Logo
TENTATIVE program subject to modifications
TENTATIVE program
Confirmed Plenary Speakers @Chile
Jan Peters

Jan Peters
German Research Center for AI (DFKI), Research Department: SAIROL
Institute for Intelligent Autonomous Systems
TU Darmstadt
Title of the talk: Inductive Biases for Robot Reinforcement Learning
Abstract:
Autonomous robots that can assist humans in situations of daily life have been a long standing vision of robotics, artificial intelligence, and cognitive sciences. A first step towards this goal is to create robots that can learn tasks triggered by environmental context or higher level instruction. However, learning techniques have yet to live up to this promise as only few methods manage to scale to high-dimensional manipulator or humanoid robots. In this talk, we investigate a general framework suitable for learning motor skills in robotics which is based on the principles behind many analytical robotics approaches. To accomplish robot reinforcement learning learning from just few trials, the learning system can no longer explore all learn-able solutions but has to prioritize one solution over others – independent of the observed data. Such prioritization requires explicit or implicit assumptions, often called ‘induction biases’ in machine learning. Extrapolation to new robot learning tasks requires induction biases deeply rooted in general principles and domain knowledge from robotics, physics and control. Empirical evaluations on a several robot systems illustrate the effectiveness and applicability to learning control on an anthropomorphic robot arm. These robot motor skills range from toy examples (e.g., paddling a ball, ball-in-a-cup) to playing robot table tennis, juggling and manipulation of various objects.

Martin Adams

Martin Adams
Department of Electrical Engineering
Universidad de Chile
Title of the talk: Advancing State Estimation: Insights from Random Finite Sets
Abstract:
Autonomous navigation, mapping, and multi-target tracking are state estimation problems relevant to a wide range of applications. In such applications, estimation has very little meaning without a clear concept of estimation error. State-of-the-art solutions have traditionally been formulated using random vectors (RVs) in stochastic filtering, smoothing, or optimization based approaches, but fail to jointly minimize both spatial and detection errors, without the use of addon heuristics. In the Simultaneous Localization and Mapping (SLAM) problem, this usually takes the form of a back end solver and an independent front end, necessary for implementing the detection and measurement-to-state association heuristics. In contrast to RV-based approaches, the use of Random Finite Sets (RFS) yields a general measurement likelihood, which allows data association and map management routines to be a native part of the entire estimation approach, effectively combining the SLAM back and front ends into a single, joint estimation framework. RFS formulations also allow the joint consideration of detection and clutter statistics within the estimator and have recently attracted considerable research interest as well as deployment in commercial applications.
As well as justifying the general application of RFS frameworks, this presentation will focus on new RFS-based solutions to SLAM. To encompass the advantages of recent Maximum likelihood (ML) batch approaches, which use sparse matrix methods such as the g2o solver, it will be shown that the SLAM state can be modelled as a mixed distribution. This distribution jointly represents the vector-valued trajectory and the RFS-valued map, and is referred to as the Vector-Generalized Labeled Multi Bernoulli (V-GLMB) distribution. This yields hybrid RFS-RV-SLAM solutions, which yield competitive, and often superior, results to their RV-SLAM counterparts, while circumventing the need for fragile data association methods. A framework for solutions which combine the SLAM back and front ends into a single ML estimation framework will be demonstrated.

Darius Burschka

Darius Burschka
Robotics, Artificial Intelligence and Embedded Systems
Technical University of Munich
Title of the talk: Robust and Efficient Coupling of Perception to Actuation in Dynamic Environments
Abstract:
I will discuss the problem of efficient and robust information exchange between the perception and actuation modules in manipulation and mobile systems. While most current approaches use three-dimensional representations of the world as an interface to generate actions in a robot, this is not the native representation neither of the sensor nor of the motion controller and it requires calibration parameters to calculate. It is prone to errors and drifts making the accuracy of the system unreliable. Direct definitions of the tasks in the sensor space or some lower dimensional abstraction allows a robust operations of the systems in dynamic environments. I will show examples of non-metric task representation for navigation, path planning, obstacle avoidance and manipulation.

Paulo Drews-Jr

Paulo Drews-Jr
Center for Computational Sciences
Federal University of Rio Grande
Title of the talk: Robotics and its Challenges in the World of Artificial Intelligence
Abstract:
Robotics brings important challenges from both a scientific and technological point of view. The relevance and practical applications make efforts necessary. Within an extremely interdisciplinary context of robotics, autonomy brings important and relevant challenges and opportunities that converse with recent advances in artificial intelligence. Practical examples developed in projects by the Automation and Intelligent Robotics group – NAUTEC and the Embrapii Center for Robotics and Artificial Intelligence – iTEC, both from the Federal University of Rio Grande – FURG, Brazil, will also be discussed. These robotics examples includes aerial, underwater, hybrid aerial-underwater, logistics, agriculture, among other applications. 

