Search

Driving simulators have been extensively used over the last decades and technological advancements have propelled their development for cars, trucks and other vehicles with four (or more) wheels. This dissertation focuses on the use of driving simulators for two wheeled vehicles and in particular on the development and evaluation of a motorcycle riding simulator for low speed maneuvering. The reason to focus on low speed maneuvers is related to the unstable nature of motorcycles at low speeds. A dedicated riding simulator could be used to train riders to cope with vehicle instabilities and develop active safety systems that can help them to maintain the vehicle balanced and avoid falling. Existing riding simulators adopt simplified vehicle models to simulate motorcycle dynamics. In some cases, advanced non-linear models are adopted, but their validation is not always sufficiently described for the simulator application. Once the model has been integrated in the complete simulator, the results of its real-time simulation are used to provide feedback to the simulator rider through the cueing systems. Motion cueing is particularly interesting due to the peculiar vehicle dynamics of two wheelers. Different approaches are found in literature, however the applied motion cueing methods are not based on understanding of human motion perception. Finally, the riding simulator should also be validated for its usage in the specific application domain and its fidelity and behavioral validity are often neglected. In this thesis, specific aspects of development and validation of a riding simulator for low speed maneuvering are investigated.

When designing driving simulation experiments with motion cueing, it is often necessary to make choices between Motion Cueing Algorithms (MCAs) without being fully able to know how well an MCA will perform during the experiment. Choices between MCAs can therefore be greatly supported by previous measurements or predictions of motion cueing quality. This paper describes a data collection experiment on a nine degree-of-freedom motion-base simulator, in which participants are asked to continuously rate the motion cueing quality during a pre-recorded drive through an urban environment. Three benchmark MCAs are compared: a Model-Predictive Control (MPC) algorithm with infinite prediction horizon, a Classical Washout Algorithm (CWA) tuned for the use-case, and the same algorithm (CWA), but with the tilt-coordination channels turned off. By comparing ratings for the whole scenario, as well as ratings for each maneuver individually, the results show a preference of the presence of tilt-coordination, as well as a preference for the optimization-based MPC algorithm over the CWA condition. The collected data will be used directly for modeling and predicting motion cueing quality for future experiments at BMW, such that the best-suited MCA and parameter setting can be selected before experiments.

When designing driving simulation experiments with motion cueing, it is often necessary to make choices between Motion Cueing Algorithms (MCAs) without being fully able to know how well an MCA will perform during the experiment. Choices between MCAs can therefore be greatly supported by previous measurements or predictions of motion cueing quality. This paper describes a data collection experiment on a nine degree-of-freedom motion-base simulator, in which participants are asked to continuously rate the motion cueing quality during a pre-recorded drive through an urban environment. Three benchmark MCAs are compared: a Model-Predictive Control (MPC) algorithm with infinite prediction horizon, a Classical Washout Algorithm (CWA) tuned for the use-case, and the same algorithm (CWA), but with the tilt-coordination channels turned off. By comparing ratings for the whole scenario, as well as ratings for each maneuver individually, the results show a preference of the presence of tilt-coordination, as well as a preference for the optimization-based MPC algorithm over the CWA condition. The collected data will be used directly for modeling and predicting motion cueing quality for future experiments at BMW, such that the best-suited MCA and parameter setting can be selected before experiments.

When designing driving simulation experiments with motion cueing, it is often necessary to make choices between Motion Cueing Algorithms (MCAs) without being fully able to know how well an MCA will perform during the experiment. Choices between MCAs can therefore be greatly supported by previous measurements or predictions of motion cueing quality. This paper describes a data collection experiment on a nine degree-of-freedom motion-base simulator, in which participants are asked to continuously rate the motion cueing quality during a pre-recorded drive through an urban environment. Three benchmark MCAs are compared: a Model-Predictive Control (MPC) algorithm with infinite prediction horizon, a Classical Washout Algorithm (CWA) tuned for the use-case, and the same algorithm (CWA), but with the tilt-coordination channels turned off. By comparing ratings for the whole scenario, as well as ratings for each maneuver individually, the results show a preference of the presence of tilt-coordination, as well as a preference for the optimization-based MPC algorithm over the CWA condition. The collected data will be used directly for modeling and predicting motion cueing quality for future experiments at BMW, such that the best-suited MCA and parameter setting can be selected before experiments.

