Andry Maykol Pinto

A Flow-based Motion Perception Technique for an Autonomous Robot System

Visual motion perception from a moving observer is the most often encountered case in real life situations. It is a complex and challenging problem, although, it can promote the arising of new applications. This article presents an innovative and autonomous robotic system designed for active surveillance and a dense optical flow technique. Several optical flow techniques have been proposed for motion perception however, most of them are too computationally demanding for autonomous mobile systems. The proposed HybridTree method is able to identify the intrinsic nature of the motion by performing two consecutive operations: expectation and sensing. Descriptive properties of the image are retrieved using a tree-based scheme and during the expectation phase. In the sensing operation, the properties of image regions are used by a hybrid and hierarchical optical flow structure to estimate the flow field. The experiments prove that the proposed method extracts reliable visual motion information in a short period of time and is more suitable for applications that do not have specialized computer devices. Therefore, the HybridTree differs from other techniques since it introduces a new perspective for the motion perception computation: high level information about the image sequence is integrated into the estimation of the optical flow. In addition, it meets most of the robotic or surveillance demands and the resulting flow field is less computationally demanding comparatively to other state-of-the-art methods.

Enhancing dynamic videos for surveillance and robotic applications: The robust bilateral and temporal filter

Over the last few decades, surveillance applications have been an extremely useful tool to prevent dangerous situations and to identify abnormal activities. Although, the majority of surveillance videos are often subjected to different noises that corrupt structured patterns and fine edges. This makes the image processing methods even more difficult, for instance, object detection, motion segmentation, tracking, identification and recognition of humans. This paper proposes a novel filtering technique named robust bilateral and temporal (RBLT), which resorts to a spatial and temporal evolution of sequences to conduct the filtering process while preserving relevant image information. A pixel value is estimated using a robust combination of spatial characteristics of the pixels neighborhood and its own temporal evolution. Thus, robust statics concepts and temporal correlation between consecutive images are incorporated together which results in a reliable and configurable filter formulation that makes it possible to reconstruct highly dynamic and degraded image sequences. The filtering is evaluated using qualitative judgments and several assessment metrics, for different Gaussian and Salt-Pepper noise conditions. Extensive experiments considering videos obtained by stationary and non-stationary cameras prove that the proposed technique achieves a good perceptual quality of filtering sequences corrupted with a strong noise component.

The SPIDERobot: A Cable-Robot System for On-site Construction in Architecture

The use of robots in architectural construction has been a research field since the 1980s. Driven by both productive and creative concerns, different systems have been devised based on large-scale robotic structures, mobile robotic units or flying robotic vehicles. By analyzing these approaches and discussing their advantages and limitations, this paper presents an alternative strategy to automate the building construction processes in on-site scenarios. The SPIDERobot is a cable-robot system developed to perform assembly operations, which is driven by a specific Feedback Dynamic Control System (FDCS) based on a vision system. By describing and illustrating this research work, the authors argue about the advantages of this cable robot system to deal with the complexity and the scale of building construction in architecture.

Evaluation of Depth Sensors for Robotic Applications

The sensors that acquire 3D data play an important role in many applications. In addition, they have been used in the robotic field for several purposes, for instance, enhancing the navigation of mobile robots, object detection, scene reconstruction, 3D inspection of parts and others. Moreover, a significant amount of devices with distinct cost, accuracy and features have been released in the recent years which increases the difficulty of comparing each sensor in a proper manner or choosing the most suitable device for a specific task and operation field. This paper compares the Kinect v1, Kinect v2, Structure Sensor and Mesa Imaging SR4000. The noise of each sensor is characterized for different distances and considering objects with different colors. Therefore, this paper proposes a simple but quantitative benchmark for evaluating 3D devices that characterizes the most relevant features for the robotic field and in accordance with different type of operations.

An architecture for visual motion perception of a surveillance-based autonomous robot

This research presents an innovative mobile robotic system designed for active surveillance operations. This mobile robot moves along a rail and is equipped with a monocular camera. Thus, it enhances the surveillance capability when compared to conventional systems (mainly composed by multiple static cameras). In addition, the paper proposes a technique for multi-object tracking called MTMP (Multi-Tracking of Motion Profiles). The MTMP resorts to a formulation based on the Kalman filter and tracks several moving objects using motion profiles. A motion profile is characterized by the dominant flow vector and is computed using the optical flow signature with removal of outliers. A similarity measure based on the Mahalanobis distance is used by the MTMP for associating the moving objects over frames. The experiments conducted in realistic environments have proved that the static perception mode of the proposed robot is able to detect and track multiple moving objects in a short period of time and without using specialized computers. In addition, the MTMP exhibits a good computational performance since it takes less than 5 milliseconds to compute. Therefore, results show that the estimation of motion profiles is suitable for analyzing motion on image sequences.

