Dynamic Map Building Algorithm for Mobile Robotnavigation

This paper aims to remove the major limitation of contemporary machines which is the inherent lack of sensory capabilities and give machines the ability to sense its environment by vision. Vision is a useful sensor since it mimics the human sense of vision and helps in non-contact measurement. A machine such as a robot whose duties include navigation may be called upon to perform in many different environments. In such fluid environments it would be unlikely that a geometric model will always be available, and arranging and testing fiducial markers may be onerous and impossible. Here we propose a framework that exploit continuity of image sequences. The image based techniques allow accurate motion without a precise geometric model of the environment and give added speed and robustness due to the predictive information available. The machine is trained in the environment in which it is supposed to function andduring that period it collects the information of it's work-space as a sequence of images. The image sequences are then used to form a map-like directed graph. When put into operation the machine is able to identify its surroundings by a comparison done with the current image it is viewing and the image stored during the training phase. By identifying the change in position, orientation and size of the markers (which are any visual entity that is in some way distinctive and meaningful) in the object, the machine is able to make sense of its domain and able to respond appropriately. This algorithm when enhanced with more powerful image comparison algorithms such as Wavelet Based Image Indexing is very fast, reliable and efficient and also has the added advantage of lesser amounts of memory for storing the image sequences. This algorithm also has the capability to adjust with changing environments by modifying the existing map whenever a change occurs around its environment. Thus our algorithm has very wide use in fields of under-sea navigation, security robots, robot delivery systems etc.