Flight simulators, or Flight Simulation Training Devices (FSTDs), offer great benefits in terms of safety and cost associated with pilot training and certification. To warrant uniform certification standards and to prevent adverse pilot training, (sub)system fidelity requirements are imposed by the Federal Aviation Authority (FAA) and European Aviation Safety Agency (EASA). While comprehensive, a notable example of an area in which these requirements are somewhat limited, are those pertaining to the Motion Cueing System (MCS) of full-flight flight simulators. The MCS comprises hardware, typically a set of actuators to enable physical motion of the platform, and software, often termed the Motion Cueing Algorithm (MCA), to process the simulated vehicle motion to prevent violation of (physical) simulator constraints. Naturally, the MCA introduces a significant mismatch between the actual (i.e., in-flight) and simulated vehicle motion perceived by the pilot. Furthermore, this mismatch often comes on top of inaccuracies in the mathematical model used to compute the simulated vehicle motion. Because of this complex interaction, the formulation of quantitative requirements pertaining to the allowed mismatch between real vehicle and simulator motion has proven cumbersome. To date, certification of flight simulator motion is therefore based predominantly on subjective evaluation by experienced pilots. To address this problem, the aim of this dissertation is to develop a unifying tool to quantify motion cueing fidelity in helicopter flight simulation and to evaluate its suitability in realistic applications.

Motion cueing algorithm design often involves a trade-off between priorities due to the limited workspace of the simulator. Such a trade-off requires a detailed understanding of human perception, which we do not yet have. For that reason, objective motion cueing quality metrics, based on the difference between vehicle and simulator signals, offer a fast and simple alternative. Next to motion cueing quality, we argue that the total motion cueing algorithm (MCA) quality is about more than only the quality of the motion, and can also entail implementation and operational aspects of an MCA for a specific use-case and simulator combination, i.e., it is a task-dependent issue. In this paper this idea is discussed by comparing three objective motion cueing quality metrics (absolute difference, delay and cross-correlation) from literature and two metrics regarding simulator operations (workspace management and energy consumption). Comparing such metrics is difficult, but is nevertheless useful to improve the process of simulator operations if various MCAs and/or simulators are available, to aid their selection pro- cess. As a first step towards such a method, a Virtual Test Environment (VTE) was developed as a versatile software environment to compare these metrics, as well as to visualize simulator motion and its characteristics in a 3D-animation. This aims at helping MCA designers in making choices between different MCA types, their configurations, simulators and use-cases, guiding them to select the best-suited motion cueing solution.

A necessity in driving simulation testing is to understand and attenuate simulator sickness, to reduce the number of undesired drop-outs. Especially urban environments, with its many turns and changes in the velocity profile, are a challenge. This paper describes the motion sickness rating results of a between-subjects experiment (n = 63), which investigated the effects of adding scaled yaw motion to a simulator on the sickness incidence and severity while being driven as passenger through an urban environment. Three cases were considered: no motion, scaled yaw motion, and including the vehicle pitch and roll rotations in addition to the scaled yaw motion. The misery scale (MISC) was obtained every minute, and the simulator sickness questionnaire (SSQ) was completed before and after the 45-minute trial. Experimental results show that less participants became sick when some form of yaw-motion was provided.

