汽车语音交互模型及设计策略

目的 在智能技术飞速发展、用户需求不断增长的背景下,以用户体验为中心,探索汽车工业“新四化”趋势下的汽车语音交互新形态,重构汽车语音交互模型,针对性地提出交互设计策略。方法 分析车载语音系统与用户之间的信息交流过程,构建汽车语音界面架构和交互模型。通过“新四化”趋势下汽车语音交互在信息主体、信息范围和信息内容等方面的变化综述,运用信息处理流程理论开展汽车语音交互模型和设计策略研究。结果 根据车载语音系统的信息处理流程,提出了以功能层、内容层、形式层、情感层为信息架构的汽车语音界面,构建了基于信息交流的汽车语音交互模型。基于语音界面的信息架构理论,提出了未来的汽车语音交互在功能拓展、内容丰富、形式优化和情感满足等方面的设计策略。结论 从信息交流角度开展的汽车语音交互设计研究,对车载语音系统交互设计和体验优化具有参考意义。     

基于临界带功率谱方差的端点检测

端点检测作为语音信号处理的关键技术,其准确性直接影响到语音识别系统的计算复杂度和识别能力。在人耳听觉特性理论研究的基础上,利用语音段和背景噪声段临界带功率谱上的差异,提出了一种基于临界带功率谱方差的端点检测方法。通过自适应门限值的选取,该方法对背景噪声具有良好的跟踪性能。在不同的信噪比条件下,进行了端点检测实验。结果表明：该方法与传统的短时能量和短时平均过零率方法、谱熵方法相比,可以有效降低背景噪声的影响,具有更好的鲁棒性和正确率。     

Driver performance while text messaging using handheld and in-vehicle systems

This study presents an evaluation of driver performance while text messaging via handheld mobile phones and an in-vehicle texting system. Participants sent and received text messages while driving with an experimenter on a closed-road course, using their personal mobile phones and the vehicle's system. The test vehicle was an instrumented 2010 Mercury Mariner equipped with an OEM in-vehicle system that supports text messaging and voice control of mobile devices via Bluetooth, which was modified to allow text message sending during driving. Twenty participants were tested, 11 younger (19–34) and 9 older (39–51). All participants were regular users of the in-vehicle system, although none had experience with the texting functions.Results indicated that handheld text message sending and receiving resulted in higher mental demand, more frequent and longer glances away from the roadway, and degraded steering measures compared to baseline. Using the in-vehicle system to send messages showed less performance degradation, but still had more task-related interior glance time and higher mental demand than baseline; using the system's text-to-speech functionality for incoming messages showed no differences from baseline. These findings suggest that using handheld phones to send and receive text messages may interfere with drivers’ visual and steering behaviors; the in-vehicle system showed improvement, but performance was not at baseline levels during message sending.     

选择性声强技术的研究及其在噪声控制中的应用

传统的声强技术容易受到强大背景噪声的干扰，针对此问题本文提出一种新的噪声测试技术--选择性声强技术，并利用自行研制的噪声自动分析系统对其进行了研究，包括原理、算法及测试设备。研究结果表明：选择性声强技术可以将目标噪声源的辐射噪声从强大的背景噪声中分离出来，测量结果有效地反映了目标噪声源的声场分布状况，为目标噪声源的噪声辐射特性研究及噪声控制提供参考依据。     

