Florin Moldoveanu

Robust camera pose and scene structure analysis for service robotics

Successful path planning and object manipulation in service robotics applications rely both on a good estimation of the robots position and orientation (pose) in the environment, as well as on a reliable understanding of the visualized scene. In this paper a robust real-time camera pose and a scene structure estimation system is proposed. First, the pose of the camera is estimated through the analysis of the so-called tracks. The tracks include key features from the imaged scene and geometric constraints which are used to solve the pose estimation problem. Second, based on the calculated pose of the camera, i.e. robot, the scene is analyzed via a robust depth segmentation and object classification approach. In order to reliably segment the objects depth, a feedback control technique at an image processing level has been used with the purpose of improving the robustness of the robotic vision system with respect to external influences, such as cluttered scenes and variable illumination conditions. The control strategy detailed in this paper is based on the traditional open-loop mathematical model of the depth estimation process. In order to control a robotic system, the obtained visual information is classified into objects of interest and obstacles. The proposed scene analysis architecture is evaluated through experimental results within a robotic collision avoidance system.

Multiple-superquadrics based object surface estimation for grasping in service robotics

A method for approximating an objects surface defined by 2.5D data is presented. The goal is to determine a compact volume which may be subjected to manipulation actions under a proper grasp planning. The object volume is obtained based on a union of parameterized geometric shapes, also known as superquadrics. In order to reduce over-segmentations and also the number of fitting combinations, the object space is defined as a 3D bounding Region of Interest (ROI). Further, via voxel decomposition, the ROI is decomposed in more meaningful smaller ROIs that can capture a large number of features from the object of interest. By using a fixed size voxel grid, the process can be slowed down for the cases of large objects. A boosting speedup of the process is proposed through dynamically adjusting the sides of the voxels for each object. Finally, a method for voxels merging is proposed. Each of the newly created ROIs, related to a particular region of the object, will be hosting a superquadric model which optimally estimates the considered surface.

GPU accelerated simulation of elliptic partial differential equations

This paper assesses the performance improvements of a GPU based implementation of an elliptic equation (steady heat conduction) over the CPU based version. An iterative method based on a finite difference approach has been used for the numerical solution (red black point successive over relaxation). The main idea is to move the computationally intensive parts of the algorithm onto the GPU. Because of the lack of communication between the blocks of the GPU grid, the computations have been included in two separate kernels. An important improvement to the GPU implementation has been the padding of the memory buffers which has led to fewer global memory transactions and faster kernel execution times. The performances of the two versions of the algorithm (CPU and CPU-GPU) have been compared on three different grained grids. The results indicate a speed-up which varies from around two for the coarsest grid up to over one order of magnitude for the finest grid.

Autonomic monitoring approach based on CEP and ML for logistic of sensitive goods

The main objective of this work is to develop a framework for supporting the development of applications for logistic companies that transport perishable goods (food and medicines). Reducing the amount of lost and damaged perishable goods during transportation and storage represents a substantial global challenge, which imply the implementation of cold chain monitoring at all levels of the supply chains. The framework contains several components that enable: (1) the real-time monitoring of goods during transportation; (2) forecast the temperatures of parcels; (3) generate real-time alerts/early warning when the product are not stored according to the acceptance criteria.

A Time-Delay Control Approach for a Stereo Vision Based Human-Machine Interaction System

In this paper, an approach to control a 6-DoF stereo camera for the purpose of actively tracking the face of a human observer in the context of Human-Robot Interaction (HRI) is proposed. The main objective in the presented work is to cope with the critical time-delay introduced by the computer vision algorithms used to acquire the feedback variable within the control system. In the studied HRI architecture, the feedback variable is represented by the 3D position of a human subject. We proposed a predictive control method which is able to handle the high time-delay inserted by the vision elements into the control system of the stereo camera. Also, along with the predictive control approach, a novel 3D nose detection algorithm is suggested for the computation of the feedback variable. The performance of the implemented platform is given through experimental results.

GPRS Based System for Atmospheric Pollution Monitoring and Warning

This paper presents the synthesis of a SCADA (supervisory control and data acquisition) system, named Pollution Guard, designed to collect and process atmospheric pollution data measured in several strategic points of a region. Pollution Guard makes use of the GPRS (general packet radio service) data communication infrastructure from a mobile communication provider that covers a very large area, practically the air pollution data being collected from every place in the country. In comparison to other similar systems, the new functionalities provided by Pollution Guard are the SMS (short messaging system) and e-mail alerts generated when the level of toxic substances exceeds some given values, chosen with regard to respiratory illness

GPU accelerated information retrieval using Bloom filters

Information retrieval is a technique used in search engines, advertisement placement and cognitive databases. With increasing amounts of data and stringent response time requirements, improving the underlying implementation of document retrieval becomes critical. To this end, we consider a Bloom filter, a simple randomized data structure that answers membership queries with no false negative and customizable false positive probability. Mainly, we focus on the speed-up of the algorithm by using a Graphics Processing Units (GPU) based implementation. Starting from a regular CPU implementation of the Bloom filter algorithm, we employ different optimization techniques on the two basic Bloom filter operations: mapping and querying. An important speed-up is achieved for both operations: over 300x for mapping, and over 20x for querying. Furthermore, we show that the number of hash functions used during the mapping operation, the number of files, and the number of query words have a significant effect on the execution time and the speed-up.

GPU accelerated simulation of the human arterial circulation

A GPU accelerated implementation of a reduced-order model of the human arterial circulation is introduced. The computationally intensive tasks of the algorithm (namely, the computation of the flow rate and area values at the interior grid points of the domain) have been migrated to the GPU. The CPU not only coordinates the actions performed by the GPU, but it also computes the inflow, bifurcation and outflow points. The GPU operations have been included in a single kernel, which has been optimized towards high computational intensity and low global memory throughput in order to obtain good results even for small arterial trees. Although the amount of data to be transferred between CPU and GPU is small, its scattered nature has lead to an approach where the entire arrays are copied back and forth. The speed-up obtained for the code migrated from the host to the device is of around 30×, while the speed-up of the entire application is of 4.5-4.7×. An important property of the proposed approach is that the speed-up obtained remains constant for arterial trees which are up to ten times smaller than e regular tree representing the large arteries of the human cardiovascular system. Hence it can be used with the same impact for simulations of entire systemic trees and of smaller localized trees.

Real-time collision avoidance for redundant manipulators

The paper presents a joints trajectories planning strategy for a redundant manipulator. The control system modifies, in real-time, the robot manipulator joints references positions, during a transition, when an unexpected moving obstacle is getting closer to one of the robot components (links and joints). The position of the moving objects is determined using a digital image processing system. The control system determines if the moving obstacle and the arm robot are in collision using an interference detection method based on oriented bounding boxes. When the arm robot control system detects a collision, it modifies the arm robot joints trajectories in order to increase the distance between the robot and the obstacle and, in the same time, maintain the initial trajectory of the end-effecter. Computer simulation and experimental results are given for a 7 DOF redundant arm and demonstrate the collision avoidance capability.

The time delay control of CAN messages for real-time communication

The latest advances in the field of microelectronics and modern control systems demand for more flexible control architectures. As a result of that, the centralized control architectures are replaced by flexible distributed control systems based on intelligent drives and powerful communication networks. CAN network was designed for car networking in order to reduce the complexity of the related wiring harnesses. To achieve the real-time communication constraints the traditional CAN application technique must be updated. This paper presents a new approach of using the dynamically priority allocation technique for a CAN network. Our solution is to increase gradually the message priority until it becomes the message with the highest priority from the network and it is transmitted. This method of message priority allocation allows us to determine its maximum transmission time.