<p>Vehicle simulators are becoming more common in vehicle industries. Company earns lot of money on simulations instead of real tests. Real tests are necessary but not made so extensively as before.</p><p>In this thesis the building of an vehice simulator will be described and a comparison between three different cruise controls. The three cruise controls are PI-regulator, a regulator who regulates after positions in the terrain and a MPC-regulator. The reason for choosing this three is to see the difference between simple regulation and more complex regulation with respect to fuel consumption, travel time and complexity.</p><p>The vehicle simulator is made in Matlab/Simulink, Visual Studio and Open Scene Graph. The facilities needed for runnning the simulator is a relative good computer with a grapics card on at least 128 MB RAM plus a steering wheel and pedals for brake and gas to achieve best feeling. A keyboard can be used but it reduces almost all feeling.</p><p>After several simulations a conclusion was made. The MPC-regulator was the regulator who consumed least fuel and travel time. The regulator who regulates after positions in the terrain was not too far away. It would be interesting to do more research about it. In fact it is only a PI-regulator who makes ``clever'' decisions when a hill with enough slope appears. With enough slope means a downhill where the vehicle can accelerate without the use of fuel and an uphill where the vehicle can not keep its speed with maximum use of fuel.</p><p>A conclusion was stated that the friction and height profile influenced on settings for the PI-regulator and with some adjustments on this settings, fuel could be saved.</p>

In recent years, there has been a push by automotive manufacturers to improve the efficiency of the vehicle development process. This can be accomplished by creating a computationally efficient vehicle model that has the capability of predicting the vehicle behavior in many different situations at a fast pace. This thesis presents a procedure to automatically generate the simulation code of vehicle systems rolling over three-dimensional (3-D) roads given a description of the model as input. The governing equations describing the vehicle can be formulated using either a numerical or symbolical formulation approach. A numerical approach will re-construct numerical matrices that describe the system at each time step. Whereas a symbolic approach will generate the governing equations that describe the system for all time. The latter method offers many advantages to obtaining the equations. They only have to be formulated once and can be simplified using symbolic simplification techniques, thus making the simulations more computationally efficient. The road model is automatically generated in the formulation stage based on the single elevation function (3-D mathematical function) that is used to represent the road. Symbolic algorithms are adopted to construct and optimize the non-linear equations that are required to determine the contact point. A Newton-Raphson iterative scheme is constructed around the optimized non-linear equations, so that they can be solved at each time step. The road is represented in tabular form when it can not be defined by a single elevation function. A simulation code structure was developed to incorporate the tire on a 3-D road in a symbolic computer implementation of vehicle systems. It was created so that the tire forces and moments that appear in the generalized force matrix can be evaluated during simulation and not during formulation. They are evaluated systematically by performing a number of procedure calls. A road model is first used to determine the contact point between the tire and the ground. Its location is used to calculate the tire intermediate variables, such as the camber angle, that are required by a tire model to evaluate the tire forces and moments. The structured simulation code was implemented in the DynaFlexPro software package by creating a linear graph representation of the tire and the road. DynaFlexPro was used to analyze a vehicle system on six different road profiles performing different braking and cornering maneuvers. The analyzes were repeated in MSC.ADAMS for validation purposes and good agreement was achieved between the two software packages. The results confirmed that the symbolic computing approach presented in this thesis is more computationally efficient than the purely numerical approach. Thus, the simulation code structure increases the versatility of vehicle models by permitting them to be analyzed on 3-D trajectories while remaining computationally efficient.