Evaluation of virtual reality interface for product shape designs

It is generally acknowledged that existing computer-aided design systems have inefficient user interfaces. Especially during the concept shape design stage, these systems prove to be cumbersome because of two reasons: (i) they require the usage of two-dimensional (2D) input devices, while the designs are typically three-dimensional (3D); and, (ii) CAD systems require the specification of dimensions, which may not be precisely known at the concept stage. To overcome these limitations, this research proposes the use of virtual reality (VR) devices to provide a physically intuitive interface for concept shape creation. The intuitiveness of the interface arises from the use of natural hand gestures and voice commands that emulate the way in which designers discuss concept shapes. In this scenario, the interface between the human and computer plays a central role with respect to usability, usefulness and accuracy. The focus of this research is on using two modalities: (i) hand input; and, (ii) voice driven commands, or a combination of these modalities to accomplish typical CAD tasks. Based on experience with a conventional CAD system, a set of typical CAD tasks are identified. A series of tests are then performed to determine the relative efficiency of the different modality combinations to achieve each task. The interface test results indicate that while voice commands are intuitive in initiating operations such as viewpoint zooming in/out and object creation/deletion, hand inputs are effective in performing spatial tasks such as interactive dimensioning and re-location of shapes. It was also found that a combination of voice and hand input can be used for accomplishing certain tasks more effectively such as, zooming in/out a particular direction (hand orientation indicates direction and voice is used for indicating zoom in/out operation). Based on the experience with the prototypical system developed it is concluded that voice and hand input are effective ways of building three-dimensional shapes in a virtual reality environment. To verify the efficiency of the VR-CAD interface, sample injection molded parts are built on the current VR-based CAD system and a traditional CAD system, and the times taken to build these parts are compared. The test results indicate that building geometry shapes containing canonical forms, such as block, cylinder, sphere, ⃛ etc, using a VR interface results in a speedup of five to ten times over traditional CAD systems.     

Noise abatement and traffic safety: The trade-off of quieter engines and pavements on vehicle detection

Road traffic sounds are a major source of noise pollution in urban areas. But recent developments such as low noise pavements and hybrid/electric engine vehicles cast an optimistic outlook over such an environmental problem. However, it can be argued that engine, tire, and road noise could be relevant sources of information to avoid road traffic conflicts and accidents. In this paper, we analyze the potential trade-offs of traffic-noise abatement approaches in an experimental study, focusing for the first time on the impact and interaction of relevant factors such as pavement type, vehicle type, listener's age, and background noise, on vehicle detection levels. Results reveal that vehicle and pavement type significantly affect vehicle detection. Age is a significant factor, as both younger and older people exhibit lower detection levels of incoming vehicles. Low noise pavements combined with all-electric and hybrid vehicles might pose a severe threat to the safety of vulnerable road users. All factors interact simultaneously, and vehicle detection is best predicted by the loudness signal-to-noise ratio.     

公路噪声源强预测模型研究进展

随着公路建设和城镇化的不断发展，车辆行驶过程中产生的环境噪声对沿线居民、学校的影响也日益明显。针对目前我国的标准规范中未明确给出公路噪声源强预测模型的问题，对国外公路环境噪声源强模型进行系统的分析和总结，并对我国公路噪声源强预测模型的研究提出建议。     

微光成象系统中实时降噪的优化设计

微光电视工作于极低照度或恶劣环境下，输出信噪比明显降低，目标完全淹没在背景噪声中，使探测率大大下降。提出了用帧间处理技术消除加性噪声。论述了字长和滤波系数的选择原则。研究了实时去噪的几项关键技术，算法和符号位的设计，各部分工作时间分配以及大背景噪声下时种同步方法。     

Cooperation schemes for heterogeneous material-handling vehicle fleets

This study deals with an organization which utilizes wire-guided vehicles as its material-handling system and undergoes an expansion, as a result additional material-handling devices are required. In the event, choosing similar vehicles to those currently operating in the manufacturing facility is not a viable alternative, the organization can still decide to maintain a homogeneous vehicle fleet by replacing all vehicles with advanced more expensive free-ranging vehicles. On the other hand, the organization can decide to reuse the old wire-guided vehicles while purchasing only a few new free-ranging vehicles, creating a heterogeneous vehicle fleet. This study presents two different cooperation schemes to control a heterogeneous vehicle fleet, which achieve similar operational performances as a homogeneous fleet of free-ranging vehicles. The study uses simulation to develop control strategies and to determine the system's parameters. Three different scenarios are used to test these control strategies. Finally, an operational and economical comparison analysis between the different cooperation schemes is performed.     

