一种仿生触须传感器的设计与分析

为了增强机器人对未知环境的感知与识别能力,设计了一种基于霍尔效应的仿生触须传感器。针对该传感器建立了触须感知模型,并在此基础上进行仿真实验,分析触须轴的不同材质和长度对触须传感器感知结果的影响。同时,提出了可通过触须的偏转幅度来估计目标物距离和角度位置的被动和主动感知方法。实验结果表明：利用该触须传感器能够快速、准确地感知外界环境信息。     

感知距离参数优化的无线传感器网络节能的研究

无线传感器网络(WSNs)覆盖是WSNs的关键问题,它反映了网络对目标区域的探测能力,节点的感知距离的优化能够很好地解决WSNs区域覆盖要求的同时使得能量消耗最低.研究了基于节点感知距离参数调节的WSNs节能区域覆盖方案,建立了感知半径之和最小的数学模型,采用遗传算法求解该模型,得到一个最优解.仿真结果表明:与传统覆盖方案比较,所提方案在满足覆盖要求的同时有效地降低了网络能量消耗.     

基于神经网络的压力传感器数据融合

针对压力传感器在实际应用中受多个非目标参量的影响而导致其输出数据不仅仅与目标参量有关,提出了应用神经网络技术对多传感器数据进行融合以消除非目标参量对传感器输出的影响;研究结果表明,利用神经网络实现多传感器数据融合消除非目标参量对传感器输出的影响是可行的.     

基于神经网络的压力传感器数据融合

针对压力传感器在实际应用中受多个非目标参量的影响而导致其输出数据不仅仅与目标参量有关，提出了应用神经网络技术对多传感器数据进行融合以消除非目标参量对传感器输出的影响；研究结果表明，利用神经网络实现多传感器数据融合消除非目标参量对传感器输出的影响是可行的。     

基于BP神经网络的压力传感器静态特性数据融合

针对压力传感器在实际应用中受非目标参量(温度)的影响而导致其输出数据不仅仅与目标参量(压力)有关，提出了采用BP神经网络技术对其进行数据融合处理，以消除非目标参量对压力传感器输出的影响。研究结果表明，利用神经网络实现传感器数据融合消除非目标参量对传感器输出的影响是可行的。     

概率模型下异构传感器网络部署算法的研究

目前采用虚拟力方法解决传感器节点部署问题的算法均基于同构传感器网络,面向异构传感器网络的部署需求,提出扩展的虚拟力算法.该算法采用概率感知模型,部署时根据感知半径的悬殊采用静态部署与动态部署相结合的策略,根据节点感知半径差异度决定最佳距离的取值,节点移动时采用接替移动法.仿真结果表明该算法能够根据应用需要将异构传感器节点合理地部署于目标区域内,同时能有效地均衡网络节点的能耗,延长网络的生存时间.     

基于改进蚁群算法的海上编队传感器资源分配模型

在信息化条件下,海上编队作战中传感器资源分配是传感器资源管理的关键内容。针对传感器资源分配模型构建因素不全面,分配算法计算量与时间随着传感器目标数目的增多而急剧增加等问题,本文提出一种考虑目标优先级、传感器对目标的匹配精度、传感器作用范围和协同能力的目标函数,建立一种基于改进蚁群算法的海上编队传感器资源分配模型。仿真结果表明,该算法显著提高了收敛速度与时间满意度,验证了分配模型的可行性。     

微传感器网络多分辨率感知模型

环境监测是微传感器网络最广泛的应用之一,它通过收集目标区域内单个传感器节点的感知数据来建立环境指标在该区域内的分布状态.由于传感器节点的能源有限以及补充困难,如何平衡能量消耗的有效性和感知精度是一个非常重要的问题.提出了一种多分辨率的感知模型,通过建立与精度要求相适应的感知分辨率层次以期有效地解决这一问题,并支持对感知数据精度的适应性调整.模拟和分析表明,这种多分辨率的感知模型可以有效地实现能耗有效性和感知精度的折中.     

