<p>The concurrence of social robots in indoor environments populated by people is increasing recently in many fields, such as medicine, education, and industry. In such scenarios, social robots must take into account the movements of humans and what places they like or are used to walk, in addition to social aspects, to move as natural and close as people do. Most existing approaches solve robot navigation by considering only collision avoidance and ignoring inference movements, while studies aiming to implement social navigation use approaches focused on scenarios involving a single human being, such as the social push and social force models. To overcome these limitations, we propose an approach based on optical flow analysis to detect movements in an area populated by people and subsequently perform path planning to create an optimal route for a social robot from a global perspective. The robot’s path planning considers social restrictions in terms of following the same directions and addresses of people. To evaluate the performance of this work, we carried out a series of experiments to analyze people movements and generate maps with planned routes for the social robot in various scenarios, from a set of videos of people walking in different directions in real environments, as well as in a simulated environment with moving avatars. Results show that our solution is able to recognize the busiest areas and directions of people and accordingly create routes for the robot that imitate such as movements.</p>