Place categorization and semantic mapping on a mobile robot

In this paper we focus on the challenging problem of place categorization and semantic mapping on a robot with-out environment-specific training. Motivated by their ongoing success in various visual recognition tasks, we build our system upon a state-of-the-art convolutional network. We overcome its closed-set limitations by complementing the network with a series of one-vs-all classifiers that can learn to recognize new semantic classes online. Prior domain knowledge is incorporated by embedding the classification system into a Bayesian filter framework that also ensures temporal coherence. We evaluate the classification accuracy of the system on a robot that maps a variety of places on our campus in real-time. We show how semantic information can boost robotic object detection performance and how the semantic map can be used to modulate the robot’s behaviour during navigation tasks. The system is made available to the community as a ROS module.

Mixed pixel analysis for flood mapping using extended support vector machine

This paper addresses the challenges of flood mapping using multispectral images. Quantitative flood mapping is critical for flood damage assessment and management. Remote sensing images obtained from various satellite or airborne sensors provide valuable data for this application, from which the information on the extent of flood can be extracted. However the great challenge involved in the data interpretation is to achieve more reliable flood extent mapping including both the fully inundated areas and the 'wet' areas where trees and houses are partly covered by water. This is a typical combined pure pixel and mixed pixel problem. In this paper, an extended Support Vector Machines method for spectral unmixing developed recently has been applied to generate an integrated map showing both pure pixels (fully inundated areas) and mixed pixels (trees and houses partly covered by water). The outputs were compared with the conventional mean based linear spectral mixture model, and better performance was demonstrated with a subset of Landsat ETM+ data recorded at the Daly River Basin, NT, Australia, on 3rd March, 2008, after a flood event.

Spectral unmixing with estimated adaptive endmember index using Extended Support Vector Machine

The most difficult operation in the flood inundation mapping using optical flood images is to separate fully inundated areas from the ‘wet’ areas where trees and houses are partly covered by water. This can be referred as a typical problem the presence of mixed pixels in the images. A number of automatic information extraction image classification algorithms have been developed over the years for flood mapping using optical remote sensing images. Most classification algorithms generally, help in selecting a pixel in a particular class label with the greatest likelihood. However, these hard classification methods often fail to generate a reliable flood inundation mapping because the presence of mixed pixels in the images. To solve the mixed pixel problem advanced image processing techniques are adopted and Linear Spectral unmixing method is one of the most popular soft classification technique used for mixed pixel analysis. The good performance of linear spectral unmixing depends on two important issues, those are, the method of selecting endmembers and the method to model the endmembers for unmixing. This paper presents an improvement in the adaptive selection of endmember subset for each pixel in spectral unmixing method for reliable flood mapping. Using a fixed set of endmembers for spectral unmixing all pixels in an entire image might cause over estimation of the endmember spectra residing in a mixed pixel and hence cause reducing the performance level of spectral unmixing. Compared to this, application of estimated adaptive subset of endmembers for each pixel can decrease the residual error in unmixing results and provide a reliable output. In this current paper, it has also been proved that this proposed method can improve the accuracy of conventional linear unmixing methods and also easy to apply. Three different linear spectral unmixing methods were applied to test the improvement in unmixing results. Experiments were conducted in three different sets of Landsat-5 TM images of three different flood events in Australia to examine the method on different flooding conditions and achieved satisfactory outcomes in flood mapping.

Mixed pixel analysis and assessment for flood mapping

The most difficult operation in flood inundation mapping using optical flood images is to map the ‘wet’ areas where trees and houses are partly covered by water. This can be referred to as a typical problem of the presence of mixed pixels in the images. A number of automatic information extracting image classification algorithms have been developed over the years for flood mapping using optical remote sensing images, with most labelling a pixel as a particular class. However, they often fail to generate reliable flood inundation mapping because of the presence of mixed pixels in the images. To solve this problem, spectral unmixing methods have been developed. In this thesis, methods for selecting endmembers and the method to model the primary classes for unmixing, the two most important issues in spectral unmixing, are investigated. We conduct comparative studies of three typical spectral unmixing algorithms, Partial Constrained Linear Spectral unmixing, Multiple Endmember Selection Mixture Analysis and spectral unmixing using the Extended Support Vector Machine method. They are analysed and assessed by error analysis in flood mapping using MODIS, Landsat and World View-2 images. The Conventional Root Mean Square Error Assessment is applied to obtain errors for estimated fractions of each primary class. Moreover, a newly developed Fuzzy Error Matrix is used to obtain a clear picture of error distributions at the pixel level. This thesis shows that the Extended Support Vector Machine method is able to provide a more reliable estimation of fractional abundances and allows the use of a complete set of training samples to model a defined pure class. Furthermore, it can be applied to analysis of both pure and mixed pixels to provide integrated hard-soft classification results. Our research also identifies and explores a serious drawback in relation to endmember selections in current spectral unmixing methods which apply fixed sets of endmember classes or pure classes for mixture analysis of every pixel in an entire image. However, as it is not accurate to assume that every pixel in an image must contain all endmember classes, these methods usually cause an over-estimation of the fractional abundances in a particular pixel. In this thesis, a subset of adaptive endmembers in every pixel is derived using the proposed methods to form an endmember index matrix. The experimental results show that using the pixel-dependent endmembers in unmixing significantly improves performance.

