Flow-field guided steering control for rigid autonomous ground vehicles in low-speed manoeuvring

ABSTRACT

This paper studies the low-speed manoeuvring problem for autono-mous ground vehicles operating in complex static environments. Making use of the intrinsic property of a fluid to naturally find its way to an outflow destination, a novel guidance method is proposed. In this approach, a reference flow field is calculated numerically through Computational Fluid Dynamics, based on which both the reference path topology and the steering reference to achieve the path are derived in a single process. Steering control considers three constraints: obstacle and boundary avoidance, rigidity of the vehicle, plus the non-holonomic velocity constraints due to the steering system. The influences of the parameters used during the flow field simulation and the control algorithm are discussed through numerical cases. A divergency field is defined to evaluate the quality of the flow field in guiding the vehicle. This is used to identify any problematic branching features of the flow, and control is adapted in the neighbourhood of such branching features to resolve possible ambiguities in the control reference. Results demonstrate the effectiveness of the method in finding smooth and feasible motion paths, even in complex environments.     

人机混驾环境下基于LSTM的无人驾驶车辆换道行为模型

道路系统中的人机混驾交通环境是指人工驾驶车辆与自动驾驶车辆混合运行的交通环境，其中换道行为建模是人机混驾环境下无人驾驶车辆行为研究的热点。基于深度学习理论，构建人机混驾环境下基于长短期记忆神经网络的无人驾驶车辆换道行为模型（Long-short-term-memory-based Autonomous Vehicles Lane Changing，LSTM-LC）。通过研究人工驾驶车辆在换道过程中与周边车辆的相互作用，对换道行为影响因素进行分析；同时，为了提升模型的迁移性，引入道路横向偏移量信息。结合LSTM神经网络的输入要求，使用美国公开交通数据集Next Generation SIMulation（NGSIM）构建换道行为样本库。针对LSTM-LC模型，以均方差MSE作为损失函数，使用RMSprop优化方法进行训练，对LSTM网络结构、历史序列长度N及训练样本量3个重要参数进行标定。最后，针对道路横向偏移量M对LSTM-LC模型性能的影响进行对比试验。研究结果表明：相比GRU-LC模型，LSTM-LC模型对换道行为的表征更准确，在模型的精度和迁移性上有着显著的提升；GRU-LC模型的均方差为4.64 m²，迁移性均方差为119.82 m²，而LSTM-LC模型的均方差为3.18 m²，迁移性均方差为79.58 m²，分别优化了31.5%和39.71%；通过引入道路横向偏移量M，可将LSTM-LC模型精度和迁移性提升约10%，且模型稳定性更强。     

路网环境下考虑大型车混入率的事故疏散诱导模型

大型车的混入对高速公路交通流产生了较大的影响，尤其是在交通事故情景下。为了引导事故条件下驾驶人和组织者做出高效准确的决断，将考虑了大型车混入率的动态空间占有率模型引入到交通波模型，构建干涉与非干涉情景下的交通事故影响模型。以郑尧高速为例，对模型的准确性和可行性进行了验证，分别对干涉情景下的疏散时间、疏散量以及事故发生的位置，车辆数等指标与事故影响程度的指标（包含事故最远排队长度，事故持续时间）关系进行分析。研究结果表明：疏散时间与事故影响程度成正相关关系，疏散量与事故影响程度成负相关关系，而事故发生点与上游匝道之间的距离与其关系不大；道路服务水平为0.456，车辆数为1 321 veh·h^-1时，为了使得分合流区不受影响，在不采取任何措施的情景下，应将大型车混入率控制在50.1%以下，使得最远排队长度在10 km内；当大型车混入率大于58%时，将很难通过干涉引导避免对上游分合流区产生影响；在35 min以内采取干涉措施的效果最为明显，而大于35 min时，事故持续时间会发生一个急剧的增加，不利于路网恢复，之后事故恢复时间将趋于平稳；对道路交通量进行模拟可知交通量每增加50 veh，疏散时间和距离增加的范围为[1.5 min，3.6 min]和[1.209 km，1.543 km]。研究结果可为高速公路事故诱导策略制定和疏散效果提升提供参考。     

道路交通事故痕迹鉴定确定肇事车辆

道路交通事故检验鉴定为公安机关交通管理部门的事故处理提供科学、专业的技术支持。检查和核实道路交通事故的影响是对道路交通事故技术评估的基本方法。已成为解决许多事故问题的重要依据。本文根据道路交通事故痕迹鉴定的实际情况,针对公安机关交通管理部门涉及复杂、疑难、特殊技术问题的痕迹鉴定情况,提出了针对性的解决方案。     

A design methodology for mechatronic vehicles: application of multidisciplinary optimization,multibody dynamics and genetic algorithms