Motion cueing algorithm design often involves a trade-off between priorities due to the limited workspace of the simulator. Such a trade-off requires a detailed understanding of human perception, which we do not yet have. For that reason, objective motion cueing quality metrics, based on the difference between vehicle and simulator signals, offer a fast and simple alternative. Next to motion cueing quality, we argue that the total motion cueing algorithm (MCA) quality is about more than only the quality of the motion, and can also entail implementation and operational aspects of an MCA for a specific use-case and simulator combination, i.e., it is a task-dependent issue. In this paper this idea is discussed by comparing three objective motion cueing quality metrics (absolute difference, delay and cross-correlation) from literature and two metrics regarding simulator operations (workspace management and energy consumption). Comparing such metrics is difficult, but is nevertheless useful to improve the process of simulator operations if various MCAs and/or simulators are available, to aid their selection pro- cess. As a first step towards such a method, a Virtual Test Environment (VTE) was developed as a versatile software environment to compare these metrics, as well as to visualize simulator motion and its characteristics in a 3D-animation. This aims at helping MCA designers in making choices between different MCA types, their configurations, simulators and use-cases, guiding them to select the best-suited motion cueing solution.

Researchers are using driving simulators to design and assess the automated driving and driver assistant systems, due to the safe nature of experimentation in the virtual environment. The motion cues and accelerations felt by the drivers are essential for an accurate perception of the events and the response of the drivers. In this paper, the vehicle dynamic model and the Motion Cueing Algorithm used for the simulation is described in detail, then driver’s performance and subjective assessments was studied for the braking, chicane and overtaking maneuver in the 3 different gear shifting scenario. The study demonstrates that the presence of the motion cueing feedback in the driving simulation was satisfactory and gave realistic cues for the participants independent of the gear shifting system, however no significant effect was found from the driver’s behavior due to different gear shifting system.

Despite gaining popularity, the use of Motion Cueing Algorithms (MCAs) based on Model Predictive Control (MPC) remains challenging due to the required tuning of a large number of cost function parameters. This paper investigates the effects of two critical MPC cost function parameters, the lateral specific force and roll rate error weights (Way and Wp), on the motion cueing quality achieved with an MPC-based MCA for a curve driving scenario. An offline sensitivity analysis, which quantified the effects of varying Way and Wp on the Root Mean Square Error (RMSE) and Pearson Correlation Coefficient (PCC) of the resulting simulator motion outputs, shows that for the same percentage-wise variation, Way has a more pronounced effect on both cueing quality predictors than Wp. In addition, for both RMSE and PCC, the effects of Way and Wp are also found to be largely independent, i.e., without interaction effects. This was further tested in a passive human-in-the-loop experiment with 20 participants and with nine different Way and Wp parameter combinations as test conditions, performed in the hexapod moving-base simulator of the Max Planck Institute for Biological Cybernetics in T¨ubingen. The collected continuous rating data, which were found to be reliable for 18/20 participants, show a statistically significant variation across all experiment conditions, and especially a strong interaction effect of Way and Wp. Somewhat surprisingly, the overall lowest continuous ratings were given to the combination of both reference weight settings from earlier research (our baseline condition). In line with the interaction effect in the continuous data, an extended post-experiment correlation analysis shows that a weighted combination of lateral specific force RMSE and and roll rate RMSE above the roll rate perception threshold strongly correlates (_ = 0.98) with the variation in mean continuous ratings across all experiment conditions. This approach can potentially be used for straightforward prediction of perceived motion cueing quality and offline MCA optimization.

This paper describes a driving simulation experiment, executed on the Daimler Driving Simulator (DDS), in which a filter-based and an optimization-based motion cueing algorithm (MCA) were compared using a newly developed motion cueing quality rating method. The goal of the comparison was to investigate whether optimization-based MCAs have, compared to filter-based approaches, the potential to improve the quality of motion simulations. The paper describes the two algorithms, discusses their strengths and weaknesses and describes the experimental methods and results. The MCAs were compared in an experiment where 18 participants rated the perceived motion mismatch, i.e., the perceived mismatch between the motion felt in the simulator and the motion one would expect from a drive in a real car. The results show that the quality of the motion cueing was rated better for the optimization-based MCA than for the filter-based MCA, indicating that there exists a potential to improve the quality of the motion simulation with optimization-based methods. Furthermore, it was shown that the rating method provides reliable and repeatable results within and between participants, which further establishes the utility of the method.

