José Gabriel Ramírez-Torres

A Statistical Study of the Effects of Neighborhood Topologies in Particle Swarm Optimization

The behavior of modern meta-heuristics is directed by both, the variation operators, and the values selected for the parameters of the approach. Particle swarm optimization (PSO) is a meta-heuristic which has been found to be very successful in a wide variety of optimization tasks. In PSO, a swarm of particles fly through hyper-dimensional search space being attracted by both, their personal best position and the best position found so far within a neighborhood.

The Exact Euclidean Distance Transform: A New Algorithm for Universal Path Planning

The Path-Planning problem is a basic issue in mobile robotics, in order to allow the robots to solve more complex tasks, for example, an exploration assignment in which the distance given by the planner is taken as a utility measure. Among the different proposed approaches, algorithms based on an exact cell decomposition of the environment are very popular. In this paper, we present a new algorithm for universal path planning in cell decomposition, using a raster scan method for computing the Exact Euclidean Distance Transform (EEDT) for each cell in the map. Our algorithm computes, for every cell in the map, the point sequence to the goal. For each sequence, the sub-goals are selected near to the vertices of the obstacles, reducing the total distance to the goal without post processing. At the end, we obtain a smooth path up to the goal without the need for post-processing. The paths are computed by visibility verification among the cells, exploiting the processing performed in the neighbouring cells.

A New Object Path Planner for the Box Pushing Problem

The box-pushing problem, where a community of mobile robots must move an object in cluttered environments, represents a very interesting challenge for the development of coordination schemes for multiagent systems, with multiple practical applications (e.g. industrial processes automatization, people rescue, movement of dangerous objects). In this paper we introduce a new path-planner so the resulting path of the object facilitates its manipulation by the robots to accomplish the task.

A Complete Closed-Form Solution to the Inverse Kinematics Problem for the P2Arm Manipulator Robot

In this paper, we propose a complete closed-form solution to the inverse kinematics problem for a 5-dof manipulator robot, where the existence conditions for all the possible solutions are clearly established and the singular configurations where infinite solutions exist are identified. We define a desired configuration for the final effector as the position for the center of the gripper and the direction of the grippers main axis. This choice seems natural for robots mainly devoted to assembling tasks, even if the rotation around this main axis cannot be freely set. Under these conditions, a straight-forward method in order to obtain all the different exact solutions is also presented.

Towards an Autonomous Airborne Robotic Agent

Commercialy available (UAVs) rely on the Global Positioning System (GPS) to define their flight plan, while assuming an obstacle-free environment. The work presented on this article aims to set the foundation towards an autonomous airborne agent, capable of locating itself by means of computer vision, modeling its environment, planning and executing a three dimensional trajectory. On the first stage of development we solved the localization problem using artificial markers and tested a PID controller to make the vehicle follows a given trajectory (a lemniscate); as results, we show flight data captured during real flights. This development would facilitate the integration of far more complex flight behaviors than GPS only guided flight plans.

An Airborne Agent

In order to change the control approach that commercially available Unmanned Aerial Vehicles (UAVs) use to execute a flight plan, which is based on the Global Positioning System (GPS) and assuming an obstacle-free environment, we propose a hierarchical multi-layered control system that permits to a UAV to define the flight plan during flight and locate itself by other means than GPS. The work presented on this article aims to set the foundation towards an autonomous airborne agent, capable of locating itself with the aid of computer vision, model its environment and plan and execute a three dimensional trajectory. On the current stage of development we locate the vehicle using a board of artificial markers, the flight plan to execute was defined as either a cubic spline or a Lemniscate. As results, we present the resultant flight data when the proposed control architecture drives the vehicle autonomously.

Real-Time Reconstruction of Heightmaps from Images Taken with an UAV

In this paper, an approach for a real-time three dimensional reconstruction of a dense point cloud, from images obtained with an unmanned aerial vehicle, is presented. The approach also estimates the model scale, using a known marker as reference, obtaining a fairly precise heightmap of the environment. The proposed image analysis framework is also capable of an off-line 3D reconstruction from monocular images, without known markers in the environment. The proposed framework is an adaptation of the Patch-based Multiview Stereo (PMVS) algorithm, in order to obtain a faster point cloud generation, suitable for real-time environment reconstruction. For the obtained heightmap, the error estimation ranges between 1–1.5 % for manual detection of marker and 4–5 % for automatic marker detection, which we consider is accurate enough for autonomous navigation and path planning.

Autonomous navigation of unmanned aerial vehicles guided by visual features of the terrain

In this work we present an autonomous control scheme for unmanned aerial vehicles (UAV), based on pure visual information obtained from the UAV on-board camera. The objective of the system is to visually follow some characteristics of terrain, specifically the existing boundary between two distinct areas, using an image processing pipeline. The motivation of this work is to provide a base system to control a UAV, over a border identified from any segmentation areas defined by the particular application of the UAV. With this reason in mind, an additional contribution is to provide a modular publisher/subscriber control scheme which can be easily reused for any particular application. The present work has been tested in an indoor environment, and its performance evaluated versus human users, measuring the orientation and distance from target.

On-Line Dense Point Cloud Generation from Monocular Images with Scale Estimation

This paper introduces an approach for on-line marker-based three dimensional modeling with scale estimation and heightmap construction from monocular images. The presented system is also capable of an off-line marker-less 3D reconstruction from monocular images with increased detail. This method is designed for the flexible use with an Unmaned Aerial Vehicle (UAV); this means that, despite being tested with a Parrot AR.Drone 1.0, it is easily portable to other more capable UAV models. The followed approach was an adaptation of the patch-based Multiview Stereo (PMVS) algorithm for on line point cloud generation. The system achieved 1.05 processed images per second on average, slightly surpassing the planed objective of 1 processed image per second. The height estimation error ranges between 1-1.5% with a manual marker detection and 4-5% with automatic marker detection, which seems accurate enough for autonomous navigation and path planning. As future work, tests with a better UAV, processing time reduction, marker-less height map construction, autonomous indoor navigation and collaborative on-line 3D modeling are planned.

Grasp quality for the object transportation by communities of mobile robots

The box-pushing problem, where a community of mobile robots must move an object in cluttered environments, represents a very interesting challenge for the development of coordination schemes for multiagent systems, with multiple practical applications. In this paper we present a brief review of the cooperative multi robot box pushing problem and the related trends, and we describe the strategy that we propose to solve it. Specifically in this work we will addresses the problem of grasping a polygonal object that will be manipulated by mobile robots that only can push the object. The strategy was inspired in the way that robotic hands grasp objects.