Deep Reinforcement Learning Methods for Navigational Aids Bijan Fakhri 1 , Aaron Keech 3 , Joel Schlosser 3 , Ethan Brooks 3 , Hemanth Venkateswara 1(B ) , Sethuraman Panchanathan 1 , and Zsolt Kira 2,3 1 School of Computing, Informatics, and Decision Systems Engineering, Arizona State University, Tempe, AZ 85281, USA hkdv1@asu.edu 2 School of Interactive Computing, Georgia Tech, 85 5th St. NW, Atlanta, GA, USA 3 Georgia Tech Research Institute, 250 15th St. NW, Atlanta, GA, USA Abstract. Navigation is one of the most complex daily activities we engage in. Partly due to its complexity, navigational abilities are vulner- able to many conditions including Topographical Agnosia, Alzheimer’s Disease, and vision impairments. While navigation using solely vision remains a difficult problem in the field of assistive technology, emerging methods in Deep Reinforcement Learning and Computer Vision show promise in producing vision-based navigational aids for those with nav- igation impairments. To this effect, we introduce GraphMem, a Neural Computing approach to navigation tasks and compare it to several state of the art Neural Computing methods in a one-shot, 3D, first-person maze solving task. Comparing GraphMem to current methods in navi- gation tasks unveils insights into navigation and represents a first step towards employing these emerging techniques in navigational assistive technology. Keywords: Navigation · Assistive technology Reinforcement learning · Topographical agnosia 1 Introduction From navigating the rooms and hallways of one’s own home to navigating a large city, the cognitive functions involved in negotiating an environment to arrive at a predetermined destination are delicate, complex, and in many ways innate. Specialized components of the brain (head direction cells, place cells, grid cells, and border cells) have been shown to be integral to navigation [25]. Although the ability to navigate endows people with independence and self determination, many circumstances can lead to complications in navigation, and a surprising number of people experience such complications. The World Health Organization estimates that 253 million people worldwide have a vision impairment [4] and 21 to 25 million people have Alzheimer’s Disease worldwide [7], both of which are known to cause navigation issues [24] among many other conditions. c  Springer Nature Switzerland AG 2018 A. Basu and S. Berretti (Eds.): ICSM 2018, LNCS 11010, pp. 66–75, 2018. https://doi.org/10.1007/978-3-030-04375-9_6