ABSTRACT. 基于网联车辆的车路协同控制是下一代智能交通技术的核心。本次报告论述了车联网中的车辆通信调度、车辆行车速度和行车路径的动态优化、车辆队列协同控制等核心问题及其解决方案。侧重结合车辆动态特性、车辆间的耦合特性、车联网通信限制因素，给出绿色节能的最优车辆行驶速度和路径动态优化设定方法、车辆流稳定安全行驶所需的车辆通信调度方法和安全高效的车辆编队控制方法。并通过仿真和半实物实验结果，探讨了相关结果的可行性。

ABSTRACT. 近年来，车载信息系统（俗称车辆“黑匣子” ），在货运车辆上的安装与使用日益广泛。该系统主要用于记录车辆行驶状态，并将数据实时传送至物流企业内部，进而为企业在车辆监管、运营调度等方面提供决策情报；在优化资源配置、降低运营成本、提升物流效率等方面拓展空间。然后，当前商用车载信息系统在数据采集方面，其灵活性较差、可采集数据类型较少、采样频率偏低，因此其在提升车辆安全、节能与环保方面所发挥的作用十分有限。剑桥大学道路货运可持续发展研究中心自主开发了一套基于安卓平台的数据采集系统。该系统能够高频采集如下行驶信息： • 车辆位置：经/纬度、海拔等 • 车辆运动状态：纵向/侧向/垂向加速度、侧倾/俯仰/横摆角速度等 • 车辆CAN总线数据：车速、油耗、发动机转矩/转速、油门踏板开度、档位、自动巡航功能开启情况等 并能定期向云端发起连接，获取天气状况（包含环境温度、风速、风向等）与交通流状况。在此基础上，该数采系统能够通过4G网络将所获取的各种数据实时传送至服务端的数据监控中心。 本报告介绍了该安卓数采系统的功能与特点，展示了其在英国所采集的车辆行驶数据，举例说明了车辆数据在节能减排分析方面的具体应用，并概述了近期在中国所开展的数据采集活动的成果。

ABSTRACT. 车辆自动驾驶经过多年发展，在核心技术方面已经获得了一定的突破，已经具有多个实验验证系统，并在多个场景下获得了示范应用；同时自动驾驶作为一个复杂系统，受软件，硬件，外部环境等多方面影响，特别是受现有人工智能发展水平的制约，这些都对自动驾驶的应用构成了一定的挑战。

ABSTRACT. The unmanned roller (UR) can improve the compactions quality and efficiency simultaneously relative to the manually operated roller. The path-following control for the UR is, however, challenging, because of 1) the increased degree of freedom in motion due to the articulated structure of UR, and 2) the disturbance in rollers positioning and orientating caused by the large rocks on road. In this work, a dual-layer active disturbance rejection path-following controller based on a kinematic model is proposed, where the discrepancy of the model from the roller is lumped as total disturbance and rejected online. To reduce the measurement uncertainties from rocks, the rollers position and orientation is corrected in real-time by measuring the attitudes of the front and rear frames. Additionally, a learning controller for road condition variations is designed to improve the accuracy of the kinematic model for improved control performance. The proposed solution is validated systematically in the real-world production application for the compaction of Rock-filled dams.

ABSTRACT. 汽车智能化是解决安全、交通、能源、环境等全球性问题的重要手段，成为世界汽车工业的技术前沿和发展方向。面对多发的交通事故，智能化技术对于提高行车安全性不可或缺。面对汽车智能化程度的逐渐提高，驾驶决策与人机协同成为智能汽车发展不可回避的技术难点。智能决策可有效减少由于驾驶员误操作和驾驶恐慌所引起的交通事故，人机协同共驾既可以吸收积极的驾驶经验，又可以利用传感器和执行器的功能，可有效提高车辆系统的安全性，是降低交通事故的一个可行途径。但各工况下如何进行驾驶决策、分配人类驾驶员与自动控制系统的权重是亟待解决的问题。本报告详细介绍目前驾驶决策与人机协同转向的方案，并报告吉林大学陈虹课题组在智能车辆换道决策与人机协同转向控制的方案以及一些研究进展。

ABSTRACT. This paper presents a stochastic model predictive control (SMPC) design scheme for gasoline engines with exhaust gas recirculation (EGR) by considering the stochastic distribution of indicated mean effective pressure (IMEP) as constraint to avoid abnormal combustion during transient conditions. Firstly, a model predictive control (MPC) problem is formulated to minimize the system state tracking error based on a second order state space model. Then, the SMPC problem under stochastic IMEP distribution constraint is formulated. Moreover, in order to make the problem solvable, a scenario-based optimization approach is applied to transform the stochastic optimal control problem into a deterministic optimal control problem. Finally, the deterministic optimal control problem is solved by a multishooting based sequence quadratic programming (SQP). The simulation validation shows that the proposed SMPC scheme can avoid misfire during the tip put transient condition while guaranteeing the system states tracking performance without depending on the exhaust manifold pressure sensor.

ABSTRACT. This paper proposes a logical control approach for the gearshift strategy design aimed to improve the fuel consumption efficiency of the commuting vehicle such as city bus or trucks. The vehicle dynamics is discretized with respect to the route segment, and then it is further modeled as the stochastic logical dynamical system in view of the probability characteristic of the driver acceleration demand at the specific route segment. The fuel consumption optimization problem is constructed as a finite horizon optimal control problem under the framework of the proposed stochastic logical dynamical systems. Then the optimal gearshift strategy is obtained by applying a logical dynamic programming algorithm. Simulation results demonstrate the performance of the proposed control scheme.

