| Abstract | U novim generacijama vozila ugrađuje se sve veći broj mehatroničkih sustava s ciljem poboljšanja sigurnosti, udobnosti i agilnosti vožnje. Tako se razvijaju razni sustavi aktivne dinamike vozila koji upravljaju ponašanjem vozila u uzdužnoj, bočnoj i vertikalnoj osi, s primarnim ciljem smanjenja broja i težine nesreća te povećane upravljivosti i udobnosti vozila. U konvencionalnim vozilima uobičajeno se koriste: 1) sustavi za sprječavanje blokiranja kotača koji upravljaju uzdužnom dinamikom vozila na način da održavaju kočnu silu u blizini maksimalno ostvarive vrijednosti, te tako maksimiziraju deceleraciju vozila, te 2) elektronički sustavi stabilnosti koji stabiliziraju vozilo tako da koče odabrani kotač, pri čemu se stvara odgovarajući moment skretanja, a u isto vrijeme usporava vozilo. S druge strane, aktivni ovjes uobičajeno se primjenjuje u svrhu upravljanja vertikalnom dinamikom vozila s ciljem poboljšanja udobnosti vožnje. U ovom se radu koriste pseudospektralni kolokacijski postupci optimiranja upravljačkih trajektorija aktivnog ovjesa radi istraživanja mogućnosti proširenja njegove primjene u širem kontekstu upravljanja dinamikom vozila. Pritom se koriste dva matematička modela dinamike vozila različite složenosti: 1) četvrtinski model vozila proširen s nelinearnostima vezanim uz limit hoda ovjesa i gubitak normalne sile autogume kod odskoka kotača, 2) potpuniji model vozila s 10 stupnjeva slobode s opcijski uključenim podmodelom neovješene mase. Prvo se istražuju mogućnosti unaprjeđenja karakteristika vertikalne dinamike vozila u prisustvu naglašenih izbočina i ulegnuća na cesti, a uzimajući u obzir realna ograničenja ovjesa i njegovog aktuatora. Zatim se analiza proširuje na istraživanje mogućnosti unaprjeđenja karakteristika bočne dinamike vozila pri slijeđenju referentne trajektorije gibanja, za tri tipa manevra dvostruke izmjene pravca. Pritom se pored isključive primjene aktivnog ovjesa istražuje i njegovo kombiniranje s aktuatorima sustava aktivnog skretanja. U radu se zatim istražuju mogućnosti unaprjeđenja karakteristika bočne stabilnosti vozila koristeći standardizirani ESP test. Konačno, analiza se proširuje na zadatak skraćenja zaustavnog puta vozila u kombinaciji sa sustavom za sprječavanje blokiranja kotača. |
| Abstract (english) | In the last couple of decades different types of active safety systems, such as those based on active brakes, active differentials, active steering and active suspensions, have been developed for vehicle controls in the longitudinal, lateral and vertical directions. The primary goal of such systems is to reduce the number and severity of accidents, and to improve the vehicle handling, comfort and agility. Anti-lock braking systems (ABS) aim to control longitudinal dynamics by keeping the longitudinal tyre force close to its peak value, thus maximising the vehicle deceleration. On the other hand, the conventional electronic stability control (ESC) systems improve the vehicle handling stability and responsiveness (i.e. oversteer and understeer compensation, respectively) by braking solely or predominantly one of the wheels, where a proper amount of active yaw torque is generated and at the same time the vehicle is decelerated. Simultaneously, the semi-active suspension (CDC) and active suspension (FAS) systems have been developed for improving the ride comfort performance, while preserving a high level of road holding ability. The main disadvantage of ESP system is related to brake activation, which produces a characteristic noise, vibration, and harsheness (NVH) content, and affects the vehicle agility. On the other hand, active steering or active differential actuators can provide superior handling performance without being intrusive to the driver (no NVH content) and without affecting agility (minimal reduction of the vehicle velocity). In advanced vehicle dynamics controls and related active safety, integrated control of several actuators can be used to improve the vehicle safety and performance. Examples include combining active rear steering and active central differential, active front steering and brakes, and a general case of independent four-wheel-steering (4WS) combined with four-wheel-braking (4WB). The fully active suspension (FAS) is considered in this thesis to analyze ultimate potential benefits of the vertical control contribution in rejecting the influence of emphasized discrete road disturbances (such as bumps and potholes), stabilizing the vehicle and increasing the path following accuracy, and reducing the braking distance. Design of a FAS-based vehicle dynamics control system can be a difficult task, because it is not intuitively clear how to coordinate the independent four-wheel FAS actions to maximize vehicle dynamics control performance, particularly in the presence of realistic constraints such as the suspension stroke limit. In that regard, it is generally convenient to first conduct numerical optimization of time response of FAS control variables in an off-line manner. Such optimal open-loop control results can be used to assess different vehicle dynamics actuator configurations, set realistic targets for achievable performance of more realistic (feedback) control systems, and guide the feedback control system design and tuning processes. The main aim of the thesis is to propose and implement an approach to optimization of vehicle dynamics control variables, and use the optimization approach to investigate at what extent active suspension actuator can improve the performance of longitudinal, lateral and vertical vehicle dynamics in different maneuvers. The optimization results are used to reveal and analyze control mechanisms of the FAS actuator and assess related FAS control