The use of dynamic driving simulators is nowadays common practice in the automotive industry. The effectiveness of such devices is strongly related to their capabilities of well reproducing the driving sensations, hence it is crucial that the motion control strategies generate both realistic and feasible inputs to the platform. Such strategies are called Motion Cueing Algorithms (MCAs). Model Predictive Control (MPC) has been successfully applied to MCAs, being well suited to solve constrained optimal control problems. However, the predictive aspect of the algorithm has not been exploited effectively yet, mainly due to the hard real-time requirement when using a significantly long prediction window. In this paper, a real time implementation of the so called Look-Ahead (LA) strategy is presented, that is based on an effective manipulation of the reference along the prediction horizon, and on an on-line switching policy to a Non-Look-Ahead strategy when the expected driver behavior is not reliable. An optimal tuning of the MCA is computed by means of a multi-objective optimization, where both performance improvement due to the prediction exploitation, and robustness to varying driver behavior are taken into account. Finally, a characterization of the driver skill level is proposed and validated in experimental environment.

Supplying aircraft pilots with adequate motion cues in flight simulators is still an important issue. The main problem is the limited space envelope of motion systems that makes an ideal replication of accelerations impossible. As a consequence those translational acceleration signals cannot be provided directly and need to be transformed to angular attitudes or scaled down depending on the time in which they appear. State-of-the-art motion drive algorithms like the classic washout filter algorithm are using sets of high- and low-pass filters in combination with gain factors to achieve such a response. But those control methods do not distinguish between the causes of accelerations with respect to flight maneuvers. The Apparent Vertical Filter developed by the German Aerospace Center provides a different approach to this problem. The current paper gives an overview of how the problem of false motion cues could be adessed within a number of idealized maneuvers. As an example the filter input data of a coordinated turn maneuver will be defined. The constraints of current filter methods using the classic washout filter will be discussed. The paper gives an overview of how the apparent vertical filter deals with the problem and how it is structured. Finally, the responses of both filter methods, the AVF anf the CWA, to aircraft model data are compared.

The use of dynamic driving simulators is constantly increasing in the automotive community, with applications ranging from vehicle development to rehab and driver training. The effectiveness of such devices is related to their capabilities of well reproducing the driving sensations, hence it is crucial that the motion control strategies generate both realistic and feasible inputs to the platform. Such strategies are called motion cueing algorithms (MCAs). In recent years several MCAs based on model predictive control (MPC) techniques have been proposed. The main drawback associated with the use of MPC is its computational burden, that may limit their application to high performance dynamic simulators. In the paper, a fast, real-time implementation of an MPC-based MCA for 9 DOF, high performance platform is proposed. Effectiveness of the approach in managing the available working area is illustrated by presenting experimental results from an implementation on a real device with a 200 Hz control fre...

In this paper we present the implementation of a model-predictive controller (MPC) for real-time control of a cable-robot-based motion simulator. The controller computes control inputs such that a desired acceleration and angular velocity at a defined point in simulator's cabin are tracked while satisfying constraints imposed by working space and allowed cable forces of the robot. In order to fully use the simulator capabilities, we propose an approach that includes the motion platform actuation in the MPC model. The tracking performance and computation time of the algorithm are investigated in computer simulations. Furthermore, for motion simulation scenarios where the reference trajectories are not known beforehand, we derive an estimate on how much motion simulation fidelity can maximally be improved by any reference prediction scheme compared to the case when no prediction scheme is applied. (C) 2017, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.

Vehicle simulators aim to provide the user with a driving experience that replicates the sensations experienced in a real vehicle. An important aspect is the replication of motion cues. The replication of motion cues in the vehicle simulator is limited by platform workspace and the selected position control systems performance. In this work a Matlab/Simulink setup that incorporates the various aspects involved in replication of motion sensations was created. This includes input data scaling, classical washout algorithm and inverse kinematic implementations. By adjusting the input data scaling and classical washout algorithm filter parameters platform motion that adheres to the actuator motion constraints was assured. Therefore this simulation system makes the design process easier since no parameter adjusting was done on the actual motion platform. This simulation system also ensures safe operation of the vehicle simulator system for the user and ensures safety of the vehicle simulator mechanical structure. Additionally the system allowed for the derived inverse kinematic equations to be verified and evaluated the effectiveness of the classical washout motion cueing algorithm.