ABSTRACT. During recently years, as the environment problem and resource shortage become serious largely due to vehicles’ fuel consumption, plug-in hybrid electric vehicles (PHEVs) with high fuel economy and reliability, catch a lot of attention. In this research, an energy management strategy (EMS), which serves as the core of PHEV’s performance, is introduced. The main part of the proposed EMS is based on fuzzy logic control, which takes the torque demand and State of Charge (SOC) as the input and engine torque as the output. Because fuzzy rules are usually determined by experience, genetic algorithm is applied to optimize the fuzzy rules to achieve better fuel economy via reproduction, crossover and mutation. The simulation is established on MATLAB/Simulink platform. The result demonstrates that the proposed strategy has 4.41% improvement on fuel consumption performance in comparison with fuzzy logic strategy before optimized. Moreover, the torque output is more stable than before optimized, while the SOC has basically realized the charge-discharge balance.

ABSTRACT. Look-up tables are usually used to characterize operating-point-dependent system variables in typical embedded applications due to the relatively low computational load. In this paper the problem of look-up table parameters updating is studied to cope with physical systems changes due to operating conditions and age. To this aim, an estimation algorithm is developed in order to update look-up table parameters in a finite-time. The convergence proofs are based on Lyapunov-based approach, and the results are obtained based on the regressor term related to the classical identifiability condition. The relationship between the estimation error and model approximation error is analyzed for the proposed algorithm. The validity of the proposed method is verified by updating the two-dimensional look-up table for the throttle discharge coefficient of a spark ignition gasoline engine form engine simulation tool enDYNA.

ABSTRACT. In this paper, the air-fuel ratio regulation problem of the compressed natural gas engines is researched. A state observer is designed based on the approximate linearization technology, since the air-fuel ratio is a nonlinear function of the total air mass and total fuel mass in cylinder. By the proposed state observer, a state feedback air-fuel ratio regulator is designed based on the dynamic model of the compressed natural gas engines, and the closed-loop system is mean-square stable. The validation of the proposed state feedback regulator is given by the numerical simulation under two working conditions. The simulation results show that the error of regulation of the air--fuel ratio can be regulated into a neighbourhood of its desired value under all of the working.

ABSTRACT. The problem of stochastic adaptive tracking control of the electronic throttle is researched, in which the spring gain and the external load torque are seen as an unknown constant and a Wiener process, respectively. A stochastic adaptive tracking controller is designed based on the backstepping technology, and the closed-loop system is proved to be bounded in probability. The validation of the proposed tracking controller is proved through numerical simulation under three working conditions. The simulation results show that the position of throttle plate can be tracked to its desired value, and the adaptive law can be tuned to a steady value under all of the working conditions.

ABSTRACT. Aiming at local obstacle avoidance path planning, a planning method combining force field(FF) and model predict control(MPC) is proposed. Firstly, the road area is divided according to the lane line, and the traffic scene is simulated by the FF. A simplified environment model including the virtual rectangular repulsion field of obstacle vehicles, the virtual gravity field at the driving target position and the virtual gravity field at the lane maintenance is established. Then, the kinematics vehicle model is established, and the multi-constrained and multi- objective optimization problem of path planning is constructed based on the MPC, and the acceleration and yaw rate control variables are solved. Finally, the validity of the path planning scheme is verified by simulation.

ABSTRACT. Abstract—Driving style is important to fuel economy, safety and intelligent transportation. With the development of intelligent vehicle and transportation, driving style needs be considered to achieve a human-centered control. In daily traffic, the drivers in a platoon are variable and the driving data is unlabeled. It is meaningful to classify driving styles from unlabeled data and apply it to human-centered control. A vehicle-following control methodology is proposed in this paper to promote fuel economy according to driving style. Firstly, car-following parameters are calibrated from unlabeled data. An unsupervised cluster is applied to divide drivers into three driving styles according to the Silhouette coefficient. Based on cluster centroids, a driving style classification model is established to recognize driving style from car-following data. Then, a four-layer BP neural network is built and its optimal topology and hyperparameters is chosen to develop a general driver model to predict vehicle following speed, and an advanced speed prediction model including Gipps model is proposed to promote adaptability for more scenarios; and the driving style based weighting coefficients are identified for online vehicle speed prediction. Finally, a car-following control strategy considering driving style is proposed and the safety speed with Gipps speed, driving style based boundary speeds and economy speed are derived. Simulation results show that: 1) the proposed speed control strategy can classify drivers into proper driving style; 2) the proposed strategy can be used for human-centered driving control, improving vehicle-following safety, traffic efficiency, fuel economy and humanity.

ABSTRACT. The current manual driving training venues and lack of human resources are one of the key issues that need to be solved in driving schools in China. Although the car driving simulator has been used in driving school training, it can only provide a driving experience, can not perceive the student's operation, and still cannot replace the real car training. In order to truly save resources and improve the efficiency of driving training, the model and decision-making research of the closed-loop coach system for driving teaching in accordance with the aptitude is introduced. Firstly determine the structure of the driving school training coach system, analyze the system influencing factors, and then establish the student model and coach model structure, and construct the driving school training coach system structure. The least squares method is used to identify the model and obtain the parameters of the student model and the coach model. Finally, the student model parameters are classified by fuzzy C-cluster algorithm. The coach model combines the student's characteristic parameters and makes targeted decision-making according to the students' classification results. The practical application results show that the proposed closed-loop teaching system for driving teaching in accordance with the material can fully carry out driving training according to the characteristics of different students.