基于数据融合的压力传感器静态特性研究

通常传感器的输出值不仅决定于目标参量,还会受到非目标参量的影响。为此,采用BP神经网络技术对其进行数据融合处理,以消除非目标参量对传感器输出值的影响。试验结果表明:该方法很好地抑制了传感器的交叉灵敏度,提高了其测量准确度。     

矩形柔性涡流阵列传感器裂纹检测研究

某矩形涡流阵列传感器可对金属结构的裂纹进行定量检测,但使用中往往会受到提离效应的影响。通过建立矩形涡流阵列传感器的有限元模型,分析了传感器对提离距离的响应特性,并搭建裂纹检测试验平台,研究了提离距离对传感器裂纹检测能力的影响。研究表明:矩形涡流阵列传感器对提离效应十分敏感,在裂纹检测过程中应尽量抑制提离效应,同时提离距离的增加会显著地降低传感器的裂纹检测能力,在裂纹检测过程中应尽量减小提离距离。     

Chemical sensing by analysing dynamics of plasma polymer film-coated sensors

A chemical sensing system that incorporates unique sensor films and uses pattern recgonition of their dynamic responses is presented. This system consists of a sensor array of quartz-crystal microbalances coated with plasma polymer films. The films, synthesized by radio-frequency sputtering, are useful because of their high density of radical sites and unsaturated bonds. When this sensor array is exposed to a gas, the adsorption and desorption of the target gas causes a dynamic frequency response in each piezoelectric sensor. The sensor response is analysed by an autoregressive model typically to estimate the parameters of dynamic systems. This model's coefficients reflect the sensor dynamics, providing pattern vectors that characterize the target gas. Based on this model, classification maps for single gases can be created with these pattern vectors. Thesse maps show that the dynamic sensor response provides useful information for gas classification. This model also confirms that our sensing system can identify the components of a gas mixture.     

基于遗传算法的无线传感器网络重新部署方法

研究了无线传感器网络在受限移动能力条件下的重新部署问题.针对节点的运动模型为跳跃式移动,提出一种基于遗传算法的重新部署算法.算法以节点的跳跃方向为遗传算法的基因,适用度函数同时考虑了最大化覆盖率和最小化移动总距离.仿真实验表明,在各向同性的感测模型中,此算法优于文献[8]提出的FBSD算法,能实现节点的最优运动规划,并且在有向感测模型中,此算法也能有效提高网络覆盖率.     

无线传感器网络中能量均衡和覆盖半径自适应调整的分布式目标覆盖算法

为了最大化目标覆盖的生命期,提出一种能量均衡和覆盖半径自适应调整的分布式目标覆盖算法.首先提出本地h 跳的目标覆盖问题的定义,分析目标覆盖的能量模型,发现最小生命期的目标是目标覆盖生命期的瓶颈,引入关键目标的概念;设计了基于覆盖收益和能耗代价的能量效用函数,在能量均衡原则和能效优先原则的基础上设置节点等待时间,并建立了节点的覆盖半径和等待时间自适应调整的机制.实验表明该算法网络生命期延长25%左右.算法复杂度低,效率高,可扩展性好.     

A high performance real-time Interferometry Sensor System Architecture

Optical feedback or self-mixing interferometry technique has been widely used for sensing vibration, displacement, velocity, distance and flow applications. Such applications require an accurate and consistent output for real-time target measurements. Furthermore, array based sensing, as opposed to point sensing, is being increasingly pursued which requires multiple-input, parallel computing platform. In this article, we have proposed and developed a real-time interferometric sensor applications based multi-core system architecture and programming toolkit. To show that our system is useful in a variety of situations, we applied three algorithms of varying complexity and accuracy for displacement and vibration measurement. In order to prove the efficiency of proposed sensor processing system, we compared it performance and power consumption with a state of the art NXP LPCX54102 sensor processing and motion system architecture. When compared with the baseline multi-core system, the results show that our system improves the system performance upto 7.55 times, draws 15.6% less dynamic power and consumes 8.9 times less energy.     

视觉传感器网络节点方向调度算法

视觉传感器网络中节点感知模型为有向感知模型,感知范围被限定在节点的视野范围内.节点拥有多个感知方向,在一个工作时刻,节点只能工作于一个方向.针对视觉传感器网络的有向感知模型的点目标覆盖问题,提出一种贪婪算法(GreedyAlgorithm),在优化网络对于目标点覆盖的同时,解决了节点自身覆盖冲突的问题.在此基础上,引入贡献率概念反映节点在其感知方向上对于覆盖的贡献,提出了基于贡献率的贪婪算法(Contribution Rate Greedy Algorithm)以增大网络覆盖率.仿真实验表明了该算法的有效性.     

