Alatise Mary B, Hancke Gerhard P
Department of Electrical, Electronic and Computer Engineering, University of Pretoria, Pretoria 0028, South Africa.
Department of Computer Science, City University of Hong Kong, Hong Kong, China.
Sensors (Basel). 2017 Sep 21;17(10):2164. doi: 10.3390/s17102164.
Using a single sensor to determine the pose estimation of a device cannot give accurate results. This paper presents a fusion of an inertial sensor of six degrees of freedom (6-DoF) which comprises the 3-axis of an accelerometer and the 3-axis of a gyroscope, and a vision to determine a low-cost and accurate position for an autonomous mobile robot. For vision, a monocular vision-based object detection algorithm speeded-up robust feature (SURF) and random sample consensus (RANSAC) algorithms were integrated and used to recognize a sample object in several images taken. As against the conventional method that depend on point-tracking, RANSAC uses an iterative method to estimate the parameters of a mathematical model from a set of captured data which contains outliers. With SURF and RANSAC, improved accuracy is certain; this is because of their ability to find interest points (features) under different viewing conditions using a Hessain matrix. This approach is proposed because of its simple implementation, low cost, and improved accuracy. With an extended Kalman filter (EKF), data from inertial sensors and a camera were fused to estimate the position and orientation of the mobile robot. All these sensors were mounted on the mobile robot to obtain an accurate localization. An indoor experiment was carried out to validate and evaluate the performance. Experimental results show that the proposed method is fast in computation, reliable and robust, and can be considered for practical applications. The performance of the experiments was verified by the ground truth data and root mean square errors (RMSEs).
使用单个传感器来确定设备的位姿估计无法得到准确结果。本文提出了一种融合六自由度（6-DoF）惯性传感器（包括加速度计的3轴和陀螺仪的3轴）与视觉的方法，以确定自主移动机器人的低成本且准确的位置。对于视觉部分，集成了基于单目视觉的目标检测算法加速鲁棒特征（SURF）和随机抽样一致性（RANSAC）算法，并用于在拍摄的多幅图像中识别样本目标。与依赖点跟踪的传统方法不同，RANSAC使用迭代方法从一组包含异常值的捕获数据中估计数学模型的参数。借助SURF和RANSAC，精度必然会提高；这是因为它们能够使用黑塞矩阵在不同观察条件下找到兴趣点（特征）。提出这种方法是因为其实现简单、成本低且精度提高。通过扩展卡尔曼滤波器（EKF），融合了来自惯性传感器和相机的数据，以估计移动机器人的位置和方向。所有这些传感器都安装在移动机器人上以获得精确的定位。进行了室内实验以验证和评估性能。实验结果表明，所提出的方法计算速度快、可靠且稳健，可考虑用于实际应用。实验性能通过地面真值数据和均方根误差（RMSE）进行了验证。