地铁车辆空调系统噪声分布规律

对地铁车辆在静止状态和不同速度下的空调系统开关工况进行噪声测试,对其空调系统噪声特性进行频谱分析,得到其分布规律：静止时空调系统的噪声声压级分布主要集中在160 Hz～2 000 Hz的频率范围内;空调系统噪声对坐与站的乘客的影响约1.5 dB(A)～2 dB(A)左右;车辆低速运行时空调噪声影响比较明显,且速度越低影响越大。该研究结果对地铁车辆减振降噪设计有一定的参考价值。     

Influence of texturing method on tyre/road noise of cement concrete pavement

The texture of pavements influences the tyre/road noise. This refers especially to cement concrete pavements. This article presents the results of studies on sound level on concrete pavements in Poland, which were textured using burlap drag, broom drag and exposed aggregate methods. It was established that on exposed aggregate cement concrete, the maximum aggregate size, which is a key factor in the surface’s macrotexture, has a significant influence on tyre/road noise. It should not exceed 8 mm in the upper layer. When using aggregate size 11 mm, the maximum sound level of a passing-by car increases by approximately 2 dB. Pavements textured using the broom drag and burlap drag methods turned out to be not a beneficial solution due to the sound levels emitted by heavy vehicles moving at a speed above 80 kph. Pavement texturing using the transverse broom drag emits sound levels from heavy vehicles which are approximately 3 dB higher in comparison to values noted on SMA11. The methods of pavement texturing using diamond grinding, used in some countries, point towards the possibility of obtaining beneficial acoustic properties.     

水下单边带语音多通道增强技术

船舶噪声是影响水下语音通信质量的主要因素。为了提高单边带语音信号的信噪比和语音质量,采用多通道自适应增强算法对语音信号进行研究。该算法运用自适应信号增强技术,具有运算量小、易实时实现、消噪效果显著等优点,并且能够有效利用不同换能器接收到的信息。从理论上对该算法进行了分析,并结合海试数据进行验证。结果表明,低信噪比、非平稳噪声环境下,该算法有效抑制了接收信号中的尖锐噪声,并且当系统收敛时,使其趋于白化,明显提高了系统的输出信噪比,显著改善了语音质量,在性能上明显优于传统的信号相干叠加算法。     

Deep Learning-Based Intrusion System for Vehicular Ad Hoc Networks

The increasing use of the Internet with vehicles has made travel more convenient. However, hackers can attack intelligent vehicles through various technical loopholes, resulting in a range of security issues. Due to these security issues, the safety protection technology of the in-vehicle system has become a focus of research. Using the advanced autoencoder network and recurrent neural network in deep learning, we investigated the intrusion detection system based on the in-vehicle system. We combined two algorithms to realize the efficient learning of the vehicle’s boundary behavior and the detection of intrusive behavior. In order to verify the accuracy and efficiency of the proposed model, it was evaluated using real vehicle data. The experimental results show that the combination of the two technologies can effectively and accurately identify abnormal boundary behavior. The parameters of the model are self-iteratively updated using the time-based back propagation algorithm. We verified that the model proposed in this study can reach a nearly 96% accurate detection rate.     

基于声源识别及声场模拟的工厂通风系统降噪研究

工厂厂房为了实现进气、排气及集尘的目的,需布设通风系统。由于其分布范围广泛,使得通风系统的噪声成为工厂厂房室内噪声中除机电设备之外的另一主要噪声源,对厂房内员工的作业声环境造成很大的冲击。通风系统的噪声源主要来自风管的固定支架振动、风管进排气噪声、管道壁振动以及管道噪声通过管壁的透射。文章建立了一套针对工厂通风系统噪声的声源与声场测量、频谱分析、诊断排序与降噪模拟评估技术,利用声场模拟,给出三种降噪方案。应用此技术对工厂噪声进行治理,可大大提高业主选择降噪方案的信心及成效预期心理。     

基于自适应波束形成和ICA的消噪系统

介绍了自适应宽带波束形成和快速独立分量分析的基本理论和算法,分析了固定波束形成的理论局限性及自适应宽带波束形成的优越性。提出了一种基于自适应波束形成和独立分量分析的消噪系统,算法显著地抑制了噪声、增强了语音,又具有稳定快速的性能。同时分析了自适应波束形成单元数或者输入信号数对算法性能的影响,对实际应用具有指导意义。     

