Analysis of multi-agent activity using petri nets

This paper presents the use of place/transition petri nets (PNs) for the recognition and evaluation of complex multi-agent activities. The PNs were built automatically from the activity templates that are routinely used by experts to encode domain-specific knowledge. The PNs were built in such a way that they encoded the complex temporal relations between the individual activity actions. We extended the original PN formalism to handle the propagation of evidence using net tokens. The evaluation of the spatial and temporal properties of the actions was carried out using trajectory-based action detectors and probabilistic models of the action durations. The presented approach was evaluated using several examples of real basketball activities. The obtained experimental results suggest that this approach can be used to determine the type of activity that a team has performed as well as the stage at which the activity ended.     

Adaptive input shaping for single-link flexible manipulators using an algebraic identification

This work proposes an adaptive control scheme applied to single link-flexible manipulators, which combines a feedback controller of the joint angle with an adaptive input shaper updated by an algebraic non-asymptotic identification. The feedback controller is designed to guarantee trajectory tracking of the joint angle, simplifying thus the input shaper, which can be designed for the arm dynamics only. The input shaper is updated by an algebraic identification of the natural frequency corresponding to the first vibration mode of the arm. In addition, the influence of the assumptions adopted to derive the algebraic identification on the performance of the estimation is studied. Finally, the proposed adaptive control strategy is implemented in practice.     

Detection of non-sea salt sulphate aerosol at a remote coastal site in South Africa: A PIXE study

An aerosol sampling site based on a single-stage streaker sampler with four-hour time resolution was established in 1993 at Brandt-se-Baai, on the north-western coast of South Africa (31.5°S 18°E, 50 m asl). The site was deployed to monitor the generation of dust from the local mining operations and in part to test the hypothesis of large-scale air mass recirculation over Southern Africa. Streaker samples were analysed by PIXE for the usual crustal derived elements and sulphur. Sulphate aerosols over southern Africa are produced in the atmosphere by oxidation of SO₂, from industry, or DMS, from the ocean. Industrial related sulphate plays a key role in tracing long-range transport of anthropogenic pollutants.

We present results from selected monthly samples collected during 1996. Measured particulate sulphur has been apportioned between contributions from sea spray and non-sea salt sulphate (NSS) aerosols. Marine particulate sulphur, present in the regional background, was traced by its correlation with Na and Cl, which were routinely detected. The NSS content of these air masses was enriched with respect to the expected sea salt sulphur by two orders of magnitude. Al, Si, K, Ca, Ti and Fe, the major crustal elements, were measured only episodically with the occurrence of strong south-easterly winds. These winds are associated with wind blown dust from a mine located 4 km from the site.

These results are the first direct aerosol characterisation of the remote region of the western seaboard of southern Africa. Specifically the sulphur measurements at this site provide insights into relative contributions of natural DMS, industrial and sea salt contributions of particulate sulphate.     

Experiments in laser-assisted visual sensing for AUV navigation

The Cranfield University torpedo-shaped underwater vehicle has been modified to accommodate a laser stripe illumination system. As well as providing enhanced viewing capabilities, this system derives real-time navigational data during the mission and gathers images to produce a post mission enhanced optical waterfall image of a surveyed area.

This paper describes a preliminary set of constrained motion trials at the IFREMER wave basin in Brest, where the system was towed through the 50 m test tank at different altitudes and orientations whilst the true trajectory was measured. A comparison is made between ground truth trajectory generated from these external measurements and that derived from the video camera and rotation sensors internal to the vehicle.     

Video traffic analysis using scene and vehicle models

This paper presents a complete system for analyzing a vehicle׳s behavior in the context of real-time traffic video surveillance applications. To obtain the best possible results, it is fundamental to exploit the scene characteristics and the predefined traffic rules. For that purpose, an initial training step is performed that involves estimating the geometrical structure of the road, i.e., the depth relative to the camera, the vanishing points, the road areas itself, the road decomposition (into normal and forbidden traffic lanes or areas), the traffic rules, the typical vehicle trajectories and speeds, and the lane-changing rules. This process leads to a scene model, which is used together with a simple vehicle geometrical model during the vehicle detection, tracking and trajectory estimation phases to improve the robustness against the perspective and occlusion effects. Shadow effects are also accounted for during the moving object detection phase. Finally, this spatio-temporal analysis is used to obtain information that concerns the vehicles׳ behaviors. Experiments show that the information obtained is reliable and can be computed in real time.     

