Gianluca Bardaro

Cristiano Battistini, Riccardo Andrea Izzo, Gianluca Bardaro et al.

Large and small language models have been widely used for robotic task planning. At the same time, vision-language models (VLMs) have successfully tackled problems such as image captioning, scene understanding, and visual question answering. In this work, we combine these two approaches by deploying a compact, open-source multimodal model to generate behavior trees for robotic task planning. The main obstacle to achieving this goal is the lack of an existing dataset that links visual observations and instructions to executable behavior trees. We propose a method to construct such a dataset starting from existing robotic episodes (i.e., Open X-Embodiment), in which a large model serves as a teacher in a multi-stage generation pipeline. We use this dataset to fine-tune VLMs ranging from 500M to 4B parameters via parameter-efficient fine-tuning (PEFT). The generated behavior trees, compatible with the BehaviorTree.CPP library, are evaluated both offline, using structural and lexical metrics, and online through the execution of household tasks in a state-of-the-art embodied simulator. Our results demonstrate that our fine-tuned 4B-parameter VLM approaches the performance of state-of-the-art closed-source models, achieving an 87\% success rate while requiring only a fraction of the computational resources.

Riccardo Andrea Izzo, Gianluca Bardaro, Matteo Matteucci

Current research on Vision-Language-Action (VLA) models predominantly focuses on enhancing generalization through established reasoning techniques. While effective, these improvements invariably increase computational complexity and inference latency. Furthermore, these mechanisms are typically applied indiscriminately, resulting in the inefficient allocation of resources for trivial tasks while simultaneously failing to provide the uncertainty estimation necessary to prevent catastrophic failure on out-of-distribution tasks. Inspired by human cognition, we propose an adaptive framework that dynamically routes VLA execution based on the complexity of the perceived state. Our approach transforms the VLA's vision-language backbone into an active detection tool by projecting latent embeddings into an ensemble of parametric and non-parametric estimators. This allows the system to execute known tasks immediately (Act), reason about ambiguous scenarios (Think), and preemptively halt execution when encountering significant physical or semantic anomalies (Abstain). In our empirical analysis, we observe a phenomenon where visual embeddings alone are superior for inferring task complexity due to the semantic invariance of language. Evaluated on the LIBERO and LIBERO-PRO benchmarks as well as on a real robot, our vision-only configuration achieves 80% F1-Score using as little as 5% of training data, establishing itself as a reliable and efficient task complexity detector.

Daniel Yezid Guarnizo Orjuela, Leonardo Scappatura, Veronica Di Gennaro et al.

Vision-Language-Action (VLA) models have emerged as a dominant paradigm for generalist robotic manipulation, unifying perception and control within a single end-to-end architecture. However, despite their success in controlled environments, reliable real-world deployment is severely hindered by their fragility to visual disturbances. While existing literature extensively addresses physical occlusions caused by scene geometry, a critical mode remains largely unexplored: image corruptions. These sensor-level artifacts, ranging from electronic noise and dead pixels to lens contaminants, directly compromise the integrity of the visual signal prior to interpretation. In this work, we quantify this vulnerability, demonstrating that state-of-the-art VLAs such as $π_{0.5}$ and SmolVLA, suffer catastrophic performance degradation, dropping from 90\% success rates to as low as 2\%, under common signal artifacts. To mitigate this, we introduce the Corruption Restoration Transformer (CRT), a plug-and-play and model-agnostic vision transformer designed to immunize VLA models against sensor disturbances. Leveraging an adversarial training objective, CRT restores clean observations from corrupted inputs without requiring computationally expensive fine-tuning of the underlying model. Extensive experiments across the LIBERO and Meta-World benchmarks demonstrate that CRT effectively recovers lost performance, enabling VLAs to maintain near-baseline success rates, even under severe visual corruption.

Juncen Long, Gianluca Bardaro, Simone Mentasti et al.

Pedestrian trajectory prediction is important in the research of mobile robot navigation in environments with pedestrians. Most pedestrian trajectory prediction algorithms require the input historical trajectories to be complete. If a pedestrian is unobservable in any frame in the past, then its historical trajectory become incomplete, the algorithm will not predict its future trajectory. To address this limitation, we propose the STGN-IT, a spatio-temporal graph network allowing incomplete trajectory input, which can predict the future trajectories of pedestrians with incomplete historical trajectories. STGN-IT uses the spatio-temporal graph with an additional encoding method to represent the historical trajectories and observation states of pedestrians. Moreover, STGN-IT introduces static obstacles in the environment that may affect the future trajectories as nodes to further improve the prediction accuracy. A clustering algorithm is also applied in the construction of spatio-temporal graphs. Experiments on public datasets show that STGN-IT outperforms state of the art algorithms on these metrics.

Agnese Chiatti, Gianluca Bardaro, Enrico Motta et al.

Service robots are expected to reliably make sense of complex, fast-changing environments. From a cognitive standpoint, they need the appropriate reasoning capabilities and background knowledge required to exhibit human-like Visual Intelligence. In particular, our prior work has shown that the ability to reason about spatial relations between objects in the world is a key requirement for the development of Visually Intelligent Agents. In this paper, we present a framework for commonsense spatial reasoning which is tailored to real-world robotic applications. Differently from prior approaches to qualitative spatial reasoning, the proposed framework is robust to variations in the robot's viewpoint and object orientation. The spatial relations in the proposed framework are also mapped to the types of commonsense predicates used to describe typical object configurations in English. In addition, we also show how this formally-defined framework can be implemented in a concrete spatial database.

Agnese Chiatti, Enrico Motta, Enrico Daga et al.

Service robots can help with many of our daily tasks, especially in those cases where it is inconvenient or unsafe for us to intervene: e.g., under extreme weather conditions or when social distance needs to be maintained. However, before we can successfully delegate complex tasks to robots, we need to enhance their ability to make sense of dynamic, real world environments. In this context, the first prerequisite to improving the Visual Intelligence of a robot is building robust and reliable object recognition systems. While object recognition solutions are traditionally based on Machine Learning methods, augmenting them with knowledge based reasoners has been shown to improve their performance. In particular, based on our prior work on identifying the epistemic requirements of Visual Intelligence, we hypothesise that knowledge of the typical size of objects could significantly improve the accuracy of an object recognition system. To verify this hypothesis, in this paper we present an approach to integrating knowledge about object sizes in a ML based architecture. Our experiments in a real world robotic scenario show that this combined approach ensures a significant performance increase over state of the art Machine Learning methods.