Jorge Solis

Jorge Solis
Faculty of Health, Science and Technology
Department of Engineering and Physics
Karlstad University
Title of the talk: Challenges towards Industry 5.0 in production systems: from collaborative robots to energy efficiency
Abstract:
Today and tomorrow's industry requires modern technology where collaboration takes place between people, between people and machines and between machines. Machines can monitor themselves, analyze the results and autonomously optimize operating conditions and production. The result is higher efficiency and productivity. In this lecture, I will present an overview of the ongoing research projects at Karlstad University within the applications areas to ageing, energy, environment and education. In particular, the challenges within Industry 5.0 will be exemplified with some ongoing research projects that deals with a sustainable human-robot synergy in assembling tasks as well as adaptive battery energy storage in the food industry. Some possible extensions and potential applications will be outlined.

Streaming Plenary Speakers @ICRA 2025
Allison Okamura

Allison Okamura
Title of the talk: Rewired: The Interplay of Robots and Society

Confirmed Speakers @Chile
Rodrigo Verschae

Rodrigo Verschae
Robotics and Intelligent Systems Lab
Google Scholar
Universidad de O'Higgins
Title of the talk: Event-based Vision (Tentative)
Abstract:
Event-based vision has shown a growing success in recent years and is gaining increasing importance in robotics. This asynchronous sensor presents distinct advantages over traditional frame-based cameras, including low latency, high dynamic range, and low power consumption. In the current presentation, we will briefly review the basics of event-based cameras, and later report recent results in three main areas: depth estimation, face and gesture recognition, and motion analysis.

Julio Godoy

Julio Godoy
Google Scholar
Universidad de Concepcion
Title of the talk: Online Action Selection Methods for Multi-Agent Navigation
Abstract:
In multi-agent navigation, agents have to move from their start positions to their goal locations while avoiding collisions with other agents and any static element in the environment. Existing methods either compute the motion of each agent centrally or allow each agent to compute its own motion. Using a central controller limits the number of agents that can be controlled in real time, while using a local method produces motions that are optimal locally but do not account for the motions of the other agents, producing inefficient global motions when many agents move in a crowded space. Recently, a set of methods has been proposed that uses Deep Reinforcement Learning to learn navigation policies. However, these approaches require extensive training for each environment considered. In this talk, we present a set of online action selection methods that each agent uses to dynamically adapt its behavior to the local conditions. These approaches are highly scalable because each agent makes its own decisions on how to move, and do not require training. We validate the approaches experimentally, with multiple simulations in a variety of environments and with different numbers of agents, as well as with a small number of robots. When compared to other techniques, the proposed approaches produce motions that are more efficient and make better use of the space, allowing agents to reach their destinations faster.

Stefan Escaida

Stefan Escaida
AI & Robotics Group
Universidad de O'Higgins
Title of the talk: Model-Based Sensing for Soft Robotics with SOFA
Abstract:
In this talk, I will report on our work on model-based sensing for soft robots. Model-based sensing addresses the challenge of enabling tactile sensing and proprioception for soft robots in a principled way. Using inverse problem solving in an interactive solid mechanics simulation, the forces/deformations that best explain the observed sensor readings can be found. The first results in this line of research were obtained with soft pads, which are passive devices. Air chambers are embedded in these devices and changes in volume or pressure are measured by pneumatic sensors. Forces magnitudes and deformations due to external interactions could be estimated using SOFA. However, for estimating contact location, machine learning had to be employed. Therefore, as a follow-up, a multi-modal sensing approach was proposed: contact location is obtained additionally using soft capacitive touchpads. With them, interactions can completely be handled by the model-based approach, i.e. without relying machine learning. In more recent works, it was studied how these approaches can be applied to actuated devices. We have found that these results can be applied to the development of anatomical soft robots, that is, novel medical phantoms having advanced functionality as well as multi-segment soft manipulators.

Early-Carreer Researchers Session

Matias Mattamala University of Oxford, United Kingdom
Title of the talk: Vision-based Legged Robot Navigation for Field Applications
Abstract: Legged Robot Navigation for Field Applications" In this talk I'll discuss the challenges of deploying autonomous legged robots in the field. I'll present three main works that tackle perception, representation, and planning challenges when deploying quadruped robots for industrial inspection, underground exploration, and natural environments. I'll conclude the talk presenting a large-scale deployment of an autonomous legged robot for forest inventory applications, providing an outlook on open questions and future research avenues