Kinetic and kinematic responses of post mortem human surrogates and the Hybrid III ATD in high-speed frontal sled tests

Despite improvements in vehicle design and safety technologies, frontal automotive collisions continue to result in a substantial number of injuries and fatalities each year. Although a considerable amount of research has been performed on PMHSs and ATDs, matched dynamic whole-body frontal testing with PMHSs and the current ATD aimed at quantifying both kinetic and kinematic data in a single controlled study is lacking in the literature. Therefore, a total of 4 dynamic matched frontal sled tests were performed with three male PMHSs and a Hybrid III 50th percentile male ATD (28.6 g, Δv = 40 kph). Each subject was restrained using a 4 kN load limiting, driver-side, 3-point seatbelt. Belt force was measured for the lap belt and shoulder belt. Reaction forces were measured at the seat pan, seat back, independent foot plates, and steering column. Linear head acceleration, angular head acceleration, and pelvic acceleration were measured for all subjects. Acceleration of C7, T7, T12, both femurs, and both tibias were also measured for the PMHSs. A Vicon motion analysis system, consisting of 12 MX-T20 2 megapixel cameras, was used to quantify subject 3D motion (±1 mm) at a rate of 1 kHz. Excursions of select anatomical regions were normalized to their respective initial positions and compared by test condition and between subject types. Notable discrepancies were observed in the responses of the PMHSs and the ATD. The reaction forces and belt loading for the ATD, particularly foot plate, seat back, steering column, and lap belt forces, were not in agreement with those of the PMHSs. The forward excursions of the ATD were consistently within those of the PMHSs with the exception of the left upper extremity. This could potentially be due to the known limitations of the Hybrid III ATD shoulder and chest. The results presented herein demonstrate that there are some limitations to the current Hybrid III ATD under the loading conditions evaluated in the current study. Overall, this study presents a comprehensive data set of belt forces, reaction forces, accelerations, and bilateral displacement data that can be used to evaluate the performance of ATDs and validate computational models.     

一种面向视觉注意区域检测的运动分割方法*

结合视觉显著区检测的特点,本文提出一种面向视觉注意区域检测的运动分割方法。该方法用一种层次聚类方法将特征点的运动轨迹进行聚类。首先用中值偏移算法扩大了不同类型运动之间特征向量的差距,同时缩小了相同运动类型的差别。继而,用一种无监督聚类算法,将不同类型的运动进行分割,同时自动获得运动分类数。最后利用运动分割结果,提出一种结合空间和颜色采样的运动显著区域生成方法。与以往方法相比,该方法能够将不同类型的运动自动进行分割,生成的视觉注意区域更为准确,而且稳定性大幅提高。实验结果证明了该方法的有效性和稳定性。     

视频识别中基于簇的在线运动分割算法研究

仿射相机模型下,运动分割问题转化为子空间分离问题,处理这类问题的算法大多是离线算法,当假设不满足时性能很不理想.针对上述问题,提出一种在线运动分割算法,通过动态标签传输和簇分割进行运动分割.首先,根据固定数量的帧进行初始化,接着,通过在线策略更新轨迹相似性,最后,利用动态标签传输技术在帧间传输信息,对簇进行评估和归一化切分成本估计,实现动态的簇分割.基于基准数据集的仿真实验结果表明,算法的运行结果与离线算法相当.     

3-Dimensional curve similarity using string matching

In this paper, we present a new approach to measuring the similarity between 3D-curves. Our approach allows the possibility of using strings, where each element is a vector rather than just a symbol. We present two different approaches for representing 3D-curves. One possibility is to represent a 3D-curve as two 2D-curves, one being the projection of the 3D-curve in the XY-plane, and the other one in the YZ-plane. For the case that we need geometric rotation invariance, we have used a second approach to the symbolic representation of the 3D-curve using the curvature and the tension as their symbolic representation. We validate this approach through experiments using synthetic and digitalized data.     

考虑扰动的开关控制法

开关控制法虽具有响应迅速的优点,但当出现扰动时容易失控以致不能应用。本文考虑了扰动因素,推导出一组算式,利用计算得出的开关曲线,可实现开关控制法的稳定控制。以瓦楞纸板生产线中的切割作业为例,给出实时控制程序和系统结构框图。