Off-the-shelf vision based mobile robot sensing

The goal of this research is to enable a mobile robot using vision sensing technology to navigate in both outdoor and indoor environments, following such as a specified path, following a specified person, and detecting doorways to enter a room. The focus is upon real-time algorithm using off-the-shelf cameras.First, a simple approach for vision-based path following for a mobile robot is presented. Based upon a novel concept called the funnel lane, the coordinates of feature points during the replay phase are compared with those obtained during the teaching phase in order to determine the turning direction. The system requires a single off-the-shelf, forward-looking camera with no calibration (either external or internal, including lens distortion). The algorithm is qualitative in nature, requiring no map of the environment, no image Jacobian, no homography, no fundamental matrix, and no assumption about a flat ground plane.Second, by fusing motion and stereo information, Binocular Sparse Feature Segmentation (BSFS) algorithm is proposed for vision-based person following with a mobile robot. BSFS uses Lucas-Kanade feature detection and matching in order to determine the location of the person in the image and thereby control the robot. Matching is performed between two images of a stereo pair, as well as between successive video frames. We use the Random Sample Consensus (RANSAC) scheme for segmenting the sparse disparity map and estimating the motion models of the person and background. This system is able to reliably follow a person in complex dynamic, cluttered environments in real time.Third, a vision-based door detection algorithm is developed based on Adaboost and Data-Driven Markov Chain Monte Carlo (DDMCMC). Doors are important landmarks for indoor mobile robot navigation. Models of doors utilizing a variety of features, including color, texture, and intensity edges are presented. The Bayesian formulations are constructed and a Markov chain is designed to sample proposals. The features are combined using Adaboost to ensure optimal linear weighting. Doors are detected based on the idea of maximizing a posterior probability (MAP). Data-Driven techniques are used to compute importance proposal probabilities, which drive the Markov Chain dynamics and achieve speedup in comparison to the traditional jump diffusion methods.