This paper presents a model predictive control (MPC) strategy specifically designed to maintain a fixed switching frequency in boost converters. Conventional MPC techniques often suffer from fluctuating switching frequencies, due to the need for frequent re-tuning of weighting factors as operating conditions change. To overcome this limitation, the proposed approach introduces an adaptive cost function which incorporates the inductor current ripple and its influence on the switching frequency. By minimising unnecessary switching operations, the controller ensures a stable and predictable frequency, even under load fluctuations, reference voltage changes, and input voltage variations. Unlike conventional methods, the proposed strategy eliminates the need for frequent parameter tuning, providing a more robust, efficient, and reliable control solution. The proposed method is implemented and validated using a hardware-in-the-loop (HIL) setup with a ZedBoard platform and external ADC for real-time measurement. Comparative results with conventional MPC approaches demonstrate that the proposed strategy achieves a stable average switching frequency and maintains robust performance under diverse operating scenarios.
To improve the stability of distributed drive electric vehicles under extreme operating situations, a hierarchical direct yaw moment controller (DYC) was constructed based on model predictive control (MPC). The upper layer uses MPC to calculate additional yaw moment, while the lower layer uses quadratic programming algorithm to allocate driving torque. Furthermore, a joint simulation model is established by using Trucksim and Simulink, the simulation results show that the MPC control proposed in this paper can effectively reduce the peak value of side slip angle, yaw rate and lateral acceleration of the vehicle. Under the condition of double lane change, compared with LQR control, the MPC control is improved by 41.10%, 10.65% and 2.93% respectively; under the condition of angle step, compared with LQR control, it increased by 28.64%, 9.38% and 3.96% respectively. Lastly, the effectiveness of the control strategy was validated through hardware-in-the-loop (HIL) test.
This paper provides a comprehensive analysis of advanced algorithms for autonomous vehicles (AVs), with a focus on lane detection, object detection, and evaluation datasets. In the first part, both traditional and deep learning-based 2D/3D lane detection approaches are discussed in detail with their methodologies and applicability. In addition, the lane detection approaches are categorised into traditional and deep learning-based approaches for 2D/3D lane detection, highlighting their characteristics and performance. Subsequently, in the second part, the study then shifts to 2D/3D traditional and deep learning-based object detection, including birds-eye-view detection and transformer network-based approaches, offering a critical analysis of these techniques and their potential impact on autonomous driving technology. Finally, an overview of state-of-the-art datasets for evaluating lane and object detection algorithms in autonomous driving research is presented. Overall, this survey consolidates the current state of research in lane and object detection algorithms for autonomous vehicles, providing insights and directions for future research in this field.
The study uses a multi-body dynamic (MBD) model built with MATLAB Simscape/Simmechanics to simulate the 2DOF quarter-vehicle system's responses during the optimisation of the structure as well as the controllers. The study applied the Pareto optimisation method in simultaneously determining elastic and damping elements' parameters of the bus suspension system by PSO to optimise at the same time the vehicle body's acceleration (acc) and suspension's relative displacement (rel). The study proposed a new design method for proportional derivative fuzzy logic (PD-FLC) and decoupled algorithm sliding-mode (DA-SMC) control systems on vehicle suspension when the membership functions' positions' parameters E/NS-PS, DE/NS-PS of the FLC set and the sliding surface's parameters eta and k of the SMC set are simultaneously optimised by PSO. The active suspension system is more effective than the passive system in reducing displacement (over 40%) and vehicle body acceleration (over 17%).
Retrofitting old or classic internal combustion engine vehicles into EV conversions is becoming an important approach to reduce urban pollution, reduce fossil fuel dependence, and improve the quality of life in urban areas, especially in Southeast Asia, where many old vehicles are in use. However, determining the right battery and drive system size for practical use remains a major challenge. This research proposes an approach to determine the right battery size based on urban dynamometer driving schedule (UDDS), which reflect urban driving behaviour with frequent acceleration, stop-start, and speed changes. The driving simulation using UDDS clearly identifies the relevant forces such as air resistance, rolling resistance, and acceleration, resulting in accurate calculation of energy requirements. The results show that using UDDS can calculate the battery size of a modified EV based on a standard driving cycle, compared to the conventional calculation method that assumes the motor is continuously running at full power. This result reflects the potential to reduce system size, lower costs, and optimise the design of modified EVs. In addition, this approach is suitable for use in regions with limited driving data and supports the development of sustainable vehicle solutions in the future.
Multi-objective optimisation of suspension kinematics based on the non-dominated sorting genetic algorithm II (NSGA-II) is a widely adopted technical approach. However, the NSGA-II algorithm has several limitations. Therefore, this study aims to address these limitations and improve the algorithm. The improvement begins by introducing chaotic mapping for population initialisation, adaptive crossover and mutation rates, and a dynamic elitism retention mechanism, resulting in the proposed chaotic adaptive non-dominated sorting genetic algorithm II (CA-NSGA-II). The performance of the CA-NSGA-II algorithm is then compared with other algorithms to validate its overall performance. Lastly, the CA-NSGA-II algorithm is applied to the multi-objective optimisation of the kinematics in a double wishbone composite rear suspension system. The results show that compared with other comparison algorithms, the CA-NSGA-II algorithm has better comprehensive performance. When solving the multi-objective optimisation problem of suspension kinematics, it is superior to NSGA-II algorithm and can provide better solutions.
Electric vehicles (EVs) have recently garnered significant attention due to their environmental friendliness, decreased running cost and superior performance. Owning to the exponential growth of electric vehicle production over the past decade, safety inspections of EVs have become more crucial. Since the effect of the restraint system on the occupant's safety has been already investigated, herein, the impact of the crumple zone characteristics on head, neck and chest injury has been comprehensively studied. In this paper safety assessment of electric quadricycle named EQ10 during the frontal crash condition has been considered. The safety parameters including HIC, N(ij )and chest deflection has been evaluated as three output parameters based on ECE-R94 standard. Taguchi and ANOVA approaches also have been applied to optimise safety parameters. The achieved data affirms that in the optimised aluminium structure, HIC has decreased by 32%. In the optimised steel, HIC has decreased by 33%.
Compared to radial and cornering fatigue tests, the biaxial fatigue test more accurately reproduces the in-service stress state of a wheel by applying simultaneous radial and lateral loads. This study proposes a novel finite element simulation method for passenger car wheels, incorporating both tyre dynamic rolling via steady-state transport analysis and the effect of tyre-wheel interference fit. The method is applied to a composite aluminium alloy wheel with different materials for the spoke and rim. Results indicate that stress distributions from the proposed and a simplified method are similar. However, the proposed method yields a higher maximum rim stress by up to 25.684% under certain loads, which is closer to the measured results by referring to the literature. This indicates that the proposed method is able to provide more accurate strength evaluation results of the passenger car wheel in biaxial fatigue test.
Air spring suspension enhances handling performance and passability by adjusting vehicle height and posture. This paper proposes a control strategy to regulate vehicle height and posture using air suspension. A mathematical model of a vehicle with seven degrees of freedom is derived to obtain the state function for the controller and the plant model for verification. An air spring suspension is integrated into the vehicle model, and a corresponding control strategy is developed. The proposed strategy consists of two control loops: an outer loop using model predictive control to regulate height and posture by adjusting the target air pressure in the air springs, and an inner loop employing a PID controller to modulate the valve and control airflow. To validate the strategy, two simulations are conducted, and the results demonstrate its effectiveness.
