Traffic congestion is an important reason for driver frustration which in turn is the main cause of human errors and accidents. Statistics reports have shown that over 90% of accidents are caused by human errors. Therefore, it is vital to improve vehicle controls to ensure adequate safety measures in order to decrease the number of accidents or to reduce the impact of accidents. An application of mathematical control techniques to the longitudinal dynamics of a vehicle equipped with an adaptive cruise control (ACC) system is presented. This study is carried out for the detailed understanding of a complex ACC vehicle model under critical transitional manoeuvres (TMs) in order to establish safe inter-vehicle distance with zero range-rate (relative velocity) behind a preceding vehicle. TMs are performed under the influence of internal complexities from vehicle dynamics and within constrained operation boundaries. The constrained boundaries refer to the control input, states, and collision avoidance constraints. The ACC vehicle is based on a nonlinear longitudinal model that includes vehicle inertial and powertrain dynamics. The overall system modelling includes: complex vehicle models, engine maps construction, first-order vehicle modelling, controllers modelling (upper-level and lower-level controllers for ACC vehicles). The upper-level controller computes the desired acceleration commands for the lower-lever controller which then provides the throttle/brake commands for the complex vehicle model. An important aspect of this study is to compare four control strategies: proportional-integral-derivative; sliding mode; constant-time-gap; and, model predictive control for the upper-level controller analysis using a first-order ACC vehicle model. The first-order model represents the lags in the vehicle actuators and sensor signal processing and it does not consider the dynamic effects of the vehicle’s sub-models. Furthermore, parameter analyses on the complex ACC vehicle for controller and vehicle parameters have been conducted. The comparison analysis of the four control strategies shows that model predictive control (MPC) is the most appropriate control strategy for upper-level control because it solves the optimal control problem on-line, rather than off-line, for the current states of the system using the prediction model, at the same time being able to take into account operation constraints. The analysis shows that the complex ACC vehicle can successfully execute TMs, tracking closely the desired acceleration and obeying the constraints, whereas the constraints are only applied in the MPC controller formulation. It is found that a higher length of the prediction horizon should be used for a closed acceleration tracking. The effect of engine and transmission dynamics on the MPC controller and ACC vehicle performance during the gear shifting is studied. A sensitivity analysis for MPC controller and vehicle parameters indicates that a length of the control horizon that is too high can seriously disturb the vehicle behaviour, and this disturbance can be only removed if a higher value of control input cost weighting is used. Furthermore, the analysis indicates that a mass within the range of 1400-2000 kg is suitable for the considered ACC vehicle. It is recommended that a variable headway time should be used for the spacing control between the two vehicles. It is found that the vehicle response is highly sensitive to the control input cost weighting; a lower value (less than one) can lead to a highly unstable vehicle response. It is recommended that the lower-level controller must take into account the road gradient information because the complex ACC vehicle is unable to achieve the control objectives while following on a slope. Based on the results, it is concluded that a first-order ACC vehicle model can be used for the controller design, but it is not sufficient to capture the complex vehicle dynamic response. Therefore, a complex vehicle model should be of use for the detailed ACC vehicle analysis. In this research study the first-order ACC vehicle model is used for the complex vehicle validation, whereas the complex ACC vehicle model can be used for the experimental validation in future work.
Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:541301 |
Date | January 2011 |
Creators | Ali, Zeeshan |
Publisher | University of Nottingham |
Source Sets | Ethos UK |
Detected Language | English |
Type | Electronic Thesis or Dissertation |
Source | http://eprints.nottingham.ac.uk/11977/ |
Page generated in 0.0021 seconds