Estimating the Causal Impact of Social Navigation Signals in Corridor Scenarios

Francesco Baldo; Luca Castri; Gloria Beraldo; Nicola Bellotto

cs.ROOct 12, 2026

Estimating the Causal Impact of Social Navigation Signals in Corridor Scenarios

Authors: Francesco Baldo, Luca Castri, Gloria Beraldo, Nicola Bellotto

Organizations: Department of Information Engineering, University of Padua · Institute of Cognitive Sciences and Technologies, National Research Council of Italy

Abstract

Deploying robots in human-shared environments requires deciding not only how to navigate, but also when dedicated social signals are worth emitting. Classic observation-based inference can be misleading in such settings: confounding factors may bias the estimated relationship between a signal and the task outcome, leading to incorrect decisions. Causal inference provides a framework for disentangling these relationships by modelling the mechanisms linking observations, actions, and outcomes, and by estimating interventional effects after accounting for confounding paths. We study a human-robot interaction scenario in a corridor and use a causal model to estimate whether a social navigation signal causally improves transit success. We represent the interaction as a causal DAG involving robot action, pedestrian occupancy, available space, task outcome, and a controlled confounder (static obstacles). We compare the interventional estimate with the observational estimate obtained by conditioning only on the robot action, which does not explicitly model interventions. Our results show that accounting for confounding variables can reverse the conclusions drawn from naive observational analysis: the observational estimate wrongly suggests that the social signal is detrimental, whereas the causal estimate shows that it is beneficial. This discrepancy highlights how failing to account for confounding can lead to incorrect conclusions and demonstrates the importance of causal models for context-aware signalling decisions in social robot navigation.

Figures & tables

Figure 1: Motivating scenario: while humans naturally resolve encounters through subtle implicit cues, a robot must evaluate the causal impact of its intent signalling (e.g., LEDs blink) to determine whether emitting a signal genuinely prompts pedestrians to step aside and clear its path to the goal (green dot).

Panel of Figure 2 — Figure 2: The corridor scenario. Left: layout and waypoint graph, with the robot path from spawn to table , the observation waypoints where the decision is taken, and the conflict zone at the junction (shaded). Right: the same corridor rendered in Gazebo, with pedestrians drawn as red cylinders and the static obstacles that constitute the confounder $O$ for the current episode.

Panel of Figure 3 — Figure 3: Causal DAG of the corridor scenario. The main causal chain runs $P_{i}\rightarrow A\rightarrow P_{e}\rightarrow S\rightarrow T$ . The controlled confounder $O$ (dashed) enters at two points, opening the backdoor path $A\leftarrow O\rightarrow S\rightarrow T$ .

Figure 4: System architecture. Online (top): the robot plan drives the episode loop, coordinating the pedsim bridge, the obstacle policy and the recorder through ROS topics and parameters. Solid arrows indicate data flows, dashed arrows indicate parameters and control. Offline (bottom): the extraction pipeline turns the recorded bags into the per-episode causal dataset.

Figure 5: Timeline of an episode. Boxes are the phases of the protocol; below each, the causal variable recorded at that point. On

A=1

the robot signals, redirects the affected agents and waits about five seconds before reading

P_{e}

; on

A=0

the reading is immediate. The gap measured before the decision and the clearance measured before transit both include the static obstacles, whereas

P_{i}

and

P_{e}

are computed over pedestrians alone; the dashed arcs mark the two points where the confounder

O

enters, as in Figure 3 .

Figure 6: Comparison between naïve and causal estimate. (a) Success rate by action: within each level of

O

the signal does not lower success, and in the obstructed context it raises it from

0.10

0.26

, while the aggregate (black, dashed) falls by 21 points. (b) Composition of the two action groups: obstructed episodes are 16% of those without a signal and 50% of those with one, and that imbalance is what inverts the aggregate.

Social Robot Navigation

Social robot navigation studies how robots move safely and comfortably around people while respecting human behavior, preferences, and shared-space norms.

See all 81 papers

A Causal Approach to Predicting and Improving Human Perceptions of Social Navigation Robots
Mar 11, 2026Maximilian Diehl, Nathan Tsoi, Gustavo Chavez +2Human-Robot Interaction Social Robot Navigation
Faculty of Electrical Engineering, Chalmers University of Technology, SE-412 96 Gothenburg, Sweden. · The University of Texas at Austin, Austin, Texas, USA. · Yale University, New Haven, Connecticut, USA.
Can Causal Models Enhance Robot Navigation? Online Causal Adaptation for Real-Robot Navigation
Jun 14, 2026Zhitao Liang, Alex Mitrevski, Emmanuel Dean +1Robot Navigation
Division for Systems and Control, Department of Electrical Engineering, Chalmers University of Technology, Gothenburg, Sweden
Social Navigation for Tour-guide Robot
Oct 4, 2026Vikram Shree, Jose NinoSafe Robot Navigation Social Robot Navigation
Work done at the department of Mechanical and Aerospace Engineering, Cornell University, Ithaca NY (USA). · SiLC Technologies, Monrovia CA (USA). · Draper, Cambridge MA (USA).

Mar 11, 2026J/K move · Enter open · S save

A Causal Approach to Predicting and Improving Human Perceptions of Social Navigation Robots

Maximilian Diehl, Nathan Tsoi, Gustavo Chavez, Karinne Ramirez-Amaro+1

Faculty of Electrical Engineering, Chalmers University of Technology, SE-412 96 Gothenburg, Sweden. · The University of Texas at Austin, Austin, Texas, USA. · Yale University, New Haven, Connecticut, USA.

Human-Robot Interaction Social Robot Navigation Explainable Motion Planning Causal Interpretability

Open paper AskPDF arXiv

As mobile robots are increasingly deployed in human environments, enabling them to predict how people perceive them is critical for socially adaptable navigation. Predicting perceptions is challenging for two main reasons: (1) HRI prediction models must learn from limited data, and (2) the obtained models must be interpretable to enable safe and effective interactions. Interpretability is particularly important when a robot is perceived as incompetent (e.g., when the robot suddenly stops or rotates away from the goal), as it allows the robot to explain its reasoning and identify controllable factors to improve performance, requiring causal rather than associative reasoning. To address these challenges, we propose a Causal Bayesian Network designed to predict how people perceive a mobile robot's competence and how they interpret its intent during navigation. Additionally, we introduce a novel method to improve perceived robot competence employing a combinatorial search, guided by the proposed causal model, to identify better navigation behaviors. Our method enhances interpretability and generates counterfactual robot motions while achieving comparable or superior predictive performance to state-of-the-art methods, reaching an F1-score of 0.78 and 0.75 for competence and intention on a binary scale. To further assess our method's ability to improve the perceived robot competence, we conducted an online evaluation in which users rated robot behaviors on a 5-point Likert scale. Our method statistically significantly increased the perceived competence of low-competent robot behavior by 83%.

Abstract

Social Robot Navigation

Explore similar work

A Causal Approach to Predicting and Improving Human Perceptions of Social Navigation Robots

Can Causal Models Enhance Robot Navigation? Online Causal Adaptation for Real-Robot Navigation

Social Navigation for Tour-guide Robot