Pedro Angelo Morais de Sousa

An olfactory-based robot swarm navigation method

This paper presents a novel robot swarming navigation algorithm in order to find the odor sources in an unknown environment, based on the ability of each swarm member to sense the odor. Each robot in the swarm has a cooperative localization system which uses wireless network as a mean of measuring the distance from the other robots. In this method, at least three robots act as stationary measurement beacons while the other robots of the swarm navigate in the environment towards the odor source. In the next step, the roles of the robots will be switched and some other robots will act as beacons. The experimental tests report a good result in finding the odor source and also the accuracy of localization system1.

Real-Time Labeling of Places using Support Vector Machines

Humans refer almost to everything by their characterization rather than their detailed descriptions. For example, in indoor environments places are specified as: rooms, corridors, etc. Such categorizations, if learned by a robot, could improve the capabilities in the areas of navigation, localization, or human- robot cooperation. This paper studies the problem of categorizing environments into semantic categories. A new approach based on Support Vector Machine (SVM) is proposed and described for learning to perform classification of environment. The SVM is trained using a supervised training algorithm. This method uses simple features extracted from laser range measures, using methodologies normally used in computer vision. In the present paper the proposed method is used to distinguish between two classes of places from sensor data: rooms and corridors. The real-time experimental architecture designed for classification is presented. Experimental results obtained with real sensor data demonstrate the feasibility and effectiveness of the proposed approach.

Cooperative chemical concentration map building using Decentralized Asynchronous Particle Swarm Optimization based search by mobile robots

In this article the main objective is to perform a search in an unknown area with multiple robots in order to determine the region with highest chemical gas concentration as well as to build the chemical gas concentration map. The searching and map building tasks are accomplished by using mobile robots equipped with smart transducers for gas sensing. Robots perform the search autonomously by using their own data and the information (position information and sensor readings) obtained from the other robots. Moreover, simultaneously the robots send their sensor readings of the chemical concentration and their position data to a remote computer (a base station), where the data is combined, interpolated, and filtered to form an real-time map of the chemical gas concentration in the environment. To achieve this task as a high-level path planning algorithm we use a decentralized and asynchronous version of the Particle Swarm Optimization (PSO) algorithm which also allows for time-varying neighborhood.

An architecture for adaptive fuzzy control in industrial environments

The paper presents an architecture for adaptive fuzzy control of industrial systems. Both conventional and adaptive fuzzy control can be designed. The control methodology can integrate a priori knowledge about the control and/or about the plant, with on-line control adaptation mechanisms to cope with time-varying and/or uncertain plant parameters. The paper presents the fuzzy control software architecture that can be integrated in industrial processing and communication structures. It includes four distinct modules: a mathematical fuzzy library, a graphical user interface (GUI), fuzzy controller, and industrial communication. Three types of adaptive fuzzy control methods have been studied, and compared: (1) direct adaptive, (2) indirect adaptive, and (3) combined direct/indirect adaptive. An experimental benchmark composed of two mechanically coupled electrical DC motors has been employed to study the performance of the presented control architectures. The first motor acts as an actuator, while the second motor is used to generate nonlinearities and/or time-varying load. Results indicate that all tested controllers have good performance in overcoming changes of DC motor load.

Player/Stage simulation of olfactory experiments

This paper describes a Player/Stage simulation framework for mobile robot olfactory experiments. The PlumeSim framework simulates odor transport in the environment, as well as olfactory detection systems. The odor transport can be based on analytical theoretical models, data generated by dedicated computational fluid dynamics (CFD) software packages, or data acquired in a real environment. The framework is validated through comparison of the results of simulated and real mobile robot olfactory experiments.