A design methodology for mechatronic vehicles is presented. With multidisciplinary optimization (MDO) methods, strongly coupled mechanical, control and other subsystems are integrated as a synergistic vehicle system. With genetic algorithms (GAs) at the system level, the mechanical, control and other relevant parameters can be optimized simultaneously. To demonstrate the feasibility and efficacy of the proposed design methodology for mechatronic vehicles, it is used to resolve the conflicting requirements for ride comfort, suspension working spaces and unsprung mass dynamic loads in the optimization of half-vehicle models with active suspensions. Both deterministic and random road excitations, both rigid and flexible vehicle bodies and both perfect measurement of full state variables and estimated limited state variables are considered. Numerical results show that the optimized vehicle systems based on the methodology have better overall performance than those using the linear quadratic Gaussian (LQG) controller. It is shown that the methodology is suitable for complex design optimization problems where: (1) there is interaction between different disciplines or subsystems; (2) there are multiple design criteria; (3) there are multiple local optima; (4) there is no need for sensitivity analysis for the optimizer at the system level; and (5) there are multiple design variables.     

A closed-loop dynamic simulation-based design method for articulated heavy vehicles with active trailer steering systems

This paper presents a closed-loop dynamic simulation-based design method for articulated heavy vehicles (AHVs) with active trailer steering (ATS) systems. AHVs have poor manoeuvrability at low speeds and exhibit low lateral stability at high speeds. From the design point of view, there exists a trade-off relationship between AHVs’ manoeuvrability and stability. For example, fewer articulation points and longer wheelbases will improve high-speed lateral stability, but they will degrade low-speed manoeuvrability. To tackle this conflicting design problem, a systematic method is proposed for the design of AHVs with ATS systems. In order to evaluate vehicle performance measures under a well-defined testing manoeuvre, a driver model is introduced and it ‘drivers’ the vehicle model to follow a prescribed route at a given speed. Considering the interactions between the mechanical trailer and the ATS system, the proposed design method simultaneously optimises the active design variables of the controllers and passive design variables of the trailer in a single design loop (SDL). Through the design optimisation of an ATS system for an AHV with a truck and a drawbar trailer combination, this SDL method is compared against a published two design loop method. The benchmark investigation shows that the former can determine better trade-off design solutions than those derived by the latter. This SDL method provides an effective approach to automatically implement the design synthesis of AHVs with ATS systems.     

地铁车辆牵引拉杆装置稳定性问题研究

受压牵引拉杆在极限承载下因忽然弯曲导致结构失稳破坏,从而影响转向架系统的运行安全性。文章从理论与工程相结合的角度剖析了杆件结构失稳的力学原理,提出了改善杆件结构稳定性的基本思路和可行措施,并在此基础上,对某型牵引拉杆装置进行优化设计。分析与实验表明,在工程可能的极限载荷下,牵引拉杆装置可有效确保稳定性,并利用有限元技术成功预测出牵引拉杆装置的失稳临界载荷。这为设计分析类似牵引拉杆装置提供了一种的思路。     

新能源汽车车载终端下线检测系统设计

在车辆的生产过程中,除了需要确保车辆装配正常,还需要确保各个零件功能正常才能确保车辆正常下线。传统的车辆下线检测系统,需要人为的对车辆的识别码、车载终端的各路参数进行人工记录和比对,再将参数录入后台系统,耗费人力较多,检测耗时也较长,车辆下线检测的效率也较低。同时,通常的下线检测过程中,未对车载设备的正常运行进行判断,如果将问题车载终端装车下发,会导致监控平台无法对该车辆进行正常监控,存在安全隐患,也不满足相关要求,同时增加售后维护成本。基于此,本文设计开发了一种新能源汽车车载终端下线检测系统,实现了新能源车辆在出厂时,通过检测时,由检测人员操作该系统来查看终端工作状态,系统自动根据终端返回平台信息判断该终端工作是否正常,然后把结果反馈至生产系统。     

考虑前后多车的网联自动驾驶车辆混流跟驰模型

随着中国新基建战略的提出及自动驾驶和网联通信技术的不断发展，网联自动驾驶车辆（CAV）、自动驾驶车辆（AV）和常规人驾车辆混行的交通流将在未来长时间存在。建立适用于网联自动驾驶车辆、自动驾驶车辆和常规人驾车辆3种类型车辆的混流跟驰模型，考虑多前后车车头间距、多前车速度差、加速度差、与主体车辆的相对距离等因素，并进行典型场景的数值仿真。刹车和起步过程的3种混流数值仿真结果显示，模型在几种典型混行场景下均具有可行性，车辆的加速度和速度变化更为平缓。不同CAV比例下的数值仿真结果显示，车队中CAV比例越高，车队整体恢复至平稳状态的时间越短，波动幅度越小。CAV均质流数值仿真结果表明，与MHVAD模型相比，该模型不稳定区域减小33.8%，所控制的车队速度波动幅度减小14%。CAV与AV混流的数值仿真结果显示，与PATH实验室模型相比，由该模型控制的车队加速度进入相对稳定状态提前5.5 s。该模型可用于不同车辆均质流及3种车辆混行的队列控制，在目前开展混行实车试验困难的情况下，也可应用该模型进行混行跟驰仿真，从而为混行交通流的道路交通管理及基础设施布设提供理论依据和模型支持。