Leandro Honorato de Souza Silva, Polytechnic School (University of Pernambuco) / Federal Institute of Education, Science and Technology of Paraíba (IFPB), Brazil
Title of the talk: Self-labeling Object Detection for Waste PCB Evaluation: Towards Sustainable Computer Vision Applications
Abstract: The increasing volume of electronic waste demands intelligent and scalable automation strategies. Printed Circuit Boards (PCBs), representing nearly 30% of all e-waste and containing a high concentration of valuable metals, are central to recycling efforts. However, the diversity of PCB structures and electronic components (ECs) poses challenges for automation. In this work, we propose a two-stage approach that bridges perception and action: a Self-labeling Electronic Component Detector and a robotic prototype for autonomous classification of PCBs. The first stage introduces the GEN Self-labeling Detector. This domain-adaptive, semi-supervised method uses iterative teacher-student training to detect ECs in images of waste PCBs. Tested with YOLOv5 and Faster R-CNN on the noisy FICS-PCB dataset, GEN achieves 67% mAP in the first generation, only 5% below fully supervised models. These detections are then used to characterize PCBs, estimating hazardous elements' presence and recycling's economic feasibility via the WPCB-EFA (Waste PCB Economic Feasibility Assessment) framework. The second stage translates perception into action by integrating this vision pipeline into a physical robotic system. The prototype consists of a conveyor-based classification machine, equipped with industrial automation components such as a PLC, HMI, and a servo-driven rack-and-pinion ejection system. All subsystems communicate through OPC protocols, ensuring closed-loop decision-making. This integrated solution advances self-supervised object detection in noisy, data-scarce domains. It is a model of how robotics and automation can drive sustainability in the circular economy.

Juan Pablo Vásconez Hurtado, Universidad Andrés Bello, Chile
Title of the talk: Human-Robot Interaction in Agriculture
Abstract: The agricultural industry has been significantly affected in recent years by labor shortages due to the migration of workers to urban areas. Additionally, the complexity of agricultural environments and tasks makes achieving high levels of autonomy challenging. Human-robot interaction (HRI) strategies can enhance efficiency and productivity in agricultural processes while reducing the workload of field workers, helping to mitigate the impact of labor shortages. These strategies aim to develop collaborative robotic systems that, through artificial intelligence (AI) algorithms, can interact more naturally, smoothly, and acceptably with humans during complex tasks such as harvesting, pruning, transport, and monitoring. HRI can play a key role in advancing sustainable and efficient agricultural technologies.

Rohit Singla, Universidad de O'Higgings, Chile,
Title of the talk: Robust Estimation in Teleoperated Robotics
Abstract: Time delays and parametric uncertainties remain key challenges in teleoperated robotics, affecting stability and performance. This work introduces a real-time estimation framework that addresses both. Time-varying delays are estimated by combining the Short-Time Fourier Transform with Taylor series expansion. Parametric uncertainties are handled via a block-processing approach integrated with Markov models and nonlinear filtering. The proposed framework is validated through simulations and experiments using a Geomagic Touch haptic interface. Results show improved tracking and estimation accuracy compared to standard methods. This approach enhances robustness and transparency in teleoperation, with relevance for remote surgery, space robotics, and hazardous environment applications.

Robert Guaman Rivera, Universidad de O'Higgings, Chile
Title of the talk: Automated Construction: Trajectory Control, Obstacle Avoidance and Automated Monitoring in 3D Printing Applications
Abstract: The workspace of a robotic platform on construction sites plays a crucial role in ensuring the progress of the work, maintaining task safety, preventing damage to the robot, and ensuring the quality of the printed profile. This talk will present the study of a trajectory tracking controller with obstacle avoidance, emphasising the importance of minimising printing errors to guarantee optimal structural quality. Despite the advancements in robotics and 3D printing, one of the main challenges remains improving quality monitoring techniques during and after printing. In this context, the study will explore using deep neural networks and computer vision to automatically detect objects and obstacles in the working environment, aiming to scale this approach and monitor the quality of the printed profile on construction sites more efficiently.

Felipe Inostroza, AMTC, Chile
Title of the talk: Use of Detection Statistics in Map Building and Localization (SLAM)
Abstract: Simultaneous Localization and Mapping (SLAM) is considered one of the fundamental problems in robotics. The most popular SLAM solutions rely on heuristic procedures, with varying levels of rigor, to handle detection uncertainty and data association. To incorporate this part of the problem into the Bayesian estimation, the theory of Random Finite Sets (RFS) can be used. RFSs allow modeling the map and its observations as sets, naturally incorporating detection uncertainty through a detection probability and a false alarm distribution. This presentation explains the Joint-Vector-Set SLAM algorithm, which uses RFSs to solve the SLAM problem through a graph-based optimization strategy. Competitive results with the state of the art are shown, using these detection statistics in the SLAM process, which reduces the number of heuristics required to solve the problem, at the cost of increased computational load.

Tutorials
Matias Mattamala

Matias Mattamala
Title of the talk: Effective design of graphics for (robotics) research

Felipe Inostroza

Felipe Inostroza
Title of the talk: Robotic Mapping and Simultaneous Localization and Mapping (SLAM) (slides)

Hands-on

Harold Valenzuela and Carolina Silva
Title of the Hands-on HO1a: Planar Subactuated Locomotion Mechanism for a Quadruped Robot Leg

Ulises Campodónico
Title of the Hands-on HO1b: Planetary reduction actuator design and control for Quadruped Robot


Ignacio Bugueno
Title of the Hands-on HO2: Gentle introduction to Event-based Robot Vision

Francisco Carcamo
Title of the Hands-on HO3: Escaneo rápido de objetos para modelamiento 3D

Ariel Zuniga and Nicolas Araya
Title of the Hands-on HO4: 3D modeling with implicit methods (DeepSDF and NeRF)