室内环境下基于SLAM的四旋翼无人机定位与控制
发布时间：2018-08-23 21:06
【摘要】：四旋翼无人机(Quadrotor Unmanned Aircraft Vehicle,QUAV)具有结构简单、成本低廉、性能卓越以及飞行控制方式独特等特点,成为近年来无人机领域的研究热点。在执行室内环境下的监视和侦察任务中,QUAV优势明显,具有广阔的应用前景。本文围绕室内环境下QUAV的定位与控制问题,给出了QUAV的非线性数学模型,使用同步定位与地图构建(Simultaneous Localization and Mapping,SLAM)技术解决QUAV的室内定位问题,并根据SLAM获得的位姿信息进一步研究受到建模不确定、外部干扰、输入饱和以及姿态受限等因素影响的QUAV控制方法。论文主要研究内容如下:首先,根据QUAV运动特点,研究了QUAV非线性数学模型。为了后续研究方便,将QUAV数学模型划分为由姿态环组成的快回路方程和由位置环组成的慢回路方程,并转换得到了具有仿射非线性方程形式的QUAV快回路和慢回路系统方程。然后,针对室内环境下QUAV的定位问题,综合使用激光测距仪、姿态传感器和高度传感器,研究了基于迭代最近点(Iterative Closest Point,ICP)方法的SLAM算法。针对激光数据和高度数据的歪斜问题,利用姿态传感器的姿态角数据进行数据修正;针对QUAV的室内定位问题,应用ICP匹配算法予以处理;针对地图构建问题,分别给出了栅格地图和几何地图的创建与更新步骤。SLAM实验结果表明了ICP定位算法稳定性好、精度高,并根据SLAM地图创建结果对比了两种地图表示方式。其次,针对QUAV的快回路系统和慢回路系统,考虑外部未知干扰,研究了一种基于干扰观测器的反步控制方法。设计干扰观测器在线逼近由系统建模误差和外部干扰构成的复合干扰,并根据干扰估计值设计反步控制策略,通过Lyapunov定理严格证明了闭环系统的稳定性。仿真结果表明,在存在建模误差和外部干扰的情况下,QUAV快慢回路系统均具有良好的跟踪性能。接着,针对QUAV的快回路系统,给出了一种具有建模不确定、外部干扰、输入饱和与姿态受限的反步控制方法。针对建模不确定,使用神经网络进行逼近;针对复合干扰,设计非线性干扰观测器对干扰进行补偿;针对输入饱和,使用双曲正切函数逼近饱和函数;针对输出受限问题,使用界限Lyapunov函数设计控制器,保证姿态满足限制条件。通过Lyapunov方法证明了闭环系统的所有信号半全局一致有界。仿真结果表明在具有建模不确定、外部干扰、输入饱和与姿态受限的情况下,所设计控制方法可得到满意的控制效果。最后,搭建了室内环境下QUAV的SLAM实验平台,分别在室内环境下的小场景区域和大场景区域中进行实验,验证了室内环境下SLAM算法的可行性和有效性。
[Abstract]:(Quadrotor Unmanned Aircraft Vehicle,QUAV) with simple structure, low cost, excellent performance and unique flight control mode has become a research hotspot in the field of UAV in recent years. QUAV has obvious advantages in carrying out surveillance and reconnaissance missions in indoor environment and has broad application prospects. In this paper, the nonlinear mathematical model of QUAV is given around the problem of positioning and control of QUAV in indoor environment, and the problem of indoor location of QUAV is solved by using synchronous location and map building (Simultaneous Localization and Mapping,SLAM) technology. According to the position and attitude information obtained by SLAM, the QUAV control method which is affected by modeling uncertainty, external interference, input saturation and attitude limitation is further studied. The main contents of this paper are as follows: firstly, according to the characteristics of QUAV motion, the nonlinear mathematical model of QUAV is studied. For the convenience of further study, the QUAV mathematical model is divided into fast loop equations composed of attitude loops and slow loop equations composed of position loops, and QUAV fast loop and slow loop system equations with affine nonlinear equations are obtained. Then, aiming at the localization problem of QUAV in indoor environment, the SLAM algorithm based on iterative nearest point (Iterative Closest Point,ICP (Iterative Closest Point,ICP) method is studied by using laser rangefinder, attitude sensor and height sensor. Aiming at the skew problem of laser data and altitude data, the attitude angle data of attitude sensor is used to correct the data; to solve the indoor positioning problem of QUAV, ICP matching algorithm is used to deal with the problem; to solve the problem of map construction, The steps of creating and updating raster map and geometric map are given respectively. The results of SLAM experiment show that the ICP localization algorithm has good stability and high precision, and the two kinds of map representation methods are compared according to the results of SLAM map creation. Secondly, for the fast loop system and slow loop system of QUAV, a backstepping control method based on disturbance observer is studied, considering the external unknown disturbance. The disturbance observer is designed to approach the complex disturbance which consists of modeling error and external disturbance. According to the disturbance estimation, the backstepping control strategy is designed. The stability of the closed-loop system is strictly proved by Lyapunov theorem. Simulation results show that QUAV fast and slow loop systems have good tracking performance in the presence of modeling errors and external disturbances. Then, for the fast loop system of QUAV, a backstepping control method with modeling uncertainty, external disturbance, input saturation and attitude limitation is proposed. For modeling uncertainty, neural network is used to approximate; for complex disturbance, nonlinear disturbance observer is designed to compensate for disturbance; for input saturation, hyperbolic tangent function is used to approximate saturation function; for the problem of limited output, The limit Lyapunov function is used to design the controller to ensure that the attitude meets the limiting conditions. It is proved that all the signals of the closed loop system are semi-globally uniformly bounded by the Lyapunov method. The simulation results show that the proposed control method can obtain satisfactory control results under the conditions of uncertain modeling, external interference, input saturation and attitude limitation. Finally, the SLAM experimental platform of QUAV in indoor environment is built, and the experiment is carried out in the small scene area and large scene area in the indoor environment, which verifies the feasibility and effectiveness of the SLAM algorithm in the indoor environment.
