Javier Laserna Moratalla

We propose a branch-and-bound algorithm for robust rigid registration of two point clouds in the presence of a large number of outlier correspondences. For this purpose, we consider a maximum consensus formulation of the registration problem and reformulate it as a (large) maximal clique search in a correspondence graph, where a clique represents a complete rigid transformation. Specifically, we use a maximum clique algorithm to enumerate large maximal cliques and a fitness procedure that evaluates each clique by solving a least-squares optimization problem. The main advantages of our approach are i) it is possible to exploit the cutting-edge optimization techniques employed by current exact maximum clique algorithms, such as partial maximum satisfiability-based bounds, branching by partitioning or the use of bitstrings, etc.; ii) the correspondence graphs are expected to be sparse in real problems (confirmed empirically in our tests), and, consequently, the maximum clique problem is expected to be easy; iii) it is possible to have a good control of suboptimality with a k-nearest neighbor analysis that determines the size of the correspondence graph as a function of k. The new algorithm is called CliReg and has been implemented in C++. To evaluate CliReg, we have carried out extensive tests both on synthetic and real public datasets. The results show that CliReg clearly dominates the state of the art (e.g., RANSAC, FGR, and TEASER++) in terms of robustness, with a running time comparable to TEASER++ and RANSAC. In addition, we have implemented a fast variant called CliRegMutual that performs similarly to the fastest heuristic FGR.

The application of Deep Neural Networks (DNN) for reinforcement learning has proven effective in solving complex problems, such as playing video games or training robots to perform human tasks. Training based on reinforcement implies the continuous interaction of the agent powered by the DNN and the environment, vanishing the typical separation between the training and inference stages in deep learning. However, the high memory and accuracy requirements of gradient-based training algorithms prevent using FPGAs for these applications. As an alternative, this work demonstrates the feasibility of using Evolutionary Algorithms (EA) for training DNNs and their usage in reinforcement learning scenarios. Unlike backpropagation, EA-based training of neural networks, referred to as neuroevolution, can be effectively implemented on FPGAs. Moreover, this paper shows how the inherent parallelism of EAs can be effectively exploited in multi-FPGA scenarios to accelerate the learning process. The proposed FPGA-based neuroevolutionary framework has been validated by building a system capable of learning autonomously to play the Pong Atari game in less than 25 generations.

The increasing integration of robots into daily life requires enhanced autonomy capabilities and better levels of robustness. To address robot localization, this work proposes Hybrid AMCL, an extension of the original Adaptive Monte Carlo Localization (AMCL) algorithm, incorporating visual cues in the form of natural rectangles, usually present indoors, to overcome difficulties posed by scenarios lacking geometric features such as corridors. To this end, this work also introduces Rectangle Intersection-based Detector using Graphs and Elongation (RIDGE), a novel algorithm based on the idea of finding quadrilaterals in images, corresponding to real-world rectangles. The effectiveness of Hybrid AMCL is validated both in simulation using Isaac Sim and in a real-world environment. Our results show its potential for more accurate localization in challenging situations.

In this paper, we present the development of an advanced social robot designed to assist older adults and children with medical conditions, focusing primarily on its potential for social interaction in these contexts. While the robot incorporates some emotional recognition and expression capabilities, the core of our work centers on the design and construction of the robot, aiming to provide a flexible platform for social engagement. The robot architecture integrates both hardware and software components optimized for interaction in care settings. Key features include various sensors, such as cameras and microphones, that enable the robot to perceive its environment and respond in ways that enhance user comfort. The robot’s design allows for future expansions in emotional recognition and empathetic interaction, although these aspects are currently in a conceptual stage. This paper details the robot’s physical construction and its overall design, emphasizing its potential to support users in social environments through interaction and communication. We discuss the various components and functionalities that make the robot a promising tool for assisting elderly individuals and children with medical conditions. Future work will involve further development of its emotional capabilities and real-world testing to evaluate its overall effectiveness.