Unsupervised flow-based motion analysis for an autonomous moving system ☆

Abstract This article discusses the motion analysis based on dense optical flow fields and for a new generation of robotic moving systems with real-time constraints. It focuses on a surveillance scenario where an especially designed autonomous mobile robot uses a monocular camera for perceiving motion in the environment. The computational resources and the processing-time are two of the most critical aspects in robotics and therefore, two non-parametric techniques are proposed, namely, the Hybrid Hierarchical Optical Flow Segmentation and the Hybrid Density-Based Optical Flow Segmentation. Both methods are able to extract the moving objects by performing two consecutive operations: refining and collecting. During the refining phase, the flow field is decomposed in a set of clusters and based on descriptive motion properties. These properties are used in the collecting stage by a hierarchical or density-based scheme to merge the set of clusters that represent different motion models. In addition, a model selection method is introduced. This novel method analyzes the flow field and estimates the number of distinct moving objects using a Bayesian formulation. The research evaluates the performance achieved by the methods in a realistic surveillance situation. The experiments conducted proved that the proposed methods extract reliable motion information in real-time and without using specialized computers. Moreover, the resulting segmentation is less computationally demanding compared to other recent methods and therefore, they are suitable for most of the robotic or surveillance applications.

Shop floor scheduling in a mobile robotic environment

Nowadays,it is far more common to see mobile robotics working in the industrial sphere due to the mandatory need to achieve a new level of productivity and increase profits by reducing production costs. Management scheduling and task scheduling are crucial for companies that incessantly seek to improve their processes, increase their efficiency, reduce their production time and capitalize on their infrastructure by increasing and improving production. However, when faced with the constant decrease in production cycles, management algorithms can no longer solely focus on the mere management of the resources available, they must attempt to optimize every interaction between them, to achieve maximum efficiency for each production resource. In this paper we focus on the presentation of the new competition called Robot@Factory, its environment and its main objectives, paying special attention to the scheduling algorithm developed for this specific case study. The findings from the simulation approach have allowed us to conclude that mobile robotic path planning and the scheduling of the associated tasks represent a complex problem that has a strong impact on the efficiency of the entire production process.

An Optimization Approach for the Inverse Kinematics of a Highly Redundant Robot

This paper describes a robot with 12 degrees of freedom for pick-and-place operations using bricks. In addition, an optimization approach is proposed, which determines the state of each joint (that establishes the pose for the robot) based on the target position while minimizing the effort of the servomotors avoiding the inverse kinematics problem, which is a hard task for a 12 DOF robot manipulator. Therefore, it is a multi-objective optimization problem that will be solved using two optimization methods: the Stretched Simulated Annealing method and the NSGA II method. The experiments conducted in a simulation environment prove that the proposed approach is able to determine a solution for the inverse kinematics problem. A real robot formed by several servomotors and a gripper is also presented in this research for validating the solutions.

Streaming Image Sequences for Vision-Based Mobile Robots

Vision-based mobile robots have severe limitations related to the computational capabilities that are required for processing their algorithms. The vision algorithms processed onboard and without resorting to specialized computing devices do not achieve the real-time constraints that are imposed by that kind of systems.

Cable robot for non-standard architecture and construction: A dynamic positioning system

In the past few years, cable-driven robots have received some attention by the scientific community and the industry. They have special characteristics that made them very reliable to operate with the level of safeness that is required by different environments, such as, handling of hazardous materials in construction sites. This paper presents a cable-driven robot called SPIDERobot, that was developed for automated construction of architectural projects. This robot has a rotating claw and it is controlled by a set of 4 cables that allow 4 degrees of freedom. In addition to the robot, this paper introduces a Dynamic Control System (DCS) that controls the positioning of the robot and assures that the length of cables is always within a safe value. Results show that traditional force-feasible approaches are more influenced by the pulling forces or the geometric arrangement of all cables and their positioning is significantly less accurate than the DCS. Therefore, the architecture of the SPIDERobot is designed to enable an easily scaling up of the solution to higher dimensions for operating in realistic environments.