When tuning motion platforms, subjective opinion is usually regarded as sufficient for most applications, as it should provide verification that no false cueing occurs. However, often systems are not configured to fully utilize the available motion for the desired purpose. This paper presents a new method to objectively tune simulation motion platforms. Enhanced usage of the available platform motion space is shown through an example of its application to a specific mission task. The difference in motion filter settings for two simulator platform geometries is shown. This displays the suitability of the platforms for completion of the two-axis lateral reposition task.

The evaluation of motion cueing quality is a difficult task. Subjective ratings can give a hint on preferable parameter settings or algorithm design choices. Though, inter-individual rating differences are huge and evaluation results are often contradicting. Hence, the aviation community tried and succeeded in defining an objective motion cueing test during the past decade. Even though the interpretation of the results is still a topic of discussion and continuous research, the test serves as a useful tool for the design process of motion cueing algorithms. This paper presents an approach to adapt this motion cueing test for driving simulators. Evaluation results of two different motion systems and motion cueing designs are presented.

This paper investigates the importance of yaw and sway motion cues in curve driving simulation. While such motion cues are known to enhance simulation realism, their function in supporting realistic driver behavior in simulators is still largely unknown. A human-in-the-loop curve driving experiment was performed in the SIMONA Research Simulator at TU Delft, in which eight participants were asked to follow a winding road’s center-line, while being subject to wind disturbances. Four motion conditions were tested: 1) no motion, 2) yaw only, 3) sway only, and 4) both yaw and sway; each was tested with 5 m and 100 m road preview for correspondence with earlier work. Results show that visual road preview is essential for adequate road-following. Although the effects of yaw and sway cues are much smaller, sway motion feedback allows for improved disturbance-rejection performance, while yaw motion feedback results in reduced control activity. These distinctly different effects suggest that both motion cues are important for evoking realistic driving behavior in simulators.

Full motion simulators are traditionally used in the flight industry to train pilots. They are used to add the sensation of acceleration in simulation to make it more "realistic". Clearly the motion envelop of the simulator is limited by physical constraints so the motion platform has to be stopped and returned to the center position after an acceleration cue, called washout. A key question is: which acceleration can a subject feel and which not, called the acceleration threshold. We are also interested in strength of accelerations for which a subject can detect the direction. Literature gives several results, but we found that some of these values seemed very low to us and the experiments were conducted on very specific groups of people like pilots, A.J.Benson and H.Vogel (1986), Schroeder (1999). Furthermore, we are simulating moving vehicles like a car or an air plane and are interested in the acceleration ranges in a noisy environment. Noisy, the noise is a result from the vibration of engines, rough roads and disturbances that are Gaussian. This thesis gives a literature review, implement the cueing procedure to make motion and vibration to do different experiment and analyze the results. Thesis Master of Computer Science (MCS)

We investigated the performance of 12 airline pilots in recovering from an asymmetrical stall in a flight simulator featuring an extended aerodynamic model of a transport-category aircraft, and a centrifuge-based motion platform capable of generating enhanced buffet motion and g-cueing. All pilots had difficulties in applying the correct recovery template, as their initial response was to counter the wing drop rather than to unload the aircraft. Reproduction of actual g-loads led to significantly longer recoveries, more secondary stick-shaker events, lower maximum load factors

