Despite a surge in robotics research dedicated to inferring and understanding human intent, a universally accepted definition remains elusive since existing works often equate human intention with specific task-related goals. This article seeks to address this gap by examining the multifaceted nature of intention. Drawing on insights from psychology, it attempts to consolidate a definition of intention into a comprehensible framework for a broader audience. The article classifies different types of intention based on psychological and communication studies, offering guidance to researchers shifting from pure technical enhancements to a more human-centric perspective in robotics. It then demonstrates how various robotics studies can be aligned with these intention categories. Finally, through in-depth analyses of collaborative search and object transport use cases, the article underscores the significance of considering the diverse facets of human intention.
Figures & tables
Figure 1: Representation of the evolution of the brain potential measured before and after the volunteer performed an action. W marks the moment indicated by the volunteer as the one at which they decide to act. RP marks the moment at which the detected potential exceeds the average threshold that is considered an activation. Figure from Haggard (2005) .
Figure 2: Relationship between desires, beliefs and intention. The person’s beliefs (based on their knowledge, experience, memory, etc.) cause them to end up choosing a particular desire which is the one that can materialize an intention.
Figure 3: A model of the relationship between the intention and intentionality concepts. To have an intention is a pre-requisite to have an intentionality. Figure based on concepts presented in Malle and Knobe (1997) .
Criterion
Type of intention
Type of goal
Goal oriented – Implementation
Communication
Implicit – Explicit
Consciousness
Conscious – Unconscious
Number of agents
Individual – Collective
Temporality
Short-term – Long-term
Table 1: Summary of the different ways of classifying human intention according to the literature
Figure 4: Representation of the main relationships between intention categories.
Figure 5: Outline of the taxonomy presented for human intent.
Figure 6: Different perspectives considered about who is the agent detecting the intention.
Figure 7: Handover example. Using a trajectory predictor, the robot can predict how much the human is going to approach (implementation intention) and even the desired Object Transfer Point (goal-oriented intention) that the robot should reach.
Figure 8: Human-robot collaborative object transportation example. Human and robot transporting a bar. Force sensor attached to the robot’s wrist with which it can the human’s implicit intention and buttons on the handle of the object for explicit intention communication.
Figure 9: Human-robot pair navigating through a narrow door. Top left : Human-Robot pair navigating side-by-side until they reach a door. Top right : Human takes the lead but keeps taking into consideration the robot. Robot moves its left arm to allow the human to move forward. Bottom left : Human waits for the robot to finish passing through the door. Robot rolled up its right arm to reduce its footprint and moved its left arm to make the human to be in front of it. Bottom right : Human-Robot pair recover their side-by-side configuration.
Figure 10: Collaborative assembly example. Robot sees that the human is holding a screw and moving his arm towards a screwdriver. It predicts that in the human is going to take the screwdriver in the short term and use it to tighten the screw in the long term. Both predictions allows it to estimate how many time it has to take the next bar and when the shared space will be free.
Figure 11: Human-Robot collaborative search example. The human is holding an smartphone through which he can communicate explicitly with the robot in a bidirectional manner. The robot can also keep track of the human movement while it can see him.
As robots transition from performing repetitive tasks to collaborating with humans, understanding human intent becomes crucial to effective interaction. Anticipation enables robots to predict human actions, while proactivity allows them to take initiative and guide human behavior toward optimal outcomes. Although research has largely focused on how robots infer and respond to human intentions, less attention has been paid to how robots communicate their own intent. This paper introduces visual proactivity, a novel, simple yet effective approach that enables robots to communicate their intentions through visual feedback, influencing human behavior and enhancing transparency and fluency. We develop and evaluate proactive robotic behaviors in a human-to-robot handover scenario, where a user study validates human perception of reactive, anticipatory, and proactive behaviors. The results demonstrate that effective visual proactivity fosters better alignment and coordination, paving the way for more intuitive human-robot collaboration.
Valerio Bo, Edison Bejarano, Anaís Garrell +1
Institut de Robotica y Informatica Industrial (CSIC-UPC) and Universitat Politecnica de Catalunya - BarcelonaTech (UPC), Barcelona, Spain
Improving the effectiveness of human-robot interaction requires social robots to accurately infer human goals through robust intention understanding. This challenge is particularly critical in multimodal settings, where agents must integrate heterogeneous signals including text, visual cues to form a coherent interpretation of user intent. This paper presents IntentVLM, a novel two-stage video-language framework designed for open-vocabulary human intention recognition. The approach is inspired by forward-inverse modeling in cognitive science by decomposing intention understanding into goal candidate generation followed by structured inference through selection, effectively reducing hallucinations in latent reasoning. Evaluated on the IntentQA and Inst-IT Bench datasets, IntentVLM achieves state-of-the-art results with up to 80% accuracy, notably surpassing the baseline performance by 30% and matches human performance. Our findings demonstrate that this structured reasoning approach enhances open-vocabulary intention understanding without catastrophic forgetting, offering a robust foundation for human-centered robotics.
As robots become increasingly integrated into everyday environments, intuitive communication paradigms such as natural language and end-user programming have become indispensable for specifying autonomous robot behavior. However, these mechanisms are ineffective at fully capturing user intent: natural language is imprecise and ambiguous, whereas end-user programming can be overly specific. As a result, understanding what users truly mean when they interact with robots remains a central challenge for human-AI communication systems. To address this issue, we propose the Distill approach for human-robot communication interfaces. Given a task specification provided by the user, Distill (1) removes unnecessary steps; (2) generalizes the meaning behind individual steps; and (3) relaxes ordering constraints between steps. We implemented Distill on a web interface and, through a crowdsourcing study, demonstrated its ability to elicit and refine user intent from initial task specifications.
Ting Li, David Porfirio
Computer Science George Mason University Fairfax, Virginia, USA