Paulo Roberto Gardel Kurka

Vision-Based Measurement for Localization of Objects in 3-D for Robotic Applications

Stereo vision is widely used in many 3-D image applications. In this paper, we present a methodology to estimate the localization of objects in 3-D scenes using collaborative robots. Using this method, the robots are able to measure the location of objects using only a single camera installed on each one of them. From one pair of images, each one acquired by a different robot, the method locates homologous points, and then rebuilds the object by using equations for coordinate transformations. The experimental results validate the application of the proposed method for the measurement of the pose of objects in autonomous robotic applications.

Velocity Estimation for Autonomous Vehicles Based on Image Analysis

Stereo vision is widely used in many 3-D image applications. In this paper, a system based on stereo vision concepts has been developed with the goal to measure the velocity of vehicles using images. Using this method, an autonomous vehicle is able to measure its velocity using only a single camera. From a sequence of images, the method uses techniques to locate homologous points and then estimates the velocity of the vehicle. The experimental results validate the application of the proposed method to autonomous vehicles.

Architecture for multi-camera vision system for automated measurement of automotive components

This paper proposes the architecture of a multi-view cameras system with the purpose of online dimensional measuring of automotive components, specifically crankshafts. The introduction of a measuring vision system in a production line allows the dimensional quality control of all manufactured items, helping to optimize the process. The system consists of six digital cameras, positioned around the object, capturing different perspective images of the product. Processing of the images yields results of precise dimensional measurements of the product. The paper presents the ideas and algorithms used in the proposed measurement system, and virtual images simulation to access its operational and precision characteristics.

vSlam experiments in a custom simulated environment

In order to perform vSLAM experiments, a camera capturing information along a certain path is necessary. What if the resources to execute this are not available or a proof of concept is needed before a real application? This paper describes a simple method of combining MATLAB and Blender to perform vSLAM experiments in a custom created environment. A virtual robot and camera are created to navigate through the simulated space to capture textured images and objects that are used for the image processing. The images or video created by Blender can be easily processed by MATLAB as shown in this work. In addition, a simple application is demonstrated in the experiments section.

On linear and nonlinear trajectory tracking control for nonholonomic integrator

This paper presents two different kinds of trajectory tracking control strategies for the nonholonomic integrator known in literature as Brockett system. The first strategy presents a time-varying linear feedback control law and the second strategy is based on State Dependent Ricatti Equation (SDRE) method. Numerical simulation results indicated that both methods can be successfully used for control of the nonholomic integrator. Keywords: Brockett integrator, Hamilton – Jacobi – Bellman equation, SDRE method.

On linear and nonlinear tracking of the wheeled mobile robot

This paper presents two different trajectory tracking control strategies for wheeled mobile robot. The first strategy presents a time-varying linear feedback control law and the second strategy is based on State Dependent Ricatti Equation (SDRE) method. Numerical simulation results indicated that both methods can be successfully used for control of the robot system.

Visual odometry in mobile robots

The paper presents an application of visual odometry, through reconstruction of the path of a mobile robot, using a stereoscopic camera system. The scale invariant feature transformation algorithm (SIFT), is used to process the images and locate keypoints in a 3D euclidian coordinates space. The path of a Pioneer mobile robot is estimated using the proposed technique.

Modal parameter identification in the time domain from a random excitation force

Abstract A time domain algorithm for calculating frequencies, damping ratios and modal shape information of a discrete dynamic system is presented. Input and output samples of the systems dynamic excitation and response signals are used to determine the parameters. The algorithm is based on Pronys solution to the complex exponential interpolation of the systems unit impulse response function. The problem of determining resonant frequencies of a steel beam is presented to illustrate a practical application of the algorithm.

Model parameter identification from the time impulse response of complex dynamic systems

Abstract A time domain method for calculating the absolute modal shapes, natural frequencies and damping ratios of a discrete dynamic system is presented. The digital records of the impulsive forcing and response of the system used in the method are obtained from sampling both signals at two different sampling rates. The method performs the analysis by calculating first the frequencies and damping ratios of the system from the records of its free response and then uses these calculated values together with the records of forcing and forced response to realise the absolute value of the modal shapes.

Dynamics Analysis of Reciprocating Compressor Crankshafts

Reciprocating compressors are used in several types of applications, such as gas distribution and refrigeration systems. One of the main causes of failure of these machines are high loads, which leads to large stresses in the crankshaft. In order to study effects of these loads, the dynamic model of a two-stage reciprocating compressor driven by asynchronous and synchronous electric motor was developed. This model is described by a rigid system composed of piston, rod, crosshead, connecting rod and crankshaft. Based on the Newton and Euler equations, the resistive loads acting on the crankshaft are calculated. The mass and the inertia of these components as well as the thermodynamic cycle are parameters for this model. The thermodynamic cycle assumed ideal gas for an isotropic system, a thermal insulation between the gas and cylinder and an instant behavior of the valves. These assumptions are reasonable because the simulated reciprocating compressor operates with compression ratio below two, resulting in the absence of large temperature rates and gradients. Since the resistive forces are cyclic, the torque applied by the drive will also oscillate cyclically, resulting in an oscillation of the angular velocity. The simulation results show that there are great amplitudes of torque due to the high pressure of the compressed gas. In the permanent regime, the loads had a good agreement with experimental data. The study proved to be important in the design and operation of the reciprocating compressor, in order to prevent failure of the crankshaft.