La version editeur de cet article est disponible à l'adresse suivante : http://onlinelibrary.wiley.com/doi/10.1002/oca.2081/abstract; International audience; This study proposes a method and an experimental validation to analyze dynamics response of the simulator's cabin and platform with respect to the type of the control used in the hexapod driving simulator. In this article, two different forms of motion platform tracking control are performed as a classical motion cueing algorithm and a discrete-time linear quadratic regulator (LQR) motion cueing algorithm. For each situation, vehicle dynamics and motion platform level data are registered from the driving simulation software. In addition, the natural frequencies of the roll accelerations are obtained in real-time by using FFT. The data are denoised by using wavelet 1D transformation. The results show that by using discrete-time LQR algorithm, the roll acceleration amplitudes that correspond to the natural frequencies and the total roll jerk have decreased at the motion platform level. Also, the natural frequencies have increased reasonably by using the discrete LQR motion cueing (1.5-2.2 Hz) compared with using the classical algorithm (0.4-1.5 Hz) at the motion platform, which is an indicator of motion sickness incidence avoidance. The literature shows that lateral motion (roll, yaw, etc.) in the frequency range of 0.1-0.5 Hz induces motion sickness. Furthermore, using discrete-time LQR motion cueing algorithm has decreased the sensation error (motion platform-vehicle (cabin) levels) two times in terms of total roll jerk. In conclusion, discrete-time LQR motion cueing has reduced the simulator sickness more than the classical motion cueing algorithm depending on sensory cue conflict theory.Copyright © 2013 John Wiley & Sons, Ltd.

A perception coherence zone (PCZ) designates the range of inertial motion levels that, although not being a one-to-one match with the visual motion levels, are still considered by the subjects as being part of a coherent movement. Two types of PCZs were studied: amplitude PCZs and phase PCZs. Amplitude and phase coherences zones were measured for different types of stimulus frequency and amplitudes, for different degrees-of-freedom and using different motion simulators. The resulting data were compiled and it was demonstrated how perception coherence zones can be used to assess motion cueing solutions, and to derive motion cueing criteria. The coherence zones criteria were shown in a modified Sinacori plot. In a Sinacori plot, motion criteria are represented in terms of acceptable motion gain and phase distortion at the frequency of 1 rad/s. Since coherence zones were measured at different frequencies and amplitudes, it was possible to expand that representation to different frequencies and amplitudes. The coherence zone assessment method and criteria provide three important additions to the already available criteria. First, it presents not only criteria for desirable motion stimuli, but also offers a systematic, objective, human-perception-based method to measure the limits of the criteria. Second, the coherence zones method and criteria add a third and fourth dimension to the Sinacori plot: frequency and amplitude. By doing so, the coherence zone criteria do not depend on a specific motion filter structure. Third, by offering a measurement method and allowing different frequencies and amplitudes to be chosen, the coherence zones method can provide simulator-based, task-specific criteria. However, coherence zones as a metric, that is, as a measure of the perceived coherence of the inertial feedback provided, is platform and task independent.

A perception coherence zone (PCZ) designates the range of inertial motion levels that, although not being a one-to-one match with the visual motion levels, are still considered by the subjects as being part of a coherent movement. Two types of PCZs were studied: amplitude PCZs and phase PCZs. Amplitude and phase coherences zones were measured for different types of stimulus frequency and amplitudes, for different degrees-of-freedom and using different motion simulators. The resulting data were compiled and it was demonstrated how perception coherence zones can be used to assess motion cueing solutions, and to derive motion cueing criteria. The coherence zones criteria were shown in a modified Sinacori plot. In a Sinacori plot, motion criteria are represented in terms of acceptable motion gain and phase distortion at the frequency of 1 rad/s. Since coherence zones were measured at different frequencies and amplitudes, it was possible to expand that representation to different frequencies and amplitudes. The coherence zone assessment method and criteria provide three important additions to the already available criteria. First, it presents not only criteria for desirable motion stimuli, but also offers a systematic, objective, human-perception-based method to measure the limits of the criteria. Second, the coherence zones method and criteria add a third and fourth dimension to the Sinacori plot: frequency and amplitude. By doing so, the coherence zone criteria do not depend on a specific motion filter structure. Third, by offering a measurement method and allowing different frequencies and amplitudes to be chosen, the coherence zones method can provide simulator-based, task-specific criteria. However, coherence zones as a metric, that is, as a measure of the perceived coherence of the inertial feedback provided, is platform and task independent.

