Publicação
Calibration of odometry systems in robotic vehicles
| Resumo: | Accurate odometry is essential for autonomous navigation in robotic vehicles. Traditional encoder odometry and visual odometry are commonly used methods, each with distinct advantages and limitations. Encoder odometry, relying on wheel rotations, often suffers from cumulative errors and slippage. Visual odometry, which uses camera images to estimate movement, can be affected by environmental factors such as lighting and texture. This dissertation aims to fill a gap in the current state of the art by developing a novel methodology to calibrate robotic systems with erroneous odometry data. Building on the Atomic Transformations Optimization Method (ATOM) developed by the Laboratório de Automação e Robótica at the University of Aveiro, this work proposes enhancements to accommodate and correct odometry inaccuracies, by estimating the transformations provided by these systems. ATOM approaches the calibration problem as an extended optimization task, estimating the poses of both sensors and calibration patterns through a combination of indivisible geometric transformations, referred to as atomic transformations. Unlike pairwise calibration methods, ATOM employs a sensor-to-pattern paradigm, which significantly reduces the need for numerous error functions for each sensor pair, thereby generalizing the calibration process and making it applicable to a wide variety of robotic systems. The methodology is validated through extensive experiments on both a simulated robot (SOFTBOT) and a real robot (ZAU). The simulation results demonstrated significant improvements in calibration accuracy, confirming the efficacy of the proposed approach under controlled conditions. However, real-world experiments with ZAU revealed challenges due to unexpectedly large odometry errors, which lead to the incapability of calibrating the system. Despite these challenges, the findings contribute to advancing the field of robotic vehicles odometry calibration, providing a reliable approach for enhancing the performance of autonomous robotic systems. |
|---|---|
| Autores principais: | Silva, Bruno Filipe Amaral Vieira da |
| Assunto: | Extrinsic calibration Odometry Atomic transformations Optimization Mobile robots |
| Ano: | 2024 |
| País: | Portugal |
| Tipo de documento: | dissertação de mestrado |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade de Aveiro |
| Idioma: | inglês |
| Origem: | RIA - Repositório Institucional da Universidade de Aveiro |
| Resumo: | Accurate odometry is essential for autonomous navigation in robotic vehicles. Traditional encoder odometry and visual odometry are commonly used methods, each with distinct advantages and limitations. Encoder odometry, relying on wheel rotations, often suffers from cumulative errors and slippage. Visual odometry, which uses camera images to estimate movement, can be affected by environmental factors such as lighting and texture. This dissertation aims to fill a gap in the current state of the art by developing a novel methodology to calibrate robotic systems with erroneous odometry data. Building on the Atomic Transformations Optimization Method (ATOM) developed by the Laboratório de Automação e Robótica at the University of Aveiro, this work proposes enhancements to accommodate and correct odometry inaccuracies, by estimating the transformations provided by these systems. ATOM approaches the calibration problem as an extended optimization task, estimating the poses of both sensors and calibration patterns through a combination of indivisible geometric transformations, referred to as atomic transformations. Unlike pairwise calibration methods, ATOM employs a sensor-to-pattern paradigm, which significantly reduces the need for numerous error functions for each sensor pair, thereby generalizing the calibration process and making it applicable to a wide variety of robotic systems. The methodology is validated through extensive experiments on both a simulated robot (SOFTBOT) and a real robot (ZAU). The simulation results demonstrated significant improvements in calibration accuracy, confirming the efficacy of the proposed approach under controlled conditions. However, real-world experiments with ZAU revealed challenges due to unexpectedly large odometry errors, which lead to the incapability of calibrating the system. Despite these challenges, the findings contribute to advancing the field of robotic vehicles odometry calibration, providing a reliable approach for enhancing the performance of autonomous robotic systems. |
|---|