ABSTRACT. In the driving teaching process, the characteristics of the students change greatly, and it is difficult to establish an accurate mathematical description model for the students. For the difficulty of modeling students, the trainees and vehicles are regarded as the whole, and the cellular automata model of driving school students and vehicles is established. Train the electronic driving school coach with reinforcement learning, and give coach instructions in real time. The simulation results show that the cellular automaton model established in this paper can simulate the vehicle motion under different student characteristics based on different parameter settings. The electronic driving school coach trained by reinforcement learning can effectively improve the motion of the cellular automaton model. As a result, reasonable driving teaching decisions are provided.

ABSTRACT. Stretch reflex is the basis of human activities. The moment of inertia, damping and stiffness of the human-machine system based on the least-squares method were used to characterize quantitatively the total stretch reflex intensity of the driver’s upper limb. In order to study the influence of vehicles on the driver in the human-machine co-drive system and the muscle mechanical characteristics during the driving operation, 12 drivers were invited to take part in the experiments and were divided into three groups according to their body shapes. Through the step signal experiments under the same torque, the equivalent static stiffness obtained from the steering angle is used to characterize the static stretch reflex intensity. And the stiffness ratio is proposed to characterize the dynamic stretch reflex intensity of the driver. Finally, the correlation between mechanical parameters and EMG (Electromyography) indices, steering efficiency and co-contraction, was analyzed. This research will lay a foundation for the control strategy of the human-machine co-drive system.

ABSTRACT. Great progress has been made in the field of machine learning in recent years. And learning-based methods have been widely utilized for developing highly autonomous vehicle. To this end, we introduce a reinforcement learning based intelligent autonomous vehicle decision making method for oncoming overtaking scenario. The goal of reinforcement learning is to learn how to take optimal decision in corresponding observations through interactions with the environment using a reward function to estimate whether the decision is good or not. A Double Deep Q-learning (Double DQN) agent was used to learn policies for both longitudinal speed and lane change decision. Prioritized Experience Replay (PER) was used to accelerate convergence of the policies. A two-way 3-car scenario with oncoming traffic was established in SUMO (Simulation of Urban Mobility) to train and test the policies.

ABSTRACT. This paper proposes a multi-cell-to-multi-cell (MC2MC) equalizer for series-connected lithium-ion battery strings based on flyback conversion and its control method. An onboard 12 V lead-acid battery is used as the energy transfer station to reduce a half of switches. Different from the traditional equalizer, this equalizer can directly transfer energy from the higher-voltage cells (i.e., the optimal discharging combination) to the lead-acid battery and deliver energy from the lead-acid battery to the lower-voltage cells(i.e., the optimal charging combination), which greatly improves the equalization speed. The pulse-width modulation (PWM) duty cycle can be adjusted according to the voltage of the optimal charging or discharging combination during equalization, so that the balancing current and speed will not decrease with the decrease of the voltage difference among cells. Moreover, the demagnetization for the transformer can be realized in the discontinuous mode of the bidirectional flyback converter, thereby improving the stability of the equalization system. The simulation results show that with the proposed equalizer, the cell voltage difference can be decreased from 540 mV to 2 mV within 1500s, leading to a fast balancing speed.

ABSTRACT. Pouch lithium ion batteries are widely applied in electronic vehicles, due to their high energy density, smaller internal resistance and convenient transportation. However, pouch-type batteries are easier inducing thermal inhomogeneities and cause safety problems, especially for large-sized batteries. Therefore, in order to study the thermal behaviors of Pouch Lithium Ion Battery, a distributed electro-thermal model is established in this paper, which is developed by coupling the network equivalent circuit model and the 2D thermal model. Using this model, we can obtain the distributions of temperature, local SOC and current density on the cell plates. Then, a simulation analysis based on this model has been given for the thermal behavior of a pouch cell under different discharge rates (1C, 2C, 3C, 4C). Results verify that the larger discharge rate always causes a higher temperature rise, what’s more, the highest temperature area is not invariable in the process of discharge.

ABSTRACT. The battery module of parallel-connected lithium-ion cells is extensively applied in electric vehicles to satisfy the capacity and power capability. However, the performance of the module is influenced by the interconnect resistances and the position of its current collector. An equivalent circuit simulation model is developed and validated by pulse discharge tests. Based on the model, a module of five cells connected in parallel is established, including interconnect resistances. Then the effect of interconnect resistances and different positions of module current collector in parallel-connected cells is tested and discussed by the simulation. Results indicate that the interconnect resistances lead to inhomogeneous currents, and the cell directly connected to the module current collector appears the highest current. That is, the farther the cell is away from the module current collector, the lower current is. The cause are clarified from an angle of the currents through interconnect resistances. Therefore, in order to prolong the module lifetime, the interconnect resistance should be as small as possible, and the module current collector should not be connected with the edge single cell.

ABSTRACT. An accurate mathematical model is vital for the state estimation and management of the lithium-ion battery. And the real data is the foundation of modeling. However, the noise in the measurement signals can lead to poor estimation accuracy of model parameters. Traditionally, the low-pass filter is designed to reduce the impact of noise, while the optimal cutoff frequency in the filter is difficult to select. In view of this situation, this paper proposes a voltage reconstruction method to restrain noise and ensure the reality of battery measurement data. To verify the effectiveness of the proposed method, equivalent circuit model (ECM) is established and model parameters are estimated by means of recursive least squares (RLS). Finally, the simulation results verify the feasibility of the proposed method.

ABSTRACT. The inconsistency of the battery cells has a great impact on battery grouping performance. In this paper, the inconsistency effect of internal resistance is analyzed by using the series-connected Rint battery model. And the difference of the parameter definition between the battery cell and the battery strings is analyzed. Through the theory analysis and simulation verification, some principles of battery cells grouping are achieved, which has an important reference significance to guide how to select and classify battery cells in the battery strings for better performance.