In recent years, there has been a push by automotive manufacturers to improve the efficiency of the vehicle development process. This can be accomplished by creating a computationally efficient vehicle model that has the capability of predicting the vehicle behavior in many different situations at a fast pace. This thesis presents a procedure to automatically generate the simulation code of vehicle systems rolling over three-dimensional (3-D) roads given a description of the model as input. The governing equations describing the vehicle can be formulated using either a numerical or symbolical formulation approach. A numerical approach will re-construct numerical matrices that describe the system at each time step. Whereas a symbolic approach will generate the governing equations that describe the system for all time. The latter method offers many advantages to obtaining the equations. They only have to be formulated once and can be simplified using symbolic simplification techniques, thus making the simulations more computationally efficient. The road model is automatically generated in the formulation stage based on the single elevation function (3-D mathematical function) that is used to represent the road. Symbolic algorithms are adopted to construct and optimize the non-linear equations that are required to determine the contact point. A Newton-Raphson iterative scheme is constructed around the optimized non-linear equations, so that they can be solved at each time step. The road is represented in tabular form when it can not be defined by a single elevation function. A simulation code structure was developed to incorporate the tire on a 3-D road in a symbolic computer implementation of vehicle systems. It was created so that the tire forces and moments that appear in the generalized force matrix can be evaluated during simulation and not during formulation. They are evaluated systematically by performing a number of procedure calls. A road model is first used to determine the contact point between the tire and the ground. Its location is used to calculate the tire intermediate variables, such as the camber angle, that are required by a tire model to evaluate the tire forces and moments. The structured simulation code was implemented in the DynaFlexPro software package by creating a linear graph representation of the tire and the road. DynaFlexPro was used to analyze a vehicle system on six different road profiles performing different braking and cornering maneuvers. The analyzes were repeated in MSC.ADAMS for validation purposes and good agreement was achieved between the two software packages. The results confirmed that the symbolic computing approach presented in this thesis is more computationally efficient than the purely numerical approach. Thus, the simulation code structure increases the versatility of vehicle models by permitting them to be analyzed on 3-D trajectories while remaining computationally efficient.

Vehicle simulators are becoming more common in vehicle industries. Company earns lot of money on simulations instead of real tests. Real tests are necessary but not made so extensively as before. In this thesis the building of an vehice simulator will be described and a comparison between three different cruise controls. The three cruise controls are PI-regulator, a regulator who regulates after positions in the terrain and a MPC-regulator. The reason for choosing this three is to see the difference between simple regulation and more complex regulation with respect to fuel consumption, travel time and complexity. The vehicle simulator is made in Matlab/Simulink, Visual Studio and Open Scene Graph. The facilities needed for runnning the simulator is a relative good computer with a grapics card on at least 128 MB RAM plus a steering wheel and pedals for brake and gas to achieve best feeling. A keyboard can be used but it reduces almost all feeling. After several simulations a conclusion was made. The MPC-regulator was the regulator who consumed least fuel and travel time. The regulator who regulates after positions in the terrain was not too far away. It would be interesting to do more research about it. In fact it is only a PI-regulator who makes ``clever'' decisions when a hill with enough slope appears. With enough slope means a downhill where the vehicle can accelerate without the use of fuel and an uphill where the vehicle can not keep its speed with maximum use of fuel. A conclusion was stated that the friction and height profile influenced on settings for the PI-regulator and with some adjustments on this settings, fuel could be saved.

The main goal of this thesis is to obtain lateral parameters of the Magic Formula tire model of a tire commonly used in Formula Student competition. Both the author and the supervisor of the thesis know the tire name and its specification, but the research company that did the tire testing and provided me with the date prohibits sharing this of data publicly, so the tire designation is not mentioned in this thesis. The first chapter covers main theoretical facts related to a tire, briefly describes some of the tire models and shows possibilities how to determine tire characteristics that are used in a tire model. The thesis describes how to process raw tire data measured during a laboratory experiment using scripts created in Matlab software. The inputs variables are slip angle, lateral force, normal force and inclination angle. Raw data are splitted into parts, main coefficients of the Magic formula model (B, C, D, E, Sh, Sv) are calculated and subsequently the lateral parameters are obtained using least square method to fit parameters into the measured data. The works gives two main outcomes. The first output is a set of Matlab scripts that can be used to determine lateral parameters of any tire that has the same input data format as presented. A TIR file of the Formula Student tire in case of lateral slip is the second result of the work. This can be used for vehicle dynamics simulation of Formula Student racing car. The thesis also offers a comparison between parameters, which I calculated, and those gained thanks to Optimum Tire software by Calspan research company. Additionally the work shows the effect of load and inclination angle on lateral force.

No description available.