Search Results

Lane keeping assist (LKA) is an autonomous driving technique that enables vehicles to travel along a desired line of lanes by adjusting the front steering angle. Reinforcement learning (RL) is one kind of machine learning. Agents or machines are not told how to act but instead learn from interaction with the environment. It also frees us from coding complex policies manually. But it has not yet been successfully applied to autonomous driving. Two control strategies using different deep reinforcement learning (DRL) algorithms have been proposed and used in the lane keeping assist scenario in this paper. Deep Q-network (DQN) algorithm with discrete action space and deep deterministic policy gradient (DDPG) algorithm with continuous action space have been implemented, respectively. Based on MATLAB/Simulink, deep neural networks representing the control policy are designed. The environment as well as the vehicle dynamics are also modelled in Simulink.

As a global optimization method, dynamic programming (DP) can be employed to seek the optimal velocity with minimum energy consumption for EV on given driving cycles. Due to its terrible computational burden, conventional DP is not suitable for real-time implementation especially with higher dimensions. In this paper, we propose an iterative dynamic programming (IDP) approach to reduce computing time firstly. The IDP can obtain the optimal control laws alike the conventional DP by converging the optimal control strategy iteratively and save considerable computing time. Second, the developed IDP and model predictive control (MPC) are combined to establish a real-time cruising controller called IDP-MPC for an EV with terrain preview. In the predictive controller, we use the IDP to solve a constrained finite horizon nonlinear optimization problem.

In order to reflect the actual power consumption of logistics electric vehicles in a city, sample real vehicle road data. After preprocessing, the short-stroke analysis method is used to divide it into working blocks of no less than 20 seconds. Based on principal component analysis, three of the 12 characteristic parameters were selected as the most expressive. K-means clustering algorithm is adopted to obtain the proportions of various short strokes, according to the proportion, select the short stroke with small deviation degree to combine, and construct the driving cycle, it has the characteristics of low average speed, high idle speed ratio and short driving distance. AVL-cruise software builds the vehicle model and runs the driving cycle of urban logistic EV. Compared with WLTC, the difference in power consumption is 34.3%, which is closer to the actual power consumption, the areas with the highest motor speed utilization are concentrated only in the idle area.

In this paper, for the front wall of a certain automobile, the defects of drawing splits, excessive thinning and excessive springback in the sheet metal forming process are analyzed and predicted. The stamping process has been simulated. The influence of different technical parameters (blank holder force, stamping speed, die gap and friction coefficient) on the forming results was further investigated using the center composite experiment. Through preliminary finite element simulation, the main drawing defects and trimming springback were analyzed. The second-order response surface model was established to perform the multi-objective optimization design of the stamping process with a NGSA-II genetic algorithm. Based on the relevant simulation data, multiple springback compensations are performed on the die surface to reduce the final springback of the part to meet the requirements.

This paper presents a novel torque distribution strategy (TDS) and a modified regenerative braking strategy (MRBS) for distributed-drive electric vehicle (DDEV) considering energy saving and brake stability. The presented TDS minimizes the energy consumption from battery in driving process. In order to overcome the shortcomings by using polynomial approximation for motor efficiency and the local minima problem, an exhaustive search method (ESM) is proposed to obtain the optimal front-rear torque distribution ratio. First, the power summation of four in-wheel motors is selected as the cost function of the optimization problem. Second, the ESM is designed to obtain the optimal torque distribution ratio according to current torque demand and motor speed based on motor efficiency map. Maximum motor torque and tire-road conditions are taken as constraints. Third, a MRBS is proposed to improve energy recovery performance by take ECE R13 and motor efficiency into account.

The design of aluminum foam reinforced thin-walled tubes has garnered much interest recently due to the high energy absorption capacity of these tubes. As a new kind of engineering composite material, aluminum foam can hugely increase the crashworthiness capacity without sacrificing too much weight. In this paper, axisymmetric thin-walled hollow tubes with four different kinds of cross-sections (circular, square, hexagonal and octagonal) are studied to assess their performance for crashworthiness problems. It is found that the tube with square cross-section has the best crashworthiness performance under axial impact. To seek optimal designs of square aluminum foam reinforced thin-walled tubes, a surrogate modeling technique coupled with a multi-criteria particle swarm optimization algorithm has been developed, to maximize specific energy absorption (SEA) and minimize peak crash force (PCF).

In most cases, researches on the ride performance of air suspension system are based on simplified mathematical models which could be too theoretical or not be able to consider the coupling relationship between the various components so that they behave far away from the actual vehicle system. This paper represents the study on the ride performance of an air suspension vehicle based on the accurate complex whole vehicle model which was established though ADAMS and Matlab. The applying of flexible components helped to improve the model accuracy. The stretching and compression test results of the air spring were used to establish the interconnected four-gasbag air suspension system. The vehicle ride performance was studied through the co-simulation between ADAMS and Matlab. The accuracy of the results were verified by the vehicle test results, which demonstrated the reliability of the whole model.

Gearbox fault diagnosis is one of the core research areas in the field of rotating machinery condition monitoring. The signal processing-based bearing fault diagnosis in the gearbox is considered as challenging as the vibration signals collected from acceleration transducers are, in general, a mixture of signals originating from an unknown number of sources, i.e. an underdetermined blind source separation (UBSS) problem. In this study, an effective UBSS-based algorithm solution, that combines empirical mode decomposition (EMD) and kernel independent component analysis (KICA) method, is proposed to address the technical challenge. Firstly, the nonlinear mixture signals are decomposed into a set of intrinsic mode function components (IMFs) by the EMD method, which can be combined with the original observed signals to reconstruct new observed signals. Thus, the original problem can be effectively transformed into over-determined BSS problem.

Inspired by the cross-section of a beetle’s elytron, hyperbolic lattice with double-layered feature has received increasing attention in recent years. This paper aims to investigate the compression behavior and energy absorption performance of a truss-based hyperbolic structure. The quasi-static compression simulation for the hyperbolic structure has been performed and validated with a compression test. Through compression simulation and test, the hyperbolic structure proved to show obvious twisting effect. To explore the influences of the geometric parameters in the mechanical properties of hyperbolic structure, this paper has designed 13 cells with varied rotation angle, height, and rod diameter, and investigated the mechanical properties of these configurations.

Due to the high center of gravity of medium-duty vehicles, rollover accidents can easily occur during high-speed cornering and lane changes. In order to prevent the deformation of the body structure, which would restrict the survival space and cause compression injuries to occupants, it is necessary to investigate methods for mitigating these incidents. This paper establishes a numerical model of right-side rollover for a commercial medium-duty vehicle in accordance with ECE R66 regulations, and the accuracy of the model is verified by experiment. According to the results, the material and size parameters of the key components of the right side pillar are selected as design variables. The response result matrix was constructed using the orthogonal design method for total mass, energy absorption, maximum collision acceleration, and minimum distance from the survival space.

Distributed drive electric vehicles (DDEVs), characterized by independently controllable torque at each wheel, redundant actuators, and highly integrated drive systems, are considered as the optimal platform for achieving intelligent driving with high safety and efficiency. This paper focuses on the trajectory tracking and lateral stability coordination control problems in high-speed emergency collision avoidance and autonomous lane change scenarios for DDEVs. A trajectory tracking control algorithm is proposed based on model predictive control (MPC) and coordinated optimization of distributed drive torques. The method adopts a hierarchical control architecture. Firstly, the upper-level trajectory planning layer calculates the lane change trajectory data. Based on the trajectory planning results, the middle-level controller designs a time-varying linear model predictive control method to solve the desired front wheel steering angle and additional yaw moment.