Gas distribution modelling constitutes an ideal application area for mobile robots, which - as intelligent gas sensors offer several advantages compared to stationary sensor networks. In this paper we propose the Kernel DM+V algorithm learn a statistical 2-d model from sequence of localized measurements. The does not make strong assumptions about sensing locations and can thus be applied on robot that is primarily used monitoring, also in case treats density estimation problem. contrast most...

Roboticists often take inspiration from animals for designing sensors, actuators, or algorithms that control the behavior of robots. Bio-inspiration is motivated with uncanny ability to solve complex tasks like recognizing and manipulating objects, walking on uneven terrains, navigating source an odor plume. In particular task tracking plume up its has nearly exclusively been addressed using biologically inspired robots have developed, example, mimic moths, dung beetles, lobsters. this paper...

Chemical detection systems based on chemo-resistive sensors usually include a gas chamber to control the sample air flow and minimize turbulence. However, such kind of experimental setup does not reproduce concentration fluctuations observed in natural environments destroys spatio-temporal information contained plumes. Aiming at reproducing more realistic environments, we utilize wind tunnel with two independent sources that get naturally mixed along turbulent flow. For first time, are...

In this paper we present initial experiments towards environmental monitoring with a mobile platform. A prototype of pollution robot was set up which measures the gas distribution using an ldquoelectronic noserdquo and provides three dimensional wind measurements ultrasonic anemometer. We describe design experimental setup used to run trials under varying conditions. then results mapping. The were carried out in uncontrolled environments very different properties: enclosed indoor area, part...

This paper presents a novel linear time-varying model predictive controller (LTV-MPC) using sparse clothoid-based path description: LTV-MPCC. Clothoids are used world-wide in road design since they allow smooth driving associated with low jerk values. The formulation of the MPC is based on fact that vehicle traveling at speeds defines segment clothoids if steering angle chosen to vary piecewise linearly. Therefore, we can compute motion as clothoid parameters and translate them inputs. We...

In this paper, we present an algorithm for lateral control of a vehicle—a smooth and accurate model predictive controller (MPC). The fundamental difference compared to standard MPC is that the driving smoothness directly addressed in cost function. objective based on minimization first- second-order spatial derivatives curvature. By doing so, jerky commands steering wheel, which could lead permanent damage components vehicle structure, are avoided. A good path tracking accuracy ensured by...

Due to its environmental, economical and safety implications, methane leak detection is a crucial task address in the biogas production industry. In this paper, we introduce Gasbot, robotic platform that aims automatize emission monitoring landfills sites. The distinctive characteristic of Gasbot use Tunable Laser Absorption Spectroscopy (TDLAS) sensor. This sensor provides integral concentration measurements over path laser beam. Existing gas distribution mapping algorithms can only handle...

In this paper, we address the task of gas distribution modeling in scenarios where multiple heterogeneous compounds are present. Gas is particularly useful emission monitoring applications spatial representations gaseous patches can be used to identify hot spots. realistic environments, presence chemicals expected and therefore, discrimination has incorporated process. The approach presented work addresses by combining different non selective sensors. addressed with an open sampling system,...

In this paper, we introduce a method for identification of bacteria in human blood culture samples using an electronic nose. The uses features, which capture the static (steady state) and dynamic (transient) properties signal from gas sensor array proposes means to ensemble results consecutive samples. underlying mechanism ensembling is based on estimation posterior probability, extracted support vector machine classifier. A large dataset representing ten different cultures has been used...

We address the problem of detecting changes in activity a distant gas source from response an array metal oxide (MOX) sensors deployed open sampling system. The main challenge is turbulent nature dispersion and dynamics sensors. propose change point detection approach evaluate it on individual experimental setup where intensity, compound, or mixture ratio. also introduce efficient sensor selection algorithm with selected subsets.

In this paper we present an inspection robot to produce gas distribution maps and localize sources in large outdoor environments. The is equipped with a 3D laser range finder remote sensor that returns integral concentration measurements. We apply principles of tomography create spatial model from algorithm framed as convex optimization problem it models the mean fluctuations gases. This important since dispersion not static phenomenon furthermore, areas high fluctuation can be correlated...

Applications related to industrial plant surveillance and environmental monitoring often require the creation of gas distribution maps (GDM). In this paper an approach for creating a map using Tunable Diode Laser Absorption Spectroscopy (TDLAS) sensor laser range scanner mounted on pan tilt unit is presented. The TDLAS can remotely sense target gas, in case methane, requiring novel GDM algorithms compared ones developed traditional in-situ chemical sensors. presented setup makes it possible...

This paper presents a method for optimal speed profile generation in specified clothoid-based paths with known semantic -- maximum and longitudinal lateral acceleration geometric information. A clothoid can be described using only its kink-points information, i.e. the points defining start end of clothoid. Using path representation, we formulate as convex optimization problem where objective is to produce smooth that close allowed speed. The vehicle road define constraints problem....

Gas distribution mapping is a crucial task in emission monitoring and search rescue applications. A common assumption made by state-of-the art algorithms that only one type of gaseous substance present the environment. For real world applications, this can become very restrictive. In paper we an algorithm creates gas concentration maps scenario where multiple heterogeneous sources are present. First, using array metal oxide (MOX) sensors pattern recognition algorithm, chemical compound...

The problem of gas detection is relevant to many real-world applications, such as leak in industrial settings and landfill monitoring. In this paper, we address the large areas with a mobile robotic platform equipped remote sensor. We propose an algorithm that leverages novel method based on convex relaxation for quickly solving sensor placement problems, generating efficient exploration plan robot. To demonstrate applicability our environments, performed number experimental trials, both...

This work presents a gas discrimination approach for Open Sampling Systems (OSS), composed of non-specific metal oxide sensors only. In an OSS, as used on robots or in sensor networks, the are exposed to dynamics environment and thus, most data corresponds highly diluted samples while high concentrations sparse. addition, positive correlation between class separability concentration level can be observed. The proposed computes posteriors by coupling pairwise probabilities compounds...

In this paper, we present our initial experiments to collect a database for studying mobile robot olfaction. Mobile robots with olfactory sensing capabilities are expected be used in various applications including gas distribution mapping and source localization. Owing the turbulent nature of airflow field distribution, these must equipped algorithms that can cope chaotic environments. Since it is important check applicability such diversity environments, propose build which users test...

Calibration of metal oxide (MOX) gas sensors for continuous monitoring is a complex problem due to the highly dynamic characteristics sensor signal when exposed natural environment in an Open Sampling System (OSS). This work presents probabilistic approach calibration MOX using Gaussian Processes (GP). The proposed estimates every measurement probability distribution corresponding concentration, which enables calculation confidence intervals predicted concentrations. Being able predict...

The problem of gas detection is relevant to many real-world applications, such as leak in industrial settings and surveillance. In this paper we address the large areas with a mobile robotic platform equipped remote sensor. We propose novel method based on convex relaxation for quickly finding an exploration plan that guarantees complete coverage environment. Our proves be highly efficient terms computational requirements provide nearly-optimal solutions. validate our approach both...

In this paper, we introduce a Bayesian time series model approach for gas concentration estimation using Metal Oxide (MOX) sensors in Open Sampling System (OSS). Our focuses on the compensation of slow response MOX sensors, while concurrently solving problem estimating OSS. The proposed Augmented Switching Linear allows to include all sources uncertainty arising at each step single coherent probabilistic formulation. particular, detecting on-line current sensor dynamical regime and...