Authors

Abstract

Omni-directional vision sensor can provide information within the sensor range, and the directional angle of an object can be accurately obtained through omni-directional images. Based on this characteristic, an automatic navigation and positioning system for agricultural machinery is developed, and a three-dimensional positioning algorithm for agricultural wireless sensor networks based on cross particle swarm optimization is proposed. The method mainly includes three stages: convergence node selection, measurement distance correction and node location. Using the idea of crossover operation of genetic algorithm for reference, the diversity of particles is increased, and the influence of ranging error and the number of anchor nodes on positioning results is effectively improved. The location algorithm has the ability of global search. On the positioning node, the symmetric bidirectional ranging algorithm based on LFM (Linear frequency modulation) spread spectrum technology is used to calculate the distance between the positioning node and each beacon node, and the trilateral centroid positioning algorithm is used to calculate the coordinate position information of unknown nodes. Finally, the Kalman filter algorithm is used to superimpose the observed values of the target state to solve the influence of measurement noise on the positioning accuracy.

Abstract in Chinese

全向视觉传感器可以提供传感器范围内的信息，通过全向图像可以准确获得物体的方向角。研究基于这一特点，开发了农业机械自动导航定位系统，并提出一种基于交叉粒子群优化的农业无线传感器网络三维定位算法。该方法主要包括收敛节点选择、测量距离校正和节点定位三个阶段。研究借鉴遗传算法交叉运算的思想，增加了粒子的多样性，有效地改善了测距误差和锚节点数目对定位结果的影响。定位算法具有全局搜索能力。在定位节点上，采用基于LFM扩频技术的对称双向测距算法计算定位节点与各信标节点之间的距离，采用三边质心定位算法计算未知节点的坐标位置信息。最后，利用卡尔曼滤波算法对目标状态的观测值进行叠加，解决测量噪声对定位精度的影响。