Visual Homing in Dynamic Indoor Environments

Our dissertation concerns robotic navigation in dynamic in door environments using image-based visual homing. Image-based visual homing infe rs the direction to a goal location S from the navigator’s current location C usin g the similarity between panoramic imagesIS and IC captured at those locations. There are several ways to compute this similarity. One of the contributions of our dis sertation is to identify a robust image similarity measure – mutual image information – to use in dynamic indoor environments. We crafted novel methods to speed the com putation of mutual image information with both parallel and serial processors and demonstrated that these time-savers had little negative effect on homing success. I mage-based visual homing requires a homing agent to move so as to optimise the mutual im age information signal. As the mutual information signal is corrupted by sensor nois e we turned to the stochastic optimisation literature for appropriate optimisation algorithms. We tested a number of these algorithms in both simulated and real dynamic labor tory environments and found that gradient descent (with gradients computed by one -sid d differences) works best.