different scenarios within the datasets, outperforming other models by achieving the lowest mean squared error (MSE) and the highest R2 values.

Human errors are the primary cause of powered two-wheeler crashes worldwide due to the demanding control required and the often ineffective rider-training programs. Literature on rider behaviour is limited, partly due to the lack of standard investigation methodologies. This work investigated the differences in riding style and capability of a diverse set of riders. It explored the impact of familiarisation and riding instruction through objective metrics. Correlation with experience was a particular focus. Seven riders of various experience levels performed trials on an instrumented motorcycle, following three riding instructions: ‘Free Riding’, ‘Handlebar Riding’, and ‘Body Riding’. Objective metrics assessed rider familiarisation, capability and willingness to excite motorcycle dynamics, riding style, and input preference. Results indicated that riders asymptotically converged to their motorcycle dynamics intensity level after a specific distance; both intensity and distance were positively correlated with experience. Experienced riders achieved higher longitudinal acceleration and utilised combined dynamics to a higher degree. The negative longitudinal jerk during braking varied greatly among riders and correlated with experience. A clustering approach identified two prominent trial groups concerning the motorcycle response intensity. Higher diversity emerged in the inputs, leading to five clusters with distinct riding style meanings. Instructions influenced behaviour, particularly regarding input usage. The unsupervised approach and metrics proposed should make rider behaviour research more straightforward and objective. It could be applied to naturalistic riding sessions for more conclusive evidence of inter-driver differences. The diversity that emerged concerning the command inputs used warrants a revision of training practices to promote riding safety.

Powered-Two-Wheelers (PTW) riders’ fatalities are prevalent on bends outside built-up areas due to the complexity and instability of their vehicles: countermeasures require a better understanding of the rider-PTW interaction. Analysing riding data is effective but becomes challenging when using extensive datasets; segmenting the riding data would help identify events of interest, isolate specific manoeuvres and describe the riding session. Manual segmentation would be time-consuming and subjective; automation would be beneficial. This work proposed an automatic, unsupervised tool for segmenting and clustering signals acquired during a riding session for studying motorcycle lateral dynamics in-depth. The method only requires measuring the motorcycle roll angle. An expert rider completed a closed route using an instrumented motorcycle; the algorithm divided the time series into segments categorised into clusters relative to specific riding conditions. Analysing the segmented trial revealed the effectiveness and usefulness of the approach. Then, a corner entry manoeuvre was investigated in-depth to observe each segment’s properties. The method associated each riding primitive to a cluster and described each manoeuvre through the segments’ succession. The clusters were unambiguous and easy to interpret thanks to their dynamics-based nature and minimal overlap. The algorithm identified the differences between the three corner entry manoeuvres in the trial. The segmentation simplified the in-depth corner entry analysis and allowed early detection of the manoeuvre start. The proposed tool can aid research on motorcycle dynamics, PTW-rider interaction, and riding preferences in bends. The segmented time series could be employed for rider training and pre-crash fall dynamics reconstruction.

Tremendous Research work is devoted to advancing state of art in intelligent transportation systems, with the ultimate objective of maximizing the beneficial effect of the transportation sector on the environment. One of the most critical parts of balancing energy consumption and resource exploitation in sustainable development is the introduction of decarbonized transportation in the form of electric cars. This article discusses the results of a study that evaluates the energy needed to power a battery electric vehicle (BEV) when the driver switches to the comfort ride option. We have modeled a Battery electric vehicle, thermal management system, and Adaptive Cruise (ACC) blocks using MATLAB/Simulink. To evaluate our developed model, we have conducted simulation tests under a variety of settings, including route and temperature data from Moroccan roads.

