【作者】 郑岩；

【导师】 谭庆昌；

【作者基本信息】 吉林大学 ， 机械设计及理论， 2010， 博士

【摘要】 运动平台上的跟踪在目标检测、跟踪及识别等方面有着广泛的应用,在武器装备及民用产品中有重要作用,在此背景下,对运动平台上跟踪的研究尤为重要。基于工程实际和现代控制理论,本文对运动平台上跟踪技术进行了较为深入的理论探讨和实践尝试。通过适当简化运动平台上跟踪系统结构,分析并建立了运动平台上跟踪系统动力学模型,依据现代控制理论,对建立的系统动力学模型进行了分析,采用状态反馈方法设计渐进跟踪鲁棒器,对执行装置俯仰姿态进行精确控制。分析并建立运动平台上跟踪系统的内环模型。通过跟踪算法的改进,对运动平台和机动目标的运动状态进行精确描述,建立运动平台上的跟踪系统的外环模型。利用卡尔曼滤波算法实现对目标运动轨迹进行较为精确的预测和跟踪。对本文建立的模型进行了仿真,结果证明了本文工作的有效性和显著意义。更多还原

【Abstract】 Object detecting, tracking and recognizing have a wide range of applications, for example, in military, the object’s observation, identification and trace on the battlefield, including fire control system on the platforms of armoured vehicles, aircraft, ships and others, and also the observing and aiming equipments such as anti-tank missiles, air defence missiles carried by a soldier; in civilian applications, such as the monitoring and tracking of unusual people and objects in security monitoring and controlling in some important places, including building surveillance, site security etc; and the detecting and tracking of rule-breaking vehicles on the highway or the intersection, as well as the important applications of CNC equipment, flexible manufacturing in the manufacturing field.With the continuous development of science and technology, especially the progress of information technology and controlling methods, the object tracking systems have evolved into a integrated system of multi-sensor information fusion technology and control engineering. And it has widely applications in lots of the fields, and plays important roles. So the research in tracking moving objects becomes particularly important and meaningful. The objective of this thesis is to an in-depth study and practice of trying, combined with real engineering.Usually, a dynamic tracking system on a moving platform consists of four parts: (1) the vehicle body part, i.e., footing part; (2) the implementation part, i.e., the crane and its auxiliary devices, also known as the upper part; (3) upper supporting part, i.e., the upper rotation body; (4) dynamic tracking parts, including observation and computing devices.A lots of work have been done internal and abroad in this field, especially in the research of moving platform, but there are following deficiencies:in the mechanical analysis of moving platform with caterpillar, people usually only consider the caterpillar and its forces from the ground, and ignore the forces on the loading wheels on the sides; various of methods are applied on the control of implementation devices to improve accuracy and speed of response, but the results are still not meet the actual requirements; dynamic tracking of objects are based on the relative gyroscope-based coordinate system, the accuracy of tracking the object space state is still to be improved.The moving platform is sensitive to the force from the ground, especially in the low speed case, the cause of the system vibration is very obvious. Considering the works on moving platform such as precise tank firing and precision engineering vehicles operating, mostly are implemented under low speed conditions, therefore the forces from ground and the vibration are most important factors during the design of the moving platform, and can not be ignored. Through appropriately simplifying structure and taking into account the relationship among the upper part’s vertical amplitude, vibration angle, body horizontally rotation and angle of depression and elevation, this thesis analyses and creates a mechanical model for dynamic tracking system on moving platform, and analyses the mechanical features when the moving platform get forced from ground, to reduce errors caused by system vibration, reduce response time and overshoot. On this basis, based on the modern control theory, we analysed the mechanical model of dynamical tracking on moving platform, and discussed the affections and interactions to vibrating angle of implementing device from multi factors. Using state feedback method, we convert coupled nonlinear multi input-output system into several independent single-input-output integral decoupling systems. Then design progressive tracking robust device, and accurately control the device positioning attitude. Through improving control, we analyse and create mechanical model of implementing devices on the moving platform, and then solve its control model, and analyze the model through Matlab simulation. The results of simulation verified the effectiveness of the control model and gave a conclusion that the control model can be applied to a variety of ground.The improvement of automatic tracking algorithm in this thesis can reduce the system response time and improve tracking forecast accuracy. Through the coordinate transformation based on earth coordinate system and the accurate description of the moving state of the moving platform and moving objects, we create the dynamic tracking model. After coordinate transformation, the object and our vehicle are mapped into a single reference coordinate system, and are accurately described. Thus a status equation describes the moving states of our and the enemy’s vehicles, it is very important to the further study of automatically tracking fire control system and improve the accuracy of first firing during battle. On this basis, the prediction of moving target’s track using Kalman filtering algorithm and with the improvement of the automatic tracking algorithm improvement, can shorten system response time and improve the tracking forecast accuracy. We also improve the controlling accuracy of implementing devices horizontally and vertically, and reduce the systematic errors and random disturbance sensitivity, also create the dynamic tracking system model on the moving platform, and verify the accuracy with the Monte Carlo shooting model.With analysis and study, this thesis provides the following main results:1, with the optimal control strategies, the vertical control accuracy of the implementing devices is improved. Comparing the traditional control performance, the overshoot is reduced by 90%, and angle accuracy increased by 40%.2, the existing dynamic tracking technology can simply calculate target angular velocity and speed with high precision, but cannot accurately measure the relative angle and relative angular acceleration and the position and acceleration. This thesis presents an algorithm with the use of the existing tank sensors and coordinate transformation, mapping the spatial coordinates into geodetic coordinate system, then can more accurately measure the relative target angle, the relative angular acceleration and the position and acceleration.3, Kalman filtering has been widely used for over 30 years, including robot navigation, control, sensor data fusion and even in military radar systems and missile tracking and so on. This thesis presents an automatic tracking algorithm based Kalman filter, with the recursive correction of the data of target’s real-time location, making the detection of spatial location on the target is more precise and more rapid.It is clearly seen from the simulation data that the application of the dynamic system tracking model provided by this thesis makes shooting accuracy significantly improved, and the experimental results show that the moving platform and moving state affect the model’s accuracy, in particular the object’s spatial state results in the decrease of tracking accuracy of the model. All these proved that the work on platform force analysis of platform from ground in this thesis is successful but the dynamic tracking on moving platform needs to be improved. In general, the simulation results prove the effectiveness and significance of the work in this thesis.更多还原