Legible and Proactive Robot Planning for Prosocial Human-Robot Interactions

TL;DR

The paper tackles safe and fluent human-robot interactions in dense environments by proposing legible and proactive robot planning. It introduces a robot-centric objective enhanced by a markup factor $\mu>1$ to encourage early, identifiable actions and an inconvenience budget to promote equitable collision avoidance, implemented within an MPC framework using iterated best response (IBR). The key contributions include (i) the inconvenience budget constraint, (ii) the markup term in the cost to incentivize legible behavior, (iii) an IBR-based planner with real-time feasibility, and (iv) empirical evidence that the approach yields safer, more fluent, and more prosocial interactions compared to baselines. The work demonstrates that simple optimizations to standard trajectory planning can produce robust, cooperative human-robot interactions in real-time, with potential for scaling to more agents and environments.

Abstract

Humans have a remarkable ability to fluently engage in joint collision avoidance in crowded navigation tasks despite the complexities and uncertainties inherent in human behavior. Underlying these interactions is a mutual understanding that (i) individuals are prosocial, that is, there is equitable responsibility in avoiding collisions, and (ii) individuals should behave legibly, that is, move in a way that clearly conveys their intent to reduce ambiguity in how they intend to avoid others. Toward building robots that can safely and seamlessly interact with humans, we propose a general robot trajectory planning framework for synthesizing legible and proactive behaviors and demonstrate that our robot planner naturally leads to prosocial interactions. Specifically, we introduce the notion of a markup factor to incentivize legible and proactive behaviors and an inconvenience budget constraint to ensure equitable collision avoidance responsibility. We evaluate our approach against well-established multi-agent planning algorithms and show that using our approach produces safe, fluent, and prosocial interactions. We demonstrate the real-time feasibility of our approach with human-in-the-loop simulations. Project page can be found at https://uw-ctrl.github.io/phri/.

Figures (6)

• Figure 1: Our approach enables robots to interact legibly and proactively, providing oncoming humans with sufficient time to engage cooperatively in collision avoidance. Our experiments show that this leads to fluent, safe, and prosocial interactions.
• Figure 2: Human and robot trajectories in a head-on scenario. In our approach (green box) the robot legibly and proactively conveys its intent to pass, thus providing the human sufficient time to prepare to pass safely. In other approaches (blue boxes), the robot (SFM) does not convey its intention clearly and early and causes the human to swerve significantly, (HJ) freezes on the spot, (OC) takes on more collision avoidance responsibility than necessary, or (vIBR) confuses the human.
• Figure 3: Our approach can easily account for additional wall constraints and multiple dynamic agents.
• Figure 4: Snapshot of human-in-the-loop simulation experiment. Human interacts with the robot in first-person view.
• Figure 5: (a) Analysis of heading and heading rate to illustrate the legibility and proactivity of our proposed approach. (b) Statistics of the minimum distance between human and robot to evaluate the safety performance of our approach.