Abstract:
his project introduces navigation of robot through different pattern recognition techniques for localizing the robot in a controlled environment. Localization of robot is the ability for a robot to find out its position in the environment. Localization is the most basic and important aspect for a robot to navigate. There are two main techniques for localizing a robot: relative and absolute. Relative Localization depends on starting position of a robot; this technique is simple to use but prone errors. Other technique for localizing is absolute; it depends on observation of environmental features using different sensors. Absolute localizing is very difficult to apply but has fewer errors. For robust localizing combination of both techniques should be adopted. For absolute localization many sensors can be used to collect information about the environment and position of robot in environment i.e. GPS (Global Positioning System), Camera, Light Sensors etc.
To find position of robot through Vision, camera is used. Camera is mounted (fixed or installed) on the robot just like eyes are fixed on face. At different places camera sees and extracts different images (scenes); and algorithms are used to calculate or find current position of the robot from those images. This method is known as Self-Observing Algorithm (SOA). This project deals with making a robot that uses a Self-Observing Algorithm (SOA) technique for its navigation.