performance. The thesis is organized in eight chapters, whose content is summarized as follows: Chapter 1: Introduction. Outlines the motivation for the conducted research, presents the literature review and provides the main hypothesis and an overview of the thesis. Chapter 2: Vehicle Dynamics Models. Describes the mathematical models of vehicle dynamics, which are used in the thesis, and which include: (i) low-order, quarter-car vehicle model, extended with nonlinearities related to the suspension travel limits and the lower (zero) tire normal load limit, and (ii) a more complex, 10-DOF vehicle model with an option of including the unsprung mass dynamics. In addition, a review of active safety systems is given, which includes semi-active and active suspension for vertical vehicle dynamics control, and ESP and ABS system for lateral and longitudinal vehicle dynamics control, respectively. Chapter 3: Optimal control. Gives an overview of different optimization problems, corresponding numerical methods, and advanced programming tools. The emphasis is on collocation methods for solving differentual equations, pseudospectral methods for solving optimal control problems, and SQP method for solving nonlinear programming problem, which are implemented within the programming tool TOMLAB, used in this thesis. Through a simple theoretical example and also a FAS control example, the process of optimal control problem definition and solving using the programming tool TOMLAB is presented. The process includes the cost function and constraints definition, and transformation of optimal control problem into nonlinear programming (NLP) problem, which can be solved with SQPbased algorithm packages SNOPT and KNITRO. Chapter 4: Optimization of active suspension control inputs for improved vehicle ride performance. Describes at what extent the FAS actuator can improve the vehicle ride performance. Control variable optimization is conducted in the presence of specific discrete disturbances such as large potholes and bumps (with high amplitudes and/or sharp edges). The ride performance improvements are investigated for different road preview lengths, and different disturbance shapes and sizes. In the pothole case, in addition to the conventional cost function term which penalizes sprung mass acceleration, the cost function includes the FAS energy consumption and wheel damage penalization terms, The latter is to minimize the sensitivity of wheel/tire to damage at the pothole trailing edge impact. Chapter 5: Optimization of active suspension control inputs for improved vehicle handling performance. Analyzes the FAS authority and related mechanisms of vehicle handling control for a path following task. The control variable optimization procedure is described for different vehicle actuator configurations and three types of double-lane change maneuvers, where the more complex 10-DOF model is used. The emphasis is on investigating: 1) active suspension authority on lateral vehicle dynamics control, which is revealed to include the oversteer and understeer compensation based on front/rear tyre load transfer and lateral acceleration boost, and comparison of optimization results with corresponding results for other actuator configurations, and 2) combination with other actuators, with emphasis on the active front and rear steering systems (AFS, ARS). In addition, the influence of secondary vehicle dynamics effects is analyzed: (i) influence of the FAS actuator bandwidth, (ii) influence of the tire relaxation length, camber, and toe effects, and (iii) influence of the fast unsprung mass dynamics. Chapter 6: Optimization of active suspension control inputs for improved vehicle stability control. Extends the handling control analysis from Chapter 5 to vehicle dynamic stability anaylsis based on the standardized sine-with-dwell maneuver. The optimised FAS control performance is compared with those of a standard ESP system and optimised brake control system. The cost function takes into account the yaw rate tracking error term (which indirectly reflects the lateral stability criterion) and lateral displacement amplitude (which is maximized for improved lateral responsiveness). Full (four channel) and reduced (one-three channel) FAS control formulation are considered, in order to investigate if the ultimate FAS performance can be approached by simpler control structures. Chapter 7: Optimization of active suspension control inputs for braking distance reduction. Considers the fixed-time straight-line braking to investigate the potential for braking distance reduction by means of the FAS control action. The analysis includes different tire-road friction coefficient profiles (such as constant-, transient-, and split-). In some scenarios, a simple ABS model with a sinusoidal longitudinal slip target is included, in order to mimic the ABS limit-cycle behaviour for the purpose of more realistic control variable optimization. Chapter 8: Conclusion. Outlines the main findings and the following major contributions of the doctoral thesis: 1) formulation and implemention of control variable optimization model for different active vehicle dynamics system configurations and different maneuvers, with nonlinear and discontinuous vehicle and tire dynamics effects included; 2) revealing active suspension control mechanisms for improving vehicle comfort, handling, and wheel damage resilience in the presence of discrete road disturbances, such as large bumps and potholes, and realistic suspension and actuator constraints; and 3) revealing active suspension control mechanisms for lateral vehicle stability improvement and braking distance reduction in the interaction with active steering systems, and ESP and ABS actuators. |