Understanding driver-vehicle interactions remains a challenge, particularly in the case of cornering. This is particularly the case for powered two-wheeler vehicle (PTWs) users, perhaps because PTW drivers play a greater role in controlling the stability of their vehicles than do four-wheeled vehicle drivers. This difficulty stems from the variety of practices of this population of road users when entering, controlling their path, and exiting turns. Thus, observing the evolution of rider behavior during a cornering maneuver is an essential step in identifying road environment features that are risk factors for this category of road users. The real data set used for the experiments reported here was collected in the framework of the VIROLO++ collaborative project to improve the knowledge of the real practices of PTWs drivers, particularly during cornering. The in-depth analysis of these data in order to better understand motorcyclists’ behavior can therefore be considered a challenge. For this purpose, a two-step methodology was applied: (1) a data segmentation and feature extraction step in which the multidimensional time series of roll angle and roll velocity data were segmented using a multiple regression hidden logistic process (MRHLP), and (2) a clustering step in which the two detected segments characterizing the curve entry were assigned to different clusters regarding the curve initiation and the control actions set up during the different phases of the turning maneuver based on the hierarchical clustering algorithm. The results obtained show the effectiveness of the proposed methodology.

In the near future, autonomous vehicles (AVs) may cohabit with human drivers in mixed traffic. This cohabitation raises serious challenges, both in terms of traffic flow and individual mobility, as well as from the road safety point of view. Mixed traffic may fail to fulfill expected security requirements due to the heterogeneity and unpredictability of human drivers, and autonomous cars could then monopolize the traffic. Using multi-agent reinforcement learning (MARL) algorithms, researchers have attempted to design autonomous vehicles for both scenarios, and this paper investigates their recent advances. We focus on articles tackling decision-making problems and identify four paradigms. While some authors address mixed traffic problems with or without social-desirable AVs, others tackle the case of fully-autonomous traffic. While the latter case is essentially a communication problem, most authors addressing the mixed traffic admit some limitations. The current human driver models found in the literature are too simplistic since they do not cover the heterogeneity of the drivers’ behaviors. As a result, they fail to generalize over the wide range of possible behaviors. For each paper investigated, we analyze how the authors formulated the MARL problem in terms of observation, action, and rewards to match the paradigm they apply.

Human-in-the-loop driving simulation aims to create the illusion of driving by stimulating the driver’s sensory systems in as realistic conditions as possible. However, driving simulators can only produce a subset of the sensory stimuli that would be available in a real driving situation, depending on the degree of refinement of their design. This subset must be carefully chosen because it is crucial for human acceptability. Our focus is the design of a physical dynamic (i.e., motion-based) motorcycle-riding simulator. For its instrumentation, we focused on the rider acceptability of all sub-systems and the simulator as a whole. The significance of our work lies in this particular approach; the acceptability of the riding illusion for the rider is critical for the validity of any results acquired using a simulator. In this article, we detail the design of the hardware/software architecture of our simulator under this constraint; sensors, actuators, and dataflows allow us to (1) capture the rider’s actions in real-time; (2) render the motorcycle’s behavior to the rider; and (3) measure and study rider/simulated motorcycle interactions. We believe our methodology could be adopted by future designers of motorcycle-riding simulators and other human-in-the-loop simulators to improve their rendering (including motion) quality and acceptability.

Driving a motorcycle relies on the feedback provided by several human sensory systems, on the one hand, and anticipation of the consequences of control actions, on the other hand.Driving simulators aim to create the illusion of driving by stimulating the driver's sensory systems.However, a significant number of drivers experience simulator sickness, which hinders the usefulness of driving simulators in their applications, such as driving behavior research or training / re-training.Simulator sickness occurrence is often attributed to sensory conflict.In this work, we propose an approach to understanding simulator sickness by considering the need for coherence between the complexity of the vehicle model and the complexity of the simulator from a hardware point-of-view, which constrains the fidelity of the reproduced sensory stimuli.We then describe the design of a proof-of-concept system that considers the particular issue of haptic feedback for the handlebars of a motorcycle-riding simulator.We will use this system in further experiments to demonstrate the impact of the coherence or mismatch of those two aspects on controllability and simulator sickness occurrence.

Many motorcycle accidents occur at intersections and are caused by other vehicle drivers who misperceive the speed and time-to-arrival of an approaching motorcycle. The two experiments reported here tested different motorcycle headlight configurations likely to counteract this perceptual failure. In the first experiment, conducted on a driving simulator, car drivers turned left in front of cars and motorcycles approaching an intersection under nighttime lighting conditions. The motorcycles were equipped with either a standard white central light, or one of three vertical configurations of white and yellow lights. The results showed that the standard configuration led to significantly more unsafe accepted gaps than the vertical configurations. In the second experiment, conducted on a test track using a similar task, the most promising motorcycle headlight configuration, i.e., the vertical yellow-white light arrangement (one central white light, plus one yellow light on the helmet and two yellow lights on the fork) was evaluated and compared to a standard configuration and a car. The vertical yellow-white headlight configuration again provided significant safety benefits as compared to the standard configuration. These findings demonstrate that motorcycle safety can be improved by headlight ergonomics that accentuate the vertical dimension of motorcycles. They also suggest that the driving simulator is a valid tool for conducting research on motorcycle headlight design.