Two different topics are covered in the thesis. Model Predictive Control applied to the Motion Cueing Problem In the last years the interest about dynamic driving simulators is increasing and new commercial solutions are arising. Driving simulators play an important role in the development of new vehicles and advanced driver assistance devices: in fact, on the one hand, having a human driver on a driving simulator allows automotive manufacturers to bridge the gap between virtual prototyping and on-road testing during the vehicle development phase; on the other hand, novel driver assistance systems (such as advanced accident avoidance systems) can be safely tested by having the driver operating the vehicle in a virtual, highly realistic environment, while being exposed to hazardous situations. In both applications, it is crucial to faithfully reproduce in the simulator the driver's perception of forces acting on the vehicle and its acceleration. This has to be achieved while keeping the platform within its limited operation space. Such strategies go under the name of Motion Cueing Algorithms. In this work, a particular implementation of a Motion Cueing algorithm is described, that is based on Model Predictive Control technique. A distinctive feature of such approach is that it exploits a detailed model of the human vestibular system, and consequently differs from standard Motion Cueing strategies based on Washout Filters: such feature allows for better implementation of tilt coordination and more efficient handling of the platform limits. The algorithm has been evaluated in practice on a small-size, innovative platform, by performing tests with professional drivers. Results show that the MPC-based motion cueing algorithm allows to effectively handle the platform working area, to limit the presence of those platform movements that are typically associated with driver motion sickness, and to devise simple and intuitive tuning procedures. Moreover, the availability of an effective virtual driver allows the development of effective predictive strategies, and first simulation results are reported in the thesis. Control Techniques for a Hybrid Sport Motorcycle Reduction of the environmental impact of transportation systems is a world wide priority. Hybrid propulsion vehicles have proved to have a strong potential to this regard, and different four-wheels solutions have spread out in the market. Differently from cars, and even if they are considered the ideal solution for urban mobility, motorbikes and mopeds have not seen a wide application of hybrid propulsion yet, mostly due to the more strict constraints on available space and driving feeling. In the thesis, the problem of providing a commercial 125cc motorbike with a hybrid propulsion system is considered, by adding an electric engine to its standard internal combustion engine. The aim for the prototype is to use the electrical machine (directly keyed on the drive shaft) to obtain a torque boost during accelerations, improving and regularizing the supplied power while reducing the emissions. Two different control algorithms are proposed 1) the first is based on a standard heuristic with adaptive features, simpler to implement on the ECU for the prototype; 2) the second is a torque-split optimal-control strategy, managing the different contributions from the two engines. A crucial point is the implementation of a Simulink virtual environment, realized starting from a commercial tool, VI-BikeRealTime, to test the algorithms. The hybrid engine model has been implemented in the tool from scratch, as well as a simple battery model, derived directly from data-sheet characteristics by using polynomial interpolation. The simulation system is completed by a virtual rider and a tool for build test circuits. Results of the simulations on a realistic track are included, to evaluate the different performance of the two strategies in a closed loop environment (thanks to the virtual rider). The results from on-track tests of the real prototype, using the first control strategy, are reported too.