Design and application of Soft Sensor using Ensemble Methods

Industries are faced with the choice of suitable process control policies to improve costs, quality and raw material consumption. In the paper pulp industry, it is important to estimate quickly the Chemical Oxygen Demand (COD), a parameter that is highly correlated to product quality. Soft Sensors (SSs) have been established as alternative to hardware sensors and laboratory measurements for monitoring and control purposes. However, in real setups it is often difficult to get sufficient data for SS development. This work proposes Ensemble Methods (EMs) as a way to improve the SS performance for small datasets. EMs use a set of models to obtain better prediction. Their success is usually attributed to the diversity. Bootstrap and noise injection are used to produce diverse models. Several combinations of EMs are compared. The SS is successfully applied to estimate COD in a pulp process.

Chemical concentration map building through bacterial foraging optimization based search algorithm by mobile robots

In this article we present implementation of Bacterial Foraging Optimization algorithm inspired search by multiple robots in an unknown area in order to find the region with highest chemical gas concentration as well as to build the chemical gas concentration map. The searching and map building tasks are accomplished by using mobile robots equipped with smart transducers for gas sensing called “KheNose”. Robots perform the search autonomously via bacterial chemotactic behavior. Moreover, simultaneously the robots send their sensor readings of the chemical concentration and their position data to a remote computer (a base station), where the data is combined, interpolated, and filtered to form an real-time map of the chemical gas concentration in the environment.

RFID tuning methodology applied on airport baggage tracking

This paper presents an RFID equipment tuning and configuration methodology developed in a project to support baggage tracking and feed dashboards with real time status of Service Level Agreements between the airport, the airliner and the ground operators. The project has an end-to-end travel perspective: at the departure airport, monitoring RFID equipment is used from check-in desks along pre-defined baggage circuits up to the airplane cargo entrance; at the arrival airport, RFID equipment is used from airplane cargo doors up to customs exit. The lack of adequate tooling on the market presented the opportunity to build the tool used to collect the information presented on this paper. Along with the proposed methodology to tune RFID equipment on the site survey, we also present the architecture used to implement such system.

Support Vector Machines and Features for Environment Perception in Mobile Robotics

Environment perception is one of the most challenging and underlying task which allows a mobile robot to perceive obstacles, landmarks and extract useful information to navigate safely. In this sense, classification techniques applied to sensor data may enhance the way mobile robots sense their surroundings. Amongst several techniques to classify data and to extract relevant information from the environment, Support Vector Machines (SVM) have demonstrated promising results, being used in several practical approaches. This chapter presents the core theory of SVM, and applications in two different scopes: using Lidar (Light Detection and Ranging) to label specific places, and vision-based human detection aided by Lidar.

Event-clustering for real-time data modeling

This paper proposes EventCluster, a novel approach in real-time data modeling. It deploys the Rank Order Clustering (ROC) method to automatically group all existing data sensors and actuators of the system to the Key Performance Indicators of the system. EventCluster (EC) is a cause-effect relationship data clustering tool that detects the interrelationship between field data and system performance parameters in real-time. Through its simple data filtering mechanism it can be used as a precursor to real-time sensitivity analysis. The underpinning logic of the technique is that the raw data can be obtained from field data acquisition devices and the degree of their influence on key system performance indicators can be measured in realtime with minimum computational effort. Normally monitoring and control systems are equipped with sensors and actuators that provide information for a pre-specified function regardless of other parts of the system. The global assumption of method is that a system performance or state is a function of all the inputs of the system, unless proven otherwise. In the proposed method all the inputs and outputs of the system are assumed to affect one another unless proven otherwise. In this paper, an experiment in Cement Kiln operation case demonstrates the suitability and applicability of EventClustering modeling method in industrial applications. We use the Supervisory Control and Data Acquisition (SCADA) sensors and actuators installed to monitor the operations of Kilns in Cement manufacturing process and its contagious operations as a case study for proof of concept. The sensors and actuators data collected builds the input data for measuring the performance (output) of the Kiln. The EventCluster algorithm resides within the control center of the SCADA system to assess the contribution of each input to the overall key performance indicators (output) of the process. This method improves the quality of data analysis and reduces computation overhead on the control system.