人车混行城市隧道交通噪声对行人影响的实验研究

为研究人车混行城市隧道内机动车噪声对行人的影响，文章在洞山隧道内部、口部、外部等间距布置8个测点，利用RTA840双通道实时分析仪测试了20min内的交通噪声等效声压级、频谱和车辆从行人旁边经过时3s内的交通噪声瞬时声压级、频谱。测得：(1) 隧道内部的交通噪声等效声压级为82dB(A)，比隧道外等效声压级高10dB(A)。(2) 隧道外部、口部、内部的交通噪声频谱曲线特征相同，在20~2 000Hz范围内声压级较高且呈“M”形。(3) 当车辆在隧道内部从行人旁通过时，交通噪声瞬时声压级达到86dB(A)；瞬时噪声频谱相比于背景噪声频谱在20~8 000Hz频段范围内声压级均有明显增大。最后结合测试结果和洞山隧道实际从吸声、隔声、管理三方面总结隧道内声环境改善措施。     

铁路客车车内噪声分布规律研究

运用多通道噪声测试与分析系统,对运行中的铁路客车进行了多点同步车内噪声测试和分析,得出了在不同工况和不同运行速度时的车内噪声的现状和分布规律,确定出车内噪声发生的主频带,对既有线铁道车辆和将来的高速铁道车辆进行防噪降噪设计具有较高的参考价值。     

A Highly Efficient Algorithm for Phased-Array mmWave Massive MIMO Beamforming

With the rapid development of the mobile internet and the internet of things (IoT), the fifth generation (5G) mobile communication system is seeing explosive growth in data traffic. In addition, low-frequency spectrum resources are becoming increasingly scarce and there is now an urgent need to switch to higher frequency bands. Millimeter wave (mmWave) technology has several outstanding features—it is one of the most well-known 5G technologies and has the capacity to fulfil many of the requirements of future wireless networks. Importantly, it has an abundant resource spectrum, which can significantly increase the communication rate of a mobile communication system. As such, it is now considered a key technology for future mobile communications. MmWave communication technology also has a more open network architecture; it can deliver varied services and be applied in many scenarios. By contrast, traditional, all-digital precoding systems have the drawbacks of high computational complexity and higher power consumption. This paper examines the implementation of a new hybrid precoding system that significantly reduces both calculational complexity and energy consumption. The primary idea is to generate several sub-channels with equal gain by dividing the channel by the geometric mean decomposition (GMD). In this process, the objective function of the spectral efficiency is derived, then the basic tracking principle and least square (LS) techniques are deployed to design the proposed hybrid precoding. Simulation results show that the proposed algorithm significantly improves system performance and reduces computational complexity by more than 45% compared to traditional algorithms.     

Hazard detection in noise-related incidents – the role of driving experience with battery electric vehicles

The low noise emission of battery electric vehicles (BEVs) has led to discussions about how to address potential safety issues for other road users. Legislative actions have already been undertaken to implement artificial sounds. In previous research, BEV drivers reported that due to low noise emission they paid particular attention to pedestrians and bicyclists. For the current research, we developed a hazard detection task to test whether drivers with BEV experience respond faster to incidents, which arise due to the low noise emission, than inexperienced drivers. The first study (N = 65) revealed that BEV experience only played a minor role in drivers’ response to hazards resulting from low BEV noise. The tendency to respond, reaction times and hazard evaluations were similar among experienced and inexperienced BEV drivers; only small trends in the assumed direction were observed. Still, both groups clearly differentiated between critical and non-critical scenarios and responded accordingly. In the second study (N = 58), we investigated additionally if sensitization to low noise emission of BEVs had an effect on hazard perception in incidents where the noise difference is crucial. Again, participants in all groups differentiated between critical and non-critical scenarios. Even though trends in response rates and latencies occurred, experience and sensitization to low noise seemed to only play a minor role in detecting hazards due to low BEV noise. An additional global evaluation of BEV noise further suggests that even after a short test drive, the lack of noise is perceived more as a comfort feature than a safety threat.     

高速公路隧道内交通噪声预测和降噪措施

结合公路隧道工程,利用波动声学和房间声学理论,研究公路隧道内噪声预测方法和降噪措施,给出预测模式、参数确定的方法和算例以及降噪措施的几种方案比较。