Improvement of odometry for omnidirectional vehicle using optical flow information

نویسندگان

  • Keiji Nagatani
  • Satoshi Tachibana
  • Makoto Sofne
  • Yutaka Tanaka
چکیده

Our research goal is to realize a robust navigation in indoor and outdoor environment for autonomous vehicle. An omnidirectional vehicle driven by four Mecanum wheels was chosen for our research platform. Mecanum wheel has 16 tilted rollers (45 degrees against the direction of wheel rotation) around the wheel, so the vehicle moves omnidirectionally by controlling these wheels independently. However, it has a disadvantage of odometry because of wheels’ slippage. Particularly, when the robot moves laterally, same wheels’ rotations generate different traveling distance according to a friction of a ground surface. To cope with the problem, we estimate robot’s position by detecting optical flow of ground image using vision sensor (visual dead-reckoning). The estimation method is inaccurate comparing with odometry, but it is independent from friction of ground surface. Therefore, the estimated vehicle position can be improved by fusing odometry and visual dead-reckoning based on maximum likelihood technique. This paper describes an odometry method and a visual dead-reckoning method for omnidirectional vehicle, and fusion technique to improve the estimated position of the vehicle. Finally, experimental results support above technique.

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Rotation Estimation of a Complete Omnidirectional Camera using Line Features

Rotation estimation with cameras is important for flying robot navigation. Usually, it is performed using a single camera by keeping track of some strong features such as corners or lines within the environment and observing their motion. However, with a normal camera, the field of view is quite limited and this can lead to an error because information is obtained from a small section of the sc...

متن کامل

Visual Odometry Based on Omnidirectional Images

We propose a method for visual odometry using optical flow with a single omnidirectional (catadioptric) camera. We show how omnidirectional images can be used to perform optical flow, discussing the basis of optical flow and some restrictions needed to it and how unwarp these images. In special we describe how to unwarp omnidirectional images to Bird’s eye view, that correspond to scaled orthog...

متن کامل

Correcting Vehicle Heading in Visual Odometry by Using Image Appearance

This paper describes a visual odometry algorithm for onroad vehicles under planar motion assumption. The algorithm uses as only input images taken by a single omnidirectional camera mounted on the roof of the vehicle. To achieve accurate motion estimation results, we decouple the rotation and translation estimation. In particular, we use a feature based ground plane tracker to estimate the tran...

متن کامل

A Robust Method for Ego-Motion Estimation in Urban Environment Using Stereo Camera

Visual odometry estimates the ego-motion of an agent (e.g., vehicle and robot) using image information and is a key component for autonomous vehicles and robotics. This paper proposes a robust and precise method for estimating the 6-DoF ego-motion, using a stereo rig with optical flow analysis. An objective function fitted with a set of feature points is created by establishing the mathematical...

متن کامل

One-Point-plus for 5-DOF Localization of Vehicle-Mounted Omnidirectional Camera in Long-Range Motion

This paper proposes a method for localization of vehicle using one point plus an edge matching region of monocular vision in wide urban environments. The five degree of freedom (5-DoF) localization estimated by monocular omnidirectional camera improves the planar motion assumption in most of conventional researches. In recent year, the car-like motion model with planar motion is often investiga...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2000