Alex Goldhoorn

2021

How to Simulate a Global Delivery Platform
Medium - Glovo Engineering
View article

A deep dive into building a large-scale discrete event simulation system for Glovo's global delivery network. Covers architecture decisions, performance optimization strategies, and real-world validation methods for testing matching algorithms and tuning system parameters at scale.

Searching and Tracking of Humans in Urban Environments by Humanoid Robots
PhD Thesis, Universitat Politècnica de Catalunya, 2017
PDF | BibTeX | Thesis page

Algorithms for robots to find and track people in urban environments using POMCP (reinforcement learning) and particle filters. Handles noisy sensors, dynamic obstacles, and cooperative multi-robot search.

Searching and Tracking of Humans with Cooperative Mobile Robots
Autonomous Robots, vol. 42, pp. 1-25, 2017
A. Goldhoorn, A. Garrell, R. Alquézar, A. Sanfeliu
PDF | BibTeX | DOI | Project page

Presents two different techniques for coordinated multi-robot teams for searching and tracking people. A probability map of the target person's location is maintained using two methods: one based on a reinforcement learning algorithm (POMCP) and the other based on a particle filter. The validation was accomplished with extensive simulations using up to five agents and over three hours of real-life experiments with two robots.

Searching and Tracking People in Urban Environments with Static and Dynamic Obstacles
Robotics and Autonomous Systems, 2017
A. Goldhoorn, A. Garrell, R. Alquézar, A. Sanfeliu
PDF | BibTeX | DOI | Project page

Addresses the challenge of tracking people in crowded urban environments with static and dynamic obstacles. Proposes a Highest Belief Particle Filter approach that can search and track under uncertainty in continuous space and real-time, using dynamic obstacles to improve predictions of person location.

Combining Invariant Features and the ALV Homing Method for Autonomous Robot Navigation Based on Panoramas
Journal of Intelligent and Robotic Systems, vol. 64, no. 3-4, pp. 625-649, 2011
A. Ramisa, A. Goldhoorn, D. Aldavert, R. Toledo, R. López de Màntaras
PDF | BibTeX | DOI

Presents a visual homing method that combines invariant feature descriptors with the Average Landmark Vector (ALV) approach for robust autonomous robot navigation using panoramic images. Demonstrates improved performance in the presence of noise, occlusions, and dynamic environments.

Continuous Real Time POMCP to Find-and-Follow People by a Humanoid Service Robot
IEEE-RAS International Conference on Humanoid Robots (Humanoids), Madrid, Spain, 2014
A. Goldhoorn, A. Garrell, R. Alquézar, A. Sanfeliu
PDF | BibTeX | Project page

Adapts the POMCP algorithm to work in continuous action spaces for real-time operation on humanoid service robots. Enables robots to autonomously find and follow people in indoor environments, handling partial observability and dynamic human behavior with efficient online planning.

XIM-Engine: A Software Framework to Support the Development of Interactive Applications that Uses Conscious and Unconscious Reactions in Immersive Mixed Reality
Virtual Reality International Conference (VRIC), Laval, France, 2014
P. Omedas et al. (including A. Goldhoorn)
PDF | BibTeX

Presents a software framework for developing interactive applications combining conscious and unconscious reactions in immersive mixed reality environments, integrating affective computing and bio-inspired cognitive systems.

Analysis of Methods for Playing Human Robot Hide-and-Seek in a Simple Real World Urban Environment
ROBOT 2013: First Iberian Robotics Conference, Madrid, Spain, 2013
A. Goldhoorn, A. Sanfeliu, R. Alquézar
PDF | BibTeX | DOI

Analyzes different methods for enabling robots to play hide-and-seek with humans in urban environments. Compares heuristic approaches with planning algorithms under uncertainty, evaluating their performance in real-world scenarios with dynamic obstacles and occlusions.

Comparison of MOMDP and Heuristic Methods to Play Hide-and-Seek
International Conference of the Catalan Association for Artificial Intelligence (CCIA), 2013
A. Goldhoorn, A. Sanfeliu, R. Alquézar
PDF | BibTeX | DOI

Compares Mixed Observability Markov Decision Process (MOMDP) approaches with heuristic methods for the hide-and-seek problem. Shows that MOMDP-based planners can outperform heuristics by explicitly reasoning about uncertainty and information gathering.

Solving Ambiguity in Global Localization of Autonomous Robots
M.Sc. Thesis, University of Groningen, The Netherlands
A. Goldhoorn
PDF | BibTeX

Master's thesis addressing the challenge of resolving ambiguity in global robot localization, particularly in symmetric or repetitive environments where multiple hypotheses about the robot's position are equally likely.

Using the Average Landmark Vector Method for Robot Homing
Frontiers in Artificial Intelligence and Applications, vol. 163, pp. 331-338, 2007
A. Goldhoorn, A. Ramisa, R. López de Màntaras, R. Toledo
PDF | BibTeX | Interactive Demo

Evaluates the Average Landmark Vector (ALV) method for visual robot homing using panoramic feature projections. Shows robustness to noise, occlusion, and fake features, achieving 85% success rate even with 50% of features randomly removed.

Implementation of a Simultaneous Localization and Mapping System using Growing Neural Gas
B.Sc. Thesis, University of Groningen, The Netherlands
A. Goldhoorn, H. Stadman, H. Eldering
PDF | BibTeX | Interactive Demo

Bachelor's thesis presenting a SLAM (Simultaneous Localization and Mapping) implementation using Growing Neural Gas, a self-organizing neural network approach for incremental map building and robot localization.