Formalizing Trajectories in Human-Robot Encounters via Probabilistic STL Inference

Abstract

In this paper, we are interested in formalizing human trajectories in human-robot encounters. We consider a particular case where a human and a robot walk towards each other. A question that arises is whether, when, and how humans will deviate from their trajectory to avoid a collision. These human trajectories can then be used to generate socially acceptable robot trajectories. To model these trajectories, we propose a data-driven algorithm to extract a formal specification expressed in Signal Temporal Logic with probabilistic predicates. We evaluated our method on trajectories collected through an online study where participants had to avoid colliding with a robot in a shared environment. Further, we demonstrate that probabilistic STL is a suitable formalism to depict human behavior, choices and preferences in specific scenarios of social navigation.

Inference of Multi-Class STL Specifications for Multi-Label Human-Robot Encounters

“…To avoid a collision, users had to deviate from their trajectory. The goal was to learn probabilistic STL specifications [5] from these humanrobot trajectories to model distributions of human behaviors when facing the incoming robot.…”

“…STL defines specifications over multi-dimensional continuous signals and can, e.g., be used to depict properties of trajectories. Since STL allows rich expressiveness and resemblance to natural language [3], it has recently been used in human-robot interaction applications [4], [5], and human-robot encounters. Indeed, there is a growing need to depict human-robot interactions through formal and rigorous techniques to provide guarantees on the functioning of robotic systems, as well as their interactions with humans and perceived safety [6].…”

Classification of Driving Behaviors Using STL Formulas: A Comparative Study

“…On the other hand, encoding safety rules in temporal logic is an eminent method for safety-critical applications [20], [21]. Specifications in temporal logic can be used for controller synthesis [6], [7], [21], motion planning [8], [22], [23] and learning applications [9]- [11], [24]- [28] in many autonomous systems. In particular, works in [9], [10], [24], [25], [29] try to infer a temporal logic formula for classification from the data.…”