A Protocol for Validating Social Navigation Policies
Abstract
Enabling socially acceptable behavior for situated agents is a major goal of recent robotics research. Robots not only need to operate safely around humans, but also in a way that their behavior complies to social expectations. A key challenge for developing socially-compliant polices is to measure the performance of the behavior they generate for the robot. Due to the enormous complexity of social behavior, metrics for measuring the success and failure of algorithms are difficult to obtain. In this paper, we introduce a protocol to establish a social navigation benchmark that focuses on defining a set of canonical social navigation scenarios and an in-situ metric for social scenarios based on questionnaires. Our protocol can be replicated verbatim or it can be used to define a social navigation benchmark for novel scenarios. Our goal is to introduce a protocol for benchmarking social scenarios that is homogeneous and comparable.