Duration of stay based weighted scheduling framework for mobile phone sensor data collection in opportunistic crowd sensing

Advancement of mobile phone based new sensing paradigm like opportunistic and participatory crowd sensing has lead to increase in both multimedia and scalar sensor data traffic over wireless networks. In opportunistic crowd sensing, there are more chances of missing required mobile phones sensor data due to unpredictable mobility nature of users. Analysis of scheduling schemes under realistic human mobility models has become an important issue for pervasive sensing applications specifically for mobile phone based crowd sensing. This paper refers to a weighted scheduling framework for collecting mobile phone sensor data under Levy Walk mobility model. The proposed method uses duration of stay of mobile phone users as one of the important parameter to improve the scheduling performance in terms of reducing the rate of missing of mobile nodes and thereby increasing the rate of collection of required sensor data. The simulation results show that for improving the scheduling performance in opportunistic crowd sensing, high weight value should be given for duration of stay parameter, when mobile nodes stay within the data collection region for more than schedule iteration time with high probability.     

机器人触须传感器的设计

提出了一种利用触须来识别物体表面轮廓的新方法。传感器采用了二维PSD作为敏感元件,实时地测量由于与物体接触在触须根部所产生的微小位移量。介绍了机器人触须传感器具体的结构及工作原理。由实验数据可知,触须根部位移量的大小与接触的距离成反比,且距离越近位移量变化的速度越快。由此可获得待测物体的位置、距离、角度等信息。     

反射式强度补偿光纤角位移传感器研究

提出了一种基于三探头等间距排列结构的反射式光纤角位移传感器,实现具有强度补偿的大量程、高灵敏度角位移测量.理论分析了该传感器的强度补偿角位移测量机理,建立了数学模型得到角度传感调制函数的表达式;仿真分析了光纤探头端面距反射面距离h及探头旋转半径R对传感特性的影响;实验验证了单接收光纤角位移传感特性.结果表明:输出光强和角位移之间呈现较好的线性关系;距离h越大,传感灵敏度越高;R越小,灵敏度越低,但同时传感区间相对越大.     

基于角度补偿的手机多传感器数据融合测距算法

在惯性测量领域,单纯利用加速度二次积分的方法并不能准确感知目标对象移动的距离.加速度传感器在感知呈线性运动的目标对象时较为准确和实用,但在三维空间运动时它的坐标轴会随物体发生方向的改变而不断漂移.为解决该问题,提出了一种基于角度补偿的手机多传感器数据融合测距算法(ADC-R),使用加速度传感器测量物体运动的加速度,作为计算位移的原始数据;采用手机陀螺仪传感器测量运动物体的角速度,并以旋转矢量传感器输出的数据作为参数把手机动态坐标系下测得的加速度值空间坐标转换到静态的参考坐标系下,然后进行数据融合完成角度补偿计算;最后根据物理学加速度和位移的关系运用数学积分方法和进一步修正误差的技术得到最终移动的距离.实验结果表明此方法在近距离测距方面精度较高,优于加速度积分算法和加速度与陀螺仪融合算法.     

Multi-objective optimization for coverage control in wireless sensor network with adjustable sensing radius

In this paper, the problem of maintaining sensing coverage by keeping a small number of active sensor nodes and a small amount of energy consumption in a wireless sensor network is studied. As opposed to the uniform sensing model previously, we consider a large number of sensors with adjustable sensing radius that are randomly deployed to monitor a target area. A novel coverage control scheme based on elitist non-dominated sorting genetic algorithm (NSGA-II) is proposed in a heterogeneous sensor network. By devising a cluster-based architecture, the algorithm is applied in a distributed way. Furthermore, an ameliorated binary coding is addressed to represent both sensing radius adjustment and sensor selection. Numerical and simulation results validate that the procedure to find the optimal balance point among the maximum coverage rate, the least energy consumption, as well as the minimum number of active nodes is fast and effective.