|
| |
|
|
Vehicle Dynamics Measurement Analysis Using Mathematical Modelling
Blaťák, Ondřej ; Kledus, Robert (referee) ; Mazůrek, Ivan (referee) ; Kaplan, Zdeněk (advisor)
This doctoral thesis is focused in connection of measurement with a model in this meaning. Only relative easier measurable variables are measured on a vehicle and these variables make model inputs. This process can be improved by using dynamic filtering principles, where measured variables are used not only as model inputs, but they are used for the correction of a model motion too. We can obtain an optimal estimate of vehicle real behaviour during a manoeuvre with using this method. Linear discrete Kalman filter (dynamic filter) was used for the connection of the measurement with the mathematic model with regard to chosen model (planar linear single track model). The main task of Kalman filter is try to estimate the state of the discrete time depending process that is determined by the linear stochastic difference equation. Effective measurement system (based on CompactRIO unit) was built within the scope of this PhD thesis. This system is able to measure simultaneously a lot of quantities on the experimental vehicle. Velocity components (measured through optical correlative sensor Correvit), yaw velocity (measured through vibrating gyroscope), lateral acceleration, distance component (measured through MicrosatGPS system) are the basic measured parameters. Big attention was paid to obtain other important inputs of the mathematic model such centre of gravity position, moment of inertia, steering ratio etc. The last part of the thesis is oriented on the real connection of the measurement with the mathematic model. Formula Ford 1600 was used as the experimental car. The driving manoeuvre was arranged according to the norm ISO/WD 3888/2. Three versions of the Kalman filter were realised, which differs in the measured variables used for the mathematical model motion correction. Other possibilities, how to improve the process of state variables estimation eventually how to effective use this algorithm for vehicle parameters identification (cornering stiffness of the tires for example), are sketched in the end of the thesis.
|
|
|
Construction of an Experimental Vehicle Chassis Using the Lego System
Kratochvíl, Jan ; Porteš, Petr (referee) ; Svída, David (advisor)
This bachelor's thesis deals with the development of an experimental vehicle chassis using LEGO system parts with the aim of creating a platform for development of autonomous driving. To achieve this goal, the experimental vehicle chassis must meet the parameters and characteristics of real automobiles. Therefore, research was conducted to investigate the parameters and examine requirements for steering geometry, wheel bearings, arm structures, hub structures, and other key components of the chassis, as well as the approaches used to address these parameters in official LEGO models. Based on these findings, an initial model was assembled using only LEGO system parts, on which these parameters were evaluated. Limitations of the LEGO system were identified in wheel bearings, arms, and hubs. Subsequently, design modifications were made to these parts, including the use of rolling ball bearings and the production of arms and hubs using 3D printing technology. The second model, constructed using these modified parts, already met the mentioned requirements and approached the chassis of real automobiles.
|
|
|
Vehicle Dynamics Measurement Analysis Using Mathematical Modelling
Blaťák, Ondřej ; Kledus, Robert (referee) ; Mazůrek, Ivan (referee) ; Kaplan, Zdeněk (advisor)
This doctoral thesis is focused in connection of measurement with a model in this meaning. Only relative easier measurable variables are measured on a vehicle and these variables make model inputs. This process can be improved by using dynamic filtering principles, where measured variables are used not only as model inputs, but they are used for the correction of a model motion too. We can obtain an optimal estimate of vehicle real behaviour during a manoeuvre with using this method. Linear discrete Kalman filter (dynamic filter) was used for the connection of the measurement with the mathematic model with regard to chosen model (planar linear single track model). The main task of Kalman filter is try to estimate the state of the discrete time depending process that is determined by the linear stochastic difference equation. Effective measurement system (based on CompactRIO unit) was built within the scope of this PhD thesis. This system is able to measure simultaneously a lot of quantities on the experimental vehicle. Velocity components (measured through optical correlative sensor Correvit), yaw velocity (measured through vibrating gyroscope), lateral acceleration, distance component (measured through MicrosatGPS system) are the basic measured parameters. Big attention was paid to obtain other important inputs of the mathematic model such centre of gravity position, moment of inertia, steering ratio etc. The last part of the thesis is oriented on the real connection of the measurement with the mathematic model. Formula Ford 1600 was used as the experimental car. The driving manoeuvre was arranged according to the norm ISO/WD 3888/2. Three versions of the Kalman filter were realised, which differs in the measured variables used for the mathematical model motion correction. Other possibilities, how to improve the process of state variables estimation eventually how to effective use this algorithm for vehicle parameters identification (cornering stiffness of the tires for example), are sketched in the end of the thesis.
|
|
| |
guest ::
