A Bayesian grid method PCA-based for mobile robots localization in unstructured environments

Abstract

This paper presents the experimental validation of a new method for mobile robot global self-localization in unstructured environments, i.e. that does not need any beacons or other artifacts structuring the environment. The method resorts to a PCA-based positioning sensor, filtered in a Bayesian probabilistic grid and combined with linear Kalman filters to estimate the global pose of mobile robots. In the implemented system, the information of the environment is captured only with onboard sensors installed in a differential drive robot: encoders, compass, and 2D depth sensor pointed to the ceiling. The use of PCA in a Bayesian probabilistic grid allows to fuse the highly compressed PCA database information, obtained with the low computational effort, in an environment where repetitive scenarios can occur. To avoid the negative impact in the localization estimate caused by the corrupted data existing in the 2D depth sensor, an extension to the classic PCA algorithm is suggested. Thus, the proposed method allows the self-localization of mobile robots in indoor environments with bounded accuracy and working in a wide range of illumination conditions.

A Review on Challenges of Autonomous Mobile Robot and Sensor Fusion Methods

“…The geometric representation of the environment is closer to the sensor and actuator world and it is the best one to perform local navigation. In [22], the author proposed the use of Principal Components Analysis (PCA) -Bayesian based method with grid map representation to compress images and reduce computational resources. The PCA was also use to reduce dimensionality and model the parameter of the environment by considering the pixels of an image as feature vectors of the data set [23].…”

Mobile robot self-localization in unstructured environments based on observation localizability estimation with low-cost laser range-finder and RGB-D sensors

“…However, real-time monocular-based navigation in unstructured environments using front or omnidirectional cameras is difficult, because image processing for unstructured scenes is complex and computationally heavy. In the study by Carreira et al 13 and Rodrigues et al , 14 an image-based navigation system is proposed for localization of mobile robot in indoor unstructured environments, but the system is restricted to work in buildings with static ceilings containing rich visual information. Visual odometry (VO) has been widely applied for rotation and translation estimation of mobile robots.…”

“…The determinant of DðPÞ well reflects the physical meaning of the robot's laser observation localizability dðpÞ ¼ detðDðPÞÞ (14) In other words, the higher value dðpÞ is, the lower amount of unexpected objects are in the robot's neighborhood, and thus the higher localizability the robot's current laser observation will have.…”

“…, 14 an image-based navigation system is proposed for localization of mobile robot in indoor unstructured environments, but the system is restricted to work in buildings with static ceilings containing rich visual information. Visual odometry (VO) has been widely applied for rotation and translation estimation of mobile robots.…”

Mobile Robot Localisation for Indoor Environments Based on Ceiling Pattern Recognition

“…Rodrigues at al. made studies in warehouse environments using visual odometry of ceiling images [6], Markov Localization [7] and a 3D camera applied to a depth map of the ceiling. Carreira et al [8] also use depth maps and introduce a technique to deal with missing data from the 3D camera.…”