Virtual environments, and specifically motion-based driving simulators, raise the delicate question of multisensory cueing in order to produce the good "illusion" to the users. If virtual reality is well acknowledged for cost limitation, validity issue of these research (or training) tools is critical for knowledge transfer of new results.As a first step in the validation process of the IFSTTAR motorcycle simulator, this work relies upon an original design based on a double approach : by element and global. It consists in (i) optimal balancing of simulator’s immersive and interactive characteristics linked to roll motion so as to yield a believable and embraced leaning illusion, and (ii) evaluating more generally different simulator configurations (vehicle dynamic models) through objective (training length) and subjective measures (presence, simulator sickness).Taken as a whole, the seven studies conducted during this thesis enable the validation of the relevant range of roll motion cueing to produce a leaning sensation without leading to a critical unbalance or fall sensation. A leaning angle of the simulator beyond 11.4 degrees is not recommended to avoid any fall sensation, however this figure can be influenced by various factors (e.g., visual informations, roll axis location). These studies also allow to identify, in passive driving situations, the necessary parameters as well as the relative contribution of visual and inertial informations for the production of a believable leaning sensation. In an active driving situation, the evaluation of the degree of users control through different (virtual) vehicles dynamic models has permitted to highlight the weaknesses of IFSTTAR simulator in order to set several guidelines for further development., Les environnements virtuels, et en particulier les simulateurs de conduite à plate-forme dynamique, posent la délicate question de la restitution multisensorielle en vue de générer la bonne "illusion" pour leurs utilisateurs. Si la réalité virtuelle est reconnue pour la limitation des coûts, le gain temporel et la contrôlabilité et reproductibilité des situations étudiées, la problématique de la validité de ces outils de recherche (ou de formation) est cruciale pour la transférabilité des connaissances produites.Constituant une première étape dans le processus de validation du simulateur IFSTTAR, ce travail s’appuie sur une démarche originale reposant sur une double évaluation, par élément et globale. Il s’agit (i) de configurer de manière optimale les caractéristiques immersives et interactives du simulateur liées au mouvement de roulis afin de produire une illusion d’inclinaison crédible et acceptée, et (ii) d’évaluer, de manière globale, différentes configurations du simulateur (modèles dynamiques de véhicule) au moyen de mesures objectives (temps de familiarisation) et subjectives (présence, mal du simulateur).Prises dans leur ensemble, les sept études menées durant cette thèse ont permis de valider une plage de restitution du mouvement de roulis pertinente pour produire une sensation d’inclinaison sans entraîner de déséquilibre critique ou de sensation de chute. Un angle d’inclinaison du simulateur au-delà de 11,4 degrés est ainsi déconseillé pour éviter toute sensation de chute, néanmoins cette valeur est susceptible d’être influencée par divers facteurs (e.g., présence d’informations visuelles, positionnement de l’axe de roulis). Ces études ont également permis d’identifier, en conduite passive, les paramètres nécessaires ainsi que les contributions relatives des informations visuelle et inertielle pour la production d’une sensation d’inclinaison crédible. En conduite active, l’évaluation du degré de contrôle des participants en fonction du modèle dynamique de véhicule (virtuel) a permis de pointer les faiblesses actuelles du simulateur IFSTTAR afin de proposer plusieurs pistes de développement.

Compared to aircraft, flight simulators are severely limited in their motion envelopes. Presenting the true aircraft motion one-to-one on flight simulators is generally impossible and it is therefore common practice that these motion stimuli are only presented in reduced and attenuated form. Because of a limited understanding of human motion perception processes and how these affect the perceived realism of the multimodal stimuli pilots are subjected to during aircraft control (most notably, visual and physical motion stimuli), the definition of requirements for flight simulator motion cueing fidelity is a problem that researchers and legislators have struggled with for years. This thesis therefore describes and evaluates an objective method for the assessment of simulator cueing motion fidelity. The proposed method is centered around an analysis of the control dynamics adopted by pilots during manual control tasks, and how they use visual and motion stimuli in their selected control strategy, using multimodal pilot models. This approach thereby allows for the objective and quantitative evaluation of flight simulator motion fidelity, by explicitly considering how degraded motion cueing fidelity affects a simulator's ability to induce real-flight manual control behavior. This thesis describes a number of experiments in which pilot manual control behavior was measured using this approach in the Cessna Citation II laboratory aircraft and the SIMONA Research Simulator at Delft University of Technology. A comparison of the collected measurements clearly shows that variations in simulator motion cueing fidelity result in changes in pilot manual control behavior. With increased motion cueing fidelity, pilots are seen to rely significantly more on the presented motion stimuli, a control strategy that also typically results in increased manual control performance. Furthermore, these experiments also show that important behavioral parameters that characterize the weighing of visual and motion information by pilots also correspond best with those measured for in-flight pilot behavior when simulator motion cues are close to those of real flight.