Collision Avoidance in Dynamic Environment by Estimation of Velocity and Location of Object by Robot using Parallax

Collision Avoidance in Dynamic Environment by Estimation of Velocity and Location of Object by Robot using Parallax

Ajay Kumar Rai (ABV-Indian Institute of Information Technology, Gwalior, India) and Ritu Tiwari (ABV-Indian Institute of Information Technology, Gwalior, India)
Copyright: © 2015 |Pages: 13
DOI: 10.4018/IJRAT.2015070104
OnDemand PDF Download:
List Price: $37.50
10% Discount:-$3.75


Collision avoidance is a challenging problem in robot navigation. There are two type of collision avoidance approach i.e. path planning algorithm and control algorithm. Control algorithm proven very useful for dynamic environment. In control method robot use sensing to close a feedback path and interact with the environment. This sensing feedback loop can be completed by using Camera in robot.Visual parallax is very useful for generation of 3D information of environment for robotic system. Here the authors describe the use of visual parallax in humanoid robotic system for estimation of velocity and location of object that present in environment. The location and velocity of object gives the useful information to plan a collision free path by robot in that environment. Accuracy, robustness, and applicability of this method wastest by simulation. The authors show that this approach gives an accurate and robust estimation of velocity and a distance of object. This information simplifies the difficult task of collision avoidance in dynamic environment.
Article Preview

1. Introduction And Motivation

When the robot execute autonomous patrol task, it becomes very important for robot to avoid collision with the object present in surrounding. Working efficiencyof an autonomous robot heavily depend on its ability to avoid collision with the objects present in environment. Collision avoidance for robot is very ubiquitous and challenging problem. It becomes more dangerous if the environment becomes dynamic. Several research discussed the method to make collision free movements of robot in real time (J.-C. Latombe(1991); J. T. Schwartz and C. K. Yap, Eds. (1987); J. T. Schwartz and M. Sharir (1988); K. Fujimura(1991); Y. K. Hwang and N. Ahuja(1992)).

There are two type of collision avoidance algorithm i.e. Path Planning and control algorithm [Dippold, Amanda, Lili Ma& Naira Hovakimyan, (2009); S. Quinlan & O. Khatib, (1993)]. In path planning the robot has prior knowledge of environment in which it is moving so it follows a compatible path based on knowledge of environment. But applicability of path planning approach is limited in dynamic environment. It requires too much amount of computational complexity, but gives much more useful information in case that robot is moving in static environment or environment has larger area.

Contrast to path planning algorithms, Control method have been enable robot to use sensor to close a feedback loop and interact with the environment for real time collision avoidance (A. Bowling and O. Khatib,1997).In control algorithm main objective is to get the specific knowledge so that the collision being avoided [De Luca & Alessandro]. Control Algorithm is more suitable for collision free movement in dynamic environment. A control method is composed essentially by three steps [Borenstein et al, 1996]:

  • 1.

    Sensing the environment;

  • 2.

    Developing the knowledge of the environment;

  • 3.

    Robot Control.

Different sensors are used to sense the environment. A wide variety of sensors are available. It includes ultrasonic sensor, laser range finder, radar, sonar etc[Y. M. KlimKov,1996; M. A. Garcia and A. Solanas, 2004; Kal-Tai Song & Wen-Hui Tang,1996; K. Osugi, K. Miyauchi& N. Furui,1999; .-T. Shin, 2000]. Apart from their respective advantages most of them are not suitable for sensing in small distance in environment. In contrast visual sensor provides good resolution data, better result at faster rates than previously mention sensor. The visual sensor (camera) has the ability to directly capture the 3D information of the world.

To move robot in a dynamic environment, object localization is prior condition for avoid collision. Hence the using camera for knowing the environment directly contributes to a safe operation of a robot. In fact use of camera in robot to sense environment is inspired from nature. Most of the animals, birds and insect [M. V. Srinivasan, (1992)] employ vision systems for getting the knowledge about surroundings. Humans exploit its two eyes to do the same which are sensors resemble with cameras.

Complete Article List

Search this Journal:
Open Access Articles
Volume 6: 2 Issues (2018)
Volume 5: 2 Issues (2017)
Volume 4: 2 Issues (2016)
Volume 3: 2 Issues (2015)
Volume 2: 2 Issues (2014)
Volume 1: 2 Issues (2013)
View Complete Journal Contents Listing