【学位授予单位】：南京航空航天大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：V279
[Abstract]:(Quadrotor Unmanned Aircraft Vehicle,QUAV) with simple structure, low cost, excellent performance and unique flight control mode has become a research hotspot in the field of UAV in recent years. QUAV has obvious advantages in carrying out surveillance and reconnaissance missions in indoor environment and has broad application prospects. In this paper, the nonlinear mathematical model of QUAV is given around the problem of positioning and control of QUAV in indoor environment, and the problem of indoor location of QUAV is solved by using synchronous location and map building (Simultaneous Localization and Mapping,SLAM) technology. According to the position and attitude information obtained by SLAM, the QUAV control method which is affected by modeling uncertainty, external interference, input saturation and attitude limitation is further studied. The main contents of this paper are as follows: firstly, according to the characteristics of QUAV motion, the nonlinear mathematical model of QUAV is studied. For the convenience of further study, the QUAV mathematical model is divided into fast loop equations composed of attitude loops and slow loop equations composed of position loops, and QUAV fast loop and slow loop system equations with affine nonlinear equations are obtained. Then, aiming at the localization problem of QUAV in indoor environment, the SLAM algorithm based on iterative nearest point (Iterative Closest Point,ICP (Iterative Closest Point,ICP) method is studied by using laser rangefinder, attitude sensor and height sensor. Aiming at the skew problem of laser data and altitude data, the attitude angle data of attitude sensor is used to correct the data; to solve the indoor positioning problem of QUAV, ICP matching algorithm is used to deal with the problem; to solve the problem of map construction, The steps of creating and updating raster map and geometric map are given respectively. The results of SLAM experiment show that the ICP localization algorithm has good stability and high precision, and the two kinds of map representation methods are compared according to the results of SLAM map creation. Secondly, for the fast loop system and slow loop system of QUAV, a backstepping control method based on disturbance observer is studied, considering the external unknown disturbance. The disturbance observer is designed to approach the complex disturbance which consists of modeling error and external disturbance. According to the disturbance estimation, the backstepping control strategy is designed. The stability of the closed-loop system is strictly proved by Lyapunov theorem. Simulation results show that QUAV fast and slow loop systems have good tracking performance in the presence of modeling errors and external disturbances. Then, for the fast loop system of QUAV, a backstepping control method with modeling uncertainty, external disturbance, input saturation and attitude limitation is proposed. For modeling uncertainty, neural network is used to approximate; for complex disturbance, nonlinear disturbance observer is designed to compensate for disturbance; for input saturation, hyperbolic tangent function is used to approximate saturation function; for the problem of limited output, The limit Lyapunov function is used to design the controller to ensure that the attitude meets the limiting conditions. It is proved that all the signals of the closed loop system are semi-globally uniformly bounded by the Lyapunov method. The simulation results show that the proposed control method can obtain satisfactory control results under the conditions of uncertain modeling, external interference, input saturation and attitude limitation. Finally, the SLAM experimental platform of QUAV in indoor environment is built, and the experiment is carried out in the small scene area and large scene area in the indoor environment, which verifies the feasibility and effectiveness of the SLAM algorithm in the indoor environment.
【学位授予单位】：南京航空航天大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：V279
【参考文献】
相关期刊论文 前6条
1 杨荟a，
本文编号：2199915
本文链接：https://www.wllwen.com/kejilunwen/hangkongsky/2199915.html
