The Impact of Parametric Uncertainties on Mobile Robots Velocities and Pose Estimation

Abstract

This paper analyses the impact of parametric uncertainties of a mobile robot kinematic model on velocity and pose estimation, providing models and quantifiable knowledge about them. Most works neglect how the uncertainty regarding the robot's construction aspects (such as wheels radii, the distance between them, and the robot's center of mass) affects both velocities and pose estimation. To help readers understand the influence of such parametric uncertainties, we performed experiments in a simulator and used the collected data on the proposed models. The paper also analyses how the magnitude of velocities considered by the controllers and the followed path can decrease or increase the impact of the parametric uncertainties. The proposed models and presented analysis help understanding the influence (isolated or simultaneous) of different sources of uncertainty in the robot's velocities and pose estimation. This knowledge can be applied to estimate uncertainties for localization methods based on data fusion, complementing or even avoiding the experimental procedures. Also, the development of controllers, robot simulators and new methodologies for parameters' calibration and the design and construction of new robots can also profit from the results presented in this paper.

Principal Component Analysis (PCA) and Hough Transform as Tool for Simultaneous Localization and Mapping (SLAM) with Sparse and Noisy Sensors

“…θ esti (t) = θ(t − 1) + ∆θ (8) As proven in [7], orientation is the measurement with greatest error propagation for odometry, in the case of a differential track robot. This can be corrected with the developed technique.…”

“…In each step, measurements from sensors are transformed to the global reference frame by knowing the current position (5). However, instead of using (6) to find the rotation, the system of absolute localization is used to estimate R esti (7).…”

“…In this paper, the Hough Transform is used to detect line segments from the information acquired as output of sparse and noisy ultrasonic sensors in order to estimate orientation of a differential driven robot. The orientation was chosen for this analysis because it has the worst error propagation from the sensors for this type of robot, as mathematically proven in [7]. The estimation of the X and Y position is made based on the Extended Kalman Filter (EKF).…”

Low-Cost Reduced Navigation System for Mobile Robot in Indoor/Outdoor Environments

“…The unpredictability could be due to operation in harsh environments or due to corruption in the sensor's output due to noise or bias. The level of uncertainty arises if the robot lacks critical information for carrying out its tasks [16]. To address such issues, researchers often come up with algorithms that can improve the overall robustness of the system.…”

Deep Belief Network–Based Learning Algorithm for Humanoid Robot in a Pitching Game

“…There are several types of robots, such as humanoid robots [1], [2], mobile robots [3], and swarm robots [4]. They use biological mechanisms, perception, and artificial intelligence to emulate the behaviors, thoughts, feelings, and actions of human beings or other organisms.…”