ABSTRACT. This paper introduced the structure and modeling process of a pure electric vehicle integrated thermal management system containing battery-motor heat exchanger, designed control strategies for it, and did some simulation research of its thermal management performance in high and low temperature environments and the energy-saving characteristics.

ABSTRACT. 为提高汽车座椅磁流变扭转减振器的力矩和可调范围，本文设计了一种基于螺旋流动模式的改进型磁流变扭转减振器。从设计角度对磁路分布和结构参数进行了优化，对所设计的器件在不同电流和正弦激励下的阻尼特性进行了测试。为了描述减振器的非线性阻尼滞回曲线，采用基于骨架曲线的建模方法搭建了动态力学模型。通过座椅悬架垂向振动实验台，采用天棚控制算法对磁流变座椅悬架进行了实验测试，结果验证了半主动系统的有效性。考虑到基于磁流变扭转减振器的座椅系统更易集成供电与传感模块，此系统将具有较好的应用前景。

ABSTRACT. 传统磁流变阻尼器由于其结构限制，节流通道横截面积与活塞大小及阻尼器其他结构尺寸是互相关联的，导致其阻尼力及可调阻尼比不能同时兼顾,致使冲击环境中应用的传统磁流变阻尼器结构复杂、尺寸过大。对此，本文提出一种新型节流通道旁置式磁流变阻尼器，基于磁流变液的流变特性，详细讨论了新型磁流变阻尼器设计过程中主要构件参数以及磁路尺寸的确定，通过对磁路结构的磁场有限元仿真分析验证了阻尼器结构的合理性，并研制了新型阻尼器样机，通过实验测试得到的力学性能数据表明，新型阻尼器既有较大的输出阻尼力又有较高的可调阻尼比，阻尼特性优良，验证了设计方案的可行性。

ABSTRACT. 电动汽车无线充电系统可实现电能从电网到动力电池的无线传输，无需人工操作、安全可靠、易于实现充电智能化。本文旨在设计并优化无线充电系统线圈结构，以获得合适的充电功率和最优的充电效率。首先对线圈内径和匝数进行组合寻优，并对铁氧体内径和收发端铁氧体相对角度进行优化，以获得较优的耦合线圈互感值。其次，基于有限元软件分析收发端线圈之间的水平、垂直偏移对线圈自感、互感、传输功率以及传输效率的影响，研究无线充电系统耦合线圈的偏移特性。最后，设计了一款6.6kW无线充电系统的耦合线圈，通过测量线圈参数，建立电路仿真模型，对比验证无线充电系统耦合线圈的偏移特性。

ABSTRACT. Sensitive equipment such as electric vehicle control systems and vehicle electronics are susceptible to the external electromagnetic environment. The radiated electromagnetic signals interfere with the normal driving of electric vehicle and threaten human life. In order to ensure safety and reliability of electric vehicles, electromagnetic (EM) absorbent has become a key technology. RGO/Ni was prepared by in-situ reduction method in this paper. Since composition, size, and special morphology have a vital influence on microwave absorption properties, the as-obtained products were characterized by X-ray diffraction, scanning electron microscopy and vector network analysis. Different reduction methods will bring different dielectric properties to the obtained RGO. The RGO obtained by Zn powder have a more complete structure, and therefore better dielectric properties. The prepared RGO/Ni-2 composite possesses excellent impedance matching characteristics and electromagnetic complementary effect, which are beneficial to microwave absorption performance. This work investigates the mechanism of absorption for the as-obtained RGO/Ni-2, offers a promising strategy for the fabrication of RGO/Ni and introduces the application of RGO/Ni as a highly efficient absorber in the field of electric vehicle.

ABSTRACT. 线控制动技术是汽车智能化/无人化技术的基石。电子机械式制动器是线控制动技术的重要解决方案，但是长时间连续制动时的电机堵转问题是其固有缺陷。为解决该问题，本文提出基于具有结构简单、响应迅速以及可控传递转矩范围大优点的磁流变离合器，实现电子机械式制动器在全工况范围内制动力连续、快速输出。 磁流变传动线控制动器主要由交流伺服电机、行星齿轮减速器、磁流变离合器及滚珠丝杠组成。行星齿轮减速器放大交流伺服电机输出转矩，并通过控制磁流变离合器中励磁线圈电流实现对滚珠丝杠端制动力的实时控制。首先，对磁流变离合器进行结构设计并电磁仿真以验证其可行性，在Mechanical APDL中进行了结构优化，并依据优化结果完成了原型加工。其次，对磁流变离合器的力学属性进行了标定实验和基于PID控制方法的输出转矩跟踪控制实验。最后，搭建了小尺寸磁流变传动线控制动器实验台架，并设计了分层控制策略，实现了任意制动力的实时跟踪实验与评价。研究结果对开展线控制动技术研究具有基础意义。

ABSTRACT. 磁流变制动器是基于磁流变效应的智能材料磁流变液的一种典型应用，通过对磁流变制动器内部励磁线圈施加不同励磁电流改变作用于磁流变液上的磁场强度，实现改变磁流变液材料的流变属性，进一步实现制动力矩的无级控制。磁流变制动器不仅结构简单、响应迅速，而且可控范围大，在汽车线控制动器方面应用具有广阔前景。 本文设计、优化并实现了一种磁流变线控制动器，其结构主要由输入轴、制动盘、固定盘、线圈及外壳组成，搭建了道路模拟台架并进行了磁流变线控制动器原型的道路模拟实验。具体地，针对某款微型轿车进行了制动器制动力矩匹配，设计了一款多盘式磁流变制动器，并通过ANSYS APDL编程对其进行了磁路设计与多目标结构优化。根据优化后的结构参数进行了磁流变线控制动器的原型开发与力学特性标定。基于1/4车辆道路模拟实验台架，搭建了制动实验硬件系统，并利用LabVIEW进行了制动系统的控制器设计。最后，基于搭建的制动系统试验台架中完成了单轮ABS控制实验，分析了该款纯线控制动器相对于传统液压制动器以及其他形式线控制动器的优势。