Systematic bias in MODIS dust aerosol retrieval at Kanpur (AERONET), Indo-Gangetic Basin - art.no.640816

We have compared the spectral aerosol optical depth (AOD) and aerosol fine mode fraction (AFMF) derived from Moderate Resolution Imaging Spectroradiometer (MODIS) with those of Aerosol Robotic Network (AERONET) at Kanpur (26.45N, 80.35E), northern India for the pre-monsoon season (March to June, 2001-2005). We found that MODIS systematically overestimates AOD during pre-monsoon season (known to be influenced by dust transport from north-west of India). The errors in AOD were correlated with the MODIS top-of-atmosphere apparent surface reflectance in 2.1 mu m channel (rho*(2.1)). MODIS aerosol algorithm uses p*(2.1) to derive the surface reflectance in visible channels (rho(0.47), rho(0.66)) using an empirical mid IR-visible correlation (rho(0.47) = rho(2.1)/4, rho(0.66) = rho(2.1)/2). The large uncertainty in estimating surface reflectance in visible channels (Delta rho(0.66)+/- 0.04, Delta rho(0.47)+/- 0.02) at higher values of p*(2.1) (p*(2.1) > 0.18) leads to higher aerosol contribution in the total reflected radiance at top-of atmosphere to compensate for the reduced surface reflectance in visible channels and thus leads to overestimation of AOD. This was also reflected in the very low values of AFMF during pre-monsoon whose accuracy depends on the aerosol path radiance in 0.47 and 0.66 mu m channels and aerosol models. The errors in AOD were also high in the scattering angle range 110 degrees-140 degrees, where the effect of dust non-spherity on its optical properties is significant. The direct measurements of spectral surface reflectance are required over the Indo-Gangetic basin in order to validate the mid IR-visible relationship. MODIS aerosol models should also be modified to incorporate the effect of non-spherity of dust aerosols.

Improved retrieval of aerosol size-resolved properties from moderate resolution imaging spectroradiometer over India: Role of aerosol model and surface reflectance

We have compared the total as well as fine mode aerosol optical depth (tau and tau(fine)) retrieved by Moderate Resolution Imaging Spectroradiometer (MODIS) onboard Terra and Aqua (2001-2005) with the equivalent parameters derived by Aerosol Robotic Network (AERONET) at Kanpur (26.45 degrees N, 80.35 degrees E), northern India. MODIS Collection 005 (C005)-derived tau(0.55) was found to be in good agreement with the AERONET measurements. The tau(fine) and eta (tau(fine)/tau) were, however, biased low significantly in most matched cases. A new set of retrieval with the use of absorbing aerosol model (SSA similar to 0.87) with increased visible surface reflectance provided improved tau and tau(fine) at Kanpur. The new derivation of eta also compares well qualitatively with an independent set of in situ measurements of accumulation mass fraction over much of the southern India. This suggests that though MODIS land algorithm has limited information to derive size properties of aerosols over land, more accurate parameterization of aerosol and surface properties within the existing C005 algorithm may improve the accuracy of size-resolved aerosol optical properties. The results presented in this paper indicate that there is a need to reconsider the surface parameterization and assumed aerosol properties in MODIS C005 algorithm over the Indian region in order to retrieve more accurate aerosol optical and size properties, which are essential to quantify the impact of human-made aerosols on climate.

Belief revision - an axiomatic approach

Belief revision systems aim at keeping a database consistent. They mostly concentrate on how to record and maintain dependencies. We propose an axiomatic system, called MFOT, as a solution to the problem of belief revision. MFOT has a set of proper axioms which selects a set of most plausible and consistent input beliefs. The proposed nonmonotonic inference rule further maintains consistency while generating the consequences of input beliefs. It also permits multiple property inheritance with exceptions. We have also examined some important properties of the proposed axiomatic system. We also propose a belief revision model that is object-centered. The relevance of such a model in maintaining the beliefs of a physician is examined.

Search Strategies for Multiple UAV Search and Destroy Missions

Multiple UAVs are deployed to carry out a search and destroy mission in a bounded region. The UAVs have limited sensor range and can carry limited resources which reduce with use. The UAVs perform a search task to detect targets. When a target is detected which requires different type and quantities of resources to completely destroy, then a team of UAVs called as a coalition is formed to attack the target. The coalition members have to modify their route to attack the target, in the process, the search task is affected, as search and destroy tasks are coupled. The performance of the mission is a function of the search and the task allocation strategies. Therefore, for a given task allocation strategy, we need to devise search strategies that are efficient. In this paper, we propose three different search strategies namely; random search strategy, lanes based search strategy and grid based search strategy and analyