Robotic agents can continuously provide feedback to people based on their behaviors. For instance, a robot swarm can remind a group of people to respect social distancing guidelines during a pandemic or discourage unwanted behavior such as littering. However, developing a swarm robot to operate in realistic situations is challenging: a robot requires significant resources to operate in the real world, yet costs need to be kept low to produce the robots en masse. To develop a swarm robot for encouraging social distancing, we therefore, compare the performance of different robotic navigation algorithms and various vision sensors and algorithms for detecting social distances breaches.