ABSTRACT. In the talk, a new observer design method to estimate simultaneously both the vehicle dynamics and the unknown driver’s torque will be presented. To take into account the timevarying nature of the longitudinal speed, the vehicle system is transformed into a polytopic linear parameter-varying (LPV) model with a reduced level of numerical complexity. Based on Lyapunov stability arguments, we prove that the estimation errors of the system state and of the unknown input (UI) are norm-bounded, which can be made arbitrarily small by minimizing a guaranteed L∞−gain performance. The design of the LPV unknown input observer is reformulated as an LMI-based optimization which can be effectively solved via semi-definite programming. Extensive hardware experiments are carried out under various driving test scenarios to confirm the effectiveness of the proposed observer design. In particular, a comparative study is performed with a widely adopted UI observer to emphasize the practical interests of the new estimation solution

ABSTRACT. 动力电池直接提供能量来源，是电动汽车的核心部件，并与电动汽车互相促进，发展迅猛。动力电池组往往由成百上千电池单体串并而成，对其精细管理十分关键且困难。电池管理系统BMS负责完成电池的多状态估计等功能，而传统主从式设计架构，将所有的管理任务都集中在主控器中，使其不堪重负。遂提出一种“分而治之”的智能模组设计方案，不仅为各种先进算法的工程应用奠定了资源条件，而且能够大幅提升电池系统的安全可靠性，并为退役电池系统的快速重组利用提供新途径。

ABSTRACT. 混合动力与机电复合传动技术是目前军用和民用车辆重点发展的驱动形式，涵盖机械、电子、液压、电机、电池、控制等多学科关键技术，具有高功率密度、无级变速、无级转向、大功率供电等典型技术特征，是高科技复杂集成产品，对于我国自主可控核心技术与产业链发展具有重要的牵引作用。以机械功率和电力功率分流传递为主要特点的车用机电复合传动系统可实现未来高机动地面平台或重型车辆的优异的动力性能，为车辆平台搭载的高能武器、电控悬架、空调等功率流型电控系统提供大功率电能输出；同时该系统兼具传动效率高、燃油经济性好、排放低的特点，具有良好的社会、经济效益，是车辆传动发展的重点方向之一。北京理工大学开展了车辆混合动力与机电复合传动系统的基础研究与工程开发工作，并在多型车辆上进行了演示验证或推广应用。

ABSTRACT. 开发先进控制技术是实现内燃机节能减排性能的重要途径之一。作为典型的物理控制系统，汽车内燃机具有结构复杂、多变量耦合、难建模、强非线性和随机性等系统综合复杂特征，对内燃机的控制要求是最大化其燃油效率的同时尽可能减少其有害污染物（氮氧化物NOx,碳氢化合物HC）的排放,因此又是典型的优化控制问题。本报告将从内燃机的系统建模、优化控制器设计、内燃机硬件在环仿真系统的搭建与全工况下的控制算法验证等方面系统介绍内燃机的实时优化控制算法的开发工作。本报告涉及控制设计与实验应用研究，为下一代发动机实时优化控制技术的应用与发展提供了一定的参考。

ABSTRACT. The impact caused by the detonation of landmines and improvised explosive devices may lead to spine fracture and injury of the seated occupants on mine-resistant ambush-protected vehicles. The vibration transmitted from the uneven road surface is another factor affecting ride comfort/health, on the other hand. Aiming at minimizing the injury to spine and “discomfort” due to the shock and vibration from the terrain or blast, a magnetorheological (MR) energy absorber (EA)-based semi-active seat suspension system for both shock and vibration mitigation is proposed and investigated. The proposed MR semi-active seat suspension system simply consists of a coil spring supporting the seat pan and the occupant, a MREA and a fail-safe EA rod. The dynamic model of the MR semi-active seat suspension system with a 4-degree-of-freedom lumped-parameter model for seated occupant is established. A concept of integrated hybrid controller combining strategies for shock and vibration control is proposed and designed. The hybrid controller employs the skyhook control strategy to achieve vibration control and the “soft-landing” control strategy to achieve shock control, and it switches between the two control strategies according to the system dynamic states. Based on the real-time velocity of the seat, the motion process of the “vehicle-seat-human” system can be pre-judged, and the critical point for switching the two control strategies can be determined. A feedforward control strategy based on a hysteresis model with a resistor-capacitor operator is proposed and realized to high-efficiently output desired damping force of the hybrid controller from the employed MREA. Sequentially, both ride comfort (i.e., vibration control) and vertical safety (i.e., shock control) of the MR semi-active seat suspension system are tested, analyzed and evaluated under different excitations.

ABSTRACT. Based on a pure electric logistics vehicle with a single reducer, the paper matches a new electric motor, takes the electricity consumption in NEDC working condition as the optimization objective, designs the two-speed transmission gear ratios to replace the fixed gear ratio and gives a new shift schedule matching the electric motor. Then, the two gear ratios are optimized according to the new shift schedule. The optimization results indicate that the electricity consumption(Kwh/100km) in NEDC working condition decreased by 1.21%. The maximum motor working rotation speed decreased by 44%. The maximum inclination increased by 11.68%. The acceleration time of 0~50km/h is the same as single reducer vehicle, which can meets the requirements. Compare with the single reducer, the two-speed transmission has a better performance.