Simulator sickness, an adverse physiological reaction to a simulated driving situation, is often attributed exclusively to sensory conflict or sometimes to postural instability. We postulate that simulator sickness occurrence is a negative effect of poor- or non-controllability of the virtual vehicle that induces badly controlled ego-motion in the virtual scene and uneasiness. We believe that this non-controllability stems from a mismatch between the complexity of the virtual vehicle model and that of the simulator’s hardware architecture. The architecture limits the quality of the sensory stimuli that can be provided to the user, which is problematic because of the driver’s expectations based on their prior real-life driving experience. We designed and conducted a simple within-participants experiment using a small proof-of-concept system to explore our hypothesis. The experiment consisted of the stabilization of a virtual pendulum’s oscillations using a haptic-feedback actuator. Twenty-four participants faced situations where (1) the visual feedback and the dynamic behavior of the simulated pendulum were coherent, and (2) they were mismatched. Our results show a significant effect of training on motor control and task performance; mismatch between visual feedback and dynamic model on motor control, task performance, and participant’s discomfort. We interpreted these results as supporting our hypothesis.

Introduction: Motorcyclists are particularly at risk of being injured when involved in a road traffic accident. To avoid such crashes, emergency braking and/or swerving maneuvers are frequently performed. The recent development of dynamic motorcycle simulators may allow to study the influences of various disturbance factors such as sleep deprivation (SD) and time-of-day (TOD) in safe conditions. Methods: Twelve young healthy males took part in 8 tests sessions at 06:00 h, 10:00 h, 14:00 h, 18:00 h after a night with or without sleep, in a random order. Participants had to perform an emergency braking and a swerving maneuver, both realized at 20 and 40 kph on a motorcycle dynamic simulator. For each task, the total distance/time necessary to perform the maneuver was recorded. Additional analysis was conducted on reaction and execution distance/time (considered as explanatory variables). Results: Both crash avoidance maneuvers (emergency braking and swerving) were affected by increased speed, resulting in longer time and distance at 40 kph than at 20 kph. Emergency braking was mainly influenced by sleep deprivation, which significantly increased the total distance necessary to stop at 40 kph (+1.57 m; + 20%; p < 0.01). These impaired performances can be linked to an increase in reaction time (+21%; p < 0.01). Considering the swerving maneuver, TOD and SD influences remained limited. TOD only influenced the reaction time/distance measured at 40 kph with poorer performance in the early morning (+30% at 06:00 h vs 18:00 h; p < 0.05). Discussion: Our results confirm that crash avoidance capabilities of young motorcyclists were influenced by the lack of sleep, mainly because of increased reaction times. More complex tasks (swerving maneuver) remained mostly unchanged in this paradigm. Practical Applications: Prevention campaigns should focus on the dangers of motorcycling while sleepy. Motorcycling simulators can be used to sensitize safely with sleep deprivation and time-of-day influences. (c) 2021 National Safety Council and Elsevier Ltd. All rights reserved.

Analysis of motorcyclists' behaviour and the risk that this mode of transport incurs for the mobility as well as their safety have gained more attention in recent years. In the context of the European project SimuSafe, various experimentations have been done using instrumented motorbikes in a naturalistic riding study and the riders' behaviours are subjected to self-confrontation interviews with traffic psychologists. This paper aims at the identification of different riding patterns using machine learning techniques and allows for a deeper understanding of pattern-specific risk exposure from multi-source data (video footage and interviews). More specifically, we focus on the roundabout pattern analysis as it is the most important source of collisions and a set of rules are designed using a decision tree to analyse their related risks. The generated rules may also fuel a multi-agent simulator to reflect the riders' real-world behaviours.