ABSTRACT. The sliding mode observer is used to estimate the rotor position and angular velocity of the permanent magnet synchronous motor. The controller adopts TMS320F28335 chip as the main chip. After Clark coordinate transformation, the mathematical model of the permanent magnet synchronous motor static coordinate system is established. The SVPWM space vector pulse width modulation technology is used to control the three-phase full-bridge inverter power tube breaking. The speed-current double-closed-loop permanent magnet synchronous motor is operated without position sensor vector control, and the hardware is built to verify the feasibility of the sliding mode observer to estimate the rotor position and speed.

ABSTRACT. This paper proposes two different electric-continuously variable transmission (EVT) configurations for a typical hybrid electric bus (HEB). The two proposed configurations are the Output-split Configuration with Dual Planetary gearsets (OCDP) and the Input-split Configuration with Single Planetary gearset equipped with a Clutch (C-ICSP), respectively. Then the power performance and efficiency of the two configurations are analyzed and compared based on the lever analogy. The analysis indicates the OCDP cannot apply for the HEB because the motor speed exceeds the reasonable range,because the efficiency in high-speed area is low. Only the C-ICSP has good applicability to the HEB.

ABSTRACT. Autonomous vehicle motion control system includes two aspects: longitudinal motion control and lateral motion control. This paper studies the lateral motion control strategy and algorithm of the autonomous driving electric vehicles. By analyzing the two indexes of sliding stability and rollover stability, a safety evaluation model of curve driving is established. Based on the above model, the functional relationship between the safe driving speed and curve radius is established by vehicle dynamics software simulation, and the strategy of upper steering control is designed.The control algorithm of the lower layer of lateral movement matching with the upper layer control strategy is studied, which realizes the optimal control of lateral movement deviation. Finally, the real vehicle experiment verifies that the lateral motion control strategy and algorithm designed in this paper can control the steady-state error of lateral displacement within 1.14%, and it has obviously better performance than the traditional Fuzzy controller under straight road and large curvature curve with good real-time, stability and robustness.

ABSTRACT. Power battery is the main power source of plug-in hybrid electric vehicles (PHEV), and the temperature change directly affects its safety and service life. Firstly, the temperature effect on the characteristics of batteries is analyzed, which reveals that the internal resistance of batteries varies greatly at different temperatures. On the basis, the thermal model of batteries is constructed. Furthermore, the energy management strategy based on model predictive control (MPC) is designed. BP neural network is used to predict the velocity, and dynamic programming is used as receding horizon optimization. After that, three methods are used to restrict the temperature on the basis of MPC. Finally, the engine torque is obtained considering both fuel economy and battery temperature.

ABSTRACT. In view of the problem of the road transport safety of hazardous chemicals. This paper constructs the evaluation index system of road transport safety of dangerous goods by analyzing the impact of liquid sloshing on the transport safety and driving stability. People, vehicles, roads, environment and other factors that affect the safety of the transport vehicles are also considered. In this paper, the analytic hierarchy process (AHP) is used to give the weight of each evaluation facto and seven kinds of evaluation set of safety status are set up. This paper takes the vehicle handling stability as the object, and analyzes the influence of liquid sloshing on the handling stability of vehicle. The simulation model is constructed by the software of Trucksim, and the fishhook test is selected to carry out the simulation test. The simulation results show that forces and moments generated by the liquid sloshing significantly reduce the vehicle stability. The research method provides reference for practice and decision-making in road transport of dangerous goods safety evaluation.

ABSTRACT. Abstract—Motion control is one of the key technologies in the automatic parking system. The vehicle real-time positioning information plays a fundamental role in motion control. The automatic parking scene is essentially an low-speed unmanned scene in a small area.This paper mainly studies the vehicle positioning technology in this scene, namely the Closed- loop update of parking spot. Considering the high price of equipment used in general vehicle positioning technology, this paper selects the vehicle positioning technology with simple and low cost equipment. Firstly, the positioning accuracy of the traditional track reckoning is confirmed, and the simulation with carsim & simulink is used to verify the track reckoning ability. Secondly, the method of track reckoning with speed is proposed to locate the moving vehicle in real time. Finally, due to the inaccurate wheel speed of the vehicle at low speed in the real vehicle verification section,an improved method of track reckoning used multi-information is proposed. The method combines the advantages of the traditional and wheel speed methods of track reckoning to complete the track reckoning better. The software simulations and the real vehicle tsets about parallel parking trajectory show that the track reckoning with multi-information fusion is superior in the Closed- loop update of parking spot.

ABSTRACT. Many vehicle state parameters will be used in the electronic stability control (ESC) system. ESC system must estimate some important parameters based on the limited sensors and help to develop the control strategies. ESC can estimate the work state of vehicle with tire forces based on the estimating of adhesion coefficient and wheel slip ratio, then determine the control strategy on wheels. The tire forces to be estimated include vertical force, lateral force and longitudinal force on tires. This paper proposes an estimation method for tire/road forces. The method only uses standard sensors that are available on the vehicle. The vertical tire force is estimated based on vehicle geometric parameters and vehicle working state, the lateral force is estimated based on vehicle yaw rate, and the longitudinal force is estimated based on the longitudinal acceleration. Using a Matlab-Carsim co-simulation modal,the estimation process was applied and compared to Carsim simulation results. And the HIL tests also were carried out. Both simulation and experimental results show the accuracy and potential of the estimation method.

ABSTRACT. In addition to some test standards in the level 1 automated vehicles, it still lacks perfect test evaluation procedures for level 2 automated vehicles. The evaluation method of vehicle field test for L2 automated vehicles is studied, and the multi-level automated vehicles evaluation index system is preliminarily established from the aspects of safety, intelligence and experiential. The order relationship and the analytic hierarchy process are applied to empower the automated vehicles evaluation indicators at all levels. A comprehensive evaluation model of L2 automated vehicles was established by using fuzzy comprehensive and grey comprehensive evaluation method. Taking the test results of the three models of vehicles on ACC mode as an example, a multi-level fuzzy comprehensive evaluation and a gray comprehensive evaluation were carried out to conduct a single evaluation and comprehensive evaluation on the three aspects of safety, intelligence and experiential of the automated vehicles.

ABSTRACT. The robustness of vehicle path tracking is studied for high-speed emergency avoidance conditions. Using a two-degree-of-freedom vehicle model and aiming at the ideal yaw rate for controlling the actual yaw rate tracking planning of the vehicle, a non-nonlinear auto-disturbance rejection path tracking controller is designed. The non-nonlinear ADRC controller can observe and compensate the external disturbances and model parameter uncertainties such as vehicle mass changes to ensure the robustness of the system. Aiming at the problem of excessive lateral acceleration or discontinuity of step and curvature in the avoidance path, Sigmoid function is introduced to re-plan the avoidance path. In order to verify the robustness of the controller, an SUV provided by Carsim is used for real vehicle simulation. The simulation results show that the nonlinear auto-disturbance rejection path tracking controller can control the vehicle to track the ideal avoidance path and ensure the vehicle to avoid the obstacles ahead quickly and without collision.

ABSTRACT. This paper presents a method for solving parking trajectory planning problems by combining homotopic method and GPM(Gauss pseudo-spectral method). Firstly, the vehicle kinematics model is established. By combining the dynamic constraints, avoid collision constraints and terminal conditions, the parking trajectory planning problem is described as the optimal control problem with the cost function of the shortest parking time and travel distance. Then, the optimal control problem is transformed into NLP by GPM. In the process of iteration, the constraint conditions of obstacle avoidance are transformed by the homotopic method, which uses the solution of the last optimal control problem as the guess of the next solution. After being simulated under vertical parking scenario, the results show a good convergence rate for parking trajectory planning of narrow parking spaces.

ABSTRACT. A hardware-in-the-loop test system for the Advanced Driver Assistance System controller based on virtual scenario was built. Taking the automatic emergency braking system as an example, a standardized test scenario was built based on CarMaker software and the Advanced Driver Assistance System function simulation test was carried out. The test results show that by configuring the vehicle model and establishing a standardized scenario, the hardware-in-the-loop test system of the Advanced Driver Assistance System controller based on the virtual scenario can simulate the real vehicle test which can test the complex dangerous conditions and improve the efficiency safety and repeatability of the test.

ABSTRACT. With the advantages of high power density, efficiency and reliability, the single-stage AC-DC converter has become increasingly popular for use in battery charger application. To improve the level of power, this paper proposes a novel three-phase single-stage isolated symmetrical AC-DC (TPSSI-ACDC) converter and corresponding control scheme. A compact topology with only six semiconductor devices supplies high frequency pulse voltage for transformer, meanwhile the PFC function comes true. The proportional resonant (PR) controller is adopted to track the fundamental-frequency current references with zero-error quickly. The effectiveness and the performance of the proposed converter are verified by simulation.

ABSTRACT. 主要研究了动力锂离子电池组异常电池筛选及替换机制，通过其健康状态、荷电状态、开路电压、充电瞬时升压、放电瞬时降压等多因素特征参数进行初步分类，同时利用该多参数中心特征向量进行比对，初步筛选动力电池组中的异常电池；通过电池充放电的电压曲线等指标通过聚类分析及中心特征向量比对算法进行进一步分类筛选，挑选出异常电池集合。根据剔除异常后电池集合挑选出满足中心特征向量要求的电池进行替换，进而提高成组电池效率，增加使用寿命。

ABSTRACT. This paper proposes a predictive ecological control (PEC) strategy to improve the energy efficiency of electric vehicle (EV) which approaches to an intersection with traffic light. By using the traffic signal information, the range of average cruising speed which avoids stopping is determined. Within the vehicle speed range, the cruising velocity is optimized by dynamic programming to reduce the energy loss while driving on highway with varying slops. Three case studies, cruising without information, with signal information and PEC are simulated. The results indicate the significant energy efficiency improvement of PEC and its potential impact on trip time.

ABSTRACT. Changes in the structure and operating conditions of hybrid vehicles have caused new vibration and noise problems. The hybrid vehicle will start or stop the engine selectively under different road conditions for the purpose of energy saving and emission reduction. The vibration of vehicles during engine start and stop process is not expected by the drivers, whose influences for ride comfort can’t be ignored. This paper studies the excitation source of the engine, analyzes the pumping resistance, inertial resistance and combustion torque of the engine, and establishes an exhaustive engine model based on the analysis results in ADAMS. Finally, co-simulation results show that using the coordinated torque control of the large and small motors, the vibration of the whole vehicle during the start and stop of the engine is reduced significantly.

ABSTRACT. This paper proposes a deep reinforcement learning (DRL) approach for a predictive overtaking strategy for autonomous vehicles in freeway scenarios. First, the real-world driving data is extracted from the Next Generation SIMulation (NGSIM) dataset. The long short-term memory (LSTM) model is leveraged to forecast the longitudinal and lateral motion of vehicles. Then, the freeway overtaking scenario is constructed, wherein two-lane road traffic is considered. Based on the predicted driving trajectories of the surrounding vehicles, reinforcement learning is utilized to guide the target vehicle's movement. Results indicate that the presented decision-making strategy could enhance the mobility and safety of the studied automated vehicle.

ABSTRACT. In view of the problem of insufficient cruising range of electric vehicles, the supercapacitor and DC/DC converter are connected in parallel to form an fully-active vehicle-mounted hybrid energy storage system, in which the battery is used as the main power source, and the supercapacitor is used as the auxiliary power source, and they are connected to the bus through a bidirectional DC/DC converter. A power distribution control strategy for fully-active hybrid energy storage system based on sliding mode control is proposed. The proposed control strategy includes a sliding mode controller to accurately track the reference values of battery and supercapacitor current, a voltage controller is used to maintain bus voltage stability, and the stability analysis is based on the Lyapunov method. The fully-active hybrid energy storage system and the proposed control strategy are modeled and simulated in the simulation environment of Matlab/Simulink. The simulation results show that the proposed sliding mode control strategy can accurately track the reference value of the battery and supercapacitor current, and stabilize the bus voltage.The effectiveness of the proposed control strategy is fully proved.

ABSTRACT. 随着汽车电动化、智能化的发展，电子液压制动(Electronic Hydraulic Brake, EHB)系统被越来越多的应用。而如何在保证人车安全性的基础上，给驾驶员带来更好的制动体验，是其商业化应用中的关键一环。在此背景下，本文提出一种基于驾驶员制动意图识别的EHB助力系统控制策略。首先针对非解耦式EHB助力系统建模分析，其次根据制动主缸压力及压力变化率采用模糊推理对驾驶员制动意图进行识别，将制动意图分为常规减速和紧急制动两类。基于上述两种驾驶员意图，分别设计了助力电流控制和主动压力控制两种控制策略。最后基于配备此EHB系统的四轮分布式驱动实验车进行实车验证，分别验证压力和助力电流控制效果以及控制策略的有效性。实验结果表明，EHB系统处于压力主动控制模式时，在阶跃目标压力输入下，制动主缸压力能够迅速跟随目标压力值；EHB系统处于助力电流控制模式时，助力电流能够较好的跟随目标电流，踩下制动踏板无抖振现象；对比普通助力模式和基于本文所设计的控制策略在迅速踩下制动踏板时的压力曲线，后者能够更快的完成制动主缸建压过程。

ABSTRACT. 装备差动助力转向（DDAS）系统的汽车匹配主动前轮转向（AFS）系统时，两个系统会相互干扰。DDAS系统助力时附带引入的横摆力偶矩会影响车辆的转向特性；反过来，AFS系统的介入会导致转向盘转矩突变。为此，本文针对一款电动轮前轴独立驱动汽车提出了一种协调处理上述问题的差动协同主动转向（DDCAS）控制系统。本文首先介绍了差动协同主动转向系统的结构，并建立了其动力学模型。然后分别制定了DDAS和AFS的控制策略，接着在分析两个控制系统干扰作用机理的基础上，制定了差动协同主动转向控制策略。在MATLAB/Simulink与CarSim联合仿真平台上进行了双移线工况和AFS系统耦合电机转角阶跃输入工况的仿真分析，验证了控制策略的有效性。

ABSTRACT. The configuration of single-motor direct drive or single-motor coupled with a multi-speed transmission adopted by most domestic battery electric buses has problems such as low overall efficiency and power interruption during shifting. In this paper, a dual-motor coupled propulsion system is proposed for high-performance battery electric buses. Based on the principle of minimum demand electric power, the threshold value of switching between different drive modes of electric drive system is established. Based on the dynamic programming method, the torque distribution control strategy of the electric drive system is optimized in certain drive cycle. Compared with that used the original equal proportion control strategy, the overall efficiency of the electric drive system is improved, and the power consumption is reduced by 13.6%.

ABSTRACT. In order to solve the problem of straight-line running deviation of electric bulldozer caused by the difference of response characteristics between two motors, the difference of moving devices on both sides or the difference of load on both sides, the straight driving stability of tracked electric bulldozer was studied. Based on Matlab/Simulink, the simulation model of tracked electric bulldozer was established, and the influence factors of straight driving stability were analyzed by using the model. According to the analysis results, the yaw angular velocity of bulldozer was chosen as the feedback parameter of straight-line stability control, and the control strategy based on fuzzy PID was designed. According to the yaw angular velocity and its change rate, the correction of PI parameters was obtained by fuzzy controller. A comparative simulation analysis of two cases with or without straight driving stability control was carried out, and the simulation results showed that the control strategy can achieve good results when there were differences in the torque response of the motor on both sides, the crawler on both sides and the load on both sides, which verifies the feasibility of strategy.

ABSTRACT. Clutch control is a critical issue for Automated Manual Transmission (AMT) and it is related to the vehicle drivability closely, such as the jerk during gear shift and the friction loss during the clutch slip process. In this paper, a clutch control strategy is developed during gear shift process, wherein two key problems are considered. A shaft torque observer-based clutch disengagement controller is proposed during clutch release process, and a disturbance rejection controller is applied to clutch engagement control. The plant and control-oriented models of the powertrain are built, then the gear upshift process from the first gear to second gear is taken as an example to validate the proposed clutch control strategy. The simulation results show that, the drivability during the gear shift process can be guaranteed, the shift time is short enough, whilst the clutch slip loss and vehicle jerk also satisfy the requirements.

ABSTRACT. This paper presents an algorithm for estimating road surface adhesion coefficient at different slip rates. First, estimation method of slip ratio and utilization adhesion coefficient are proposed.An estimation method for road surface adhesion coefficient is proposed for different slip ratio stages. At high slip ratio, estimates are made by utilization adhesion coefficient. At low slip rate,estimates are made by Bayes' theorem and slope method. Finally, through experimental verification, the algorithm can accurately and quickly estimate the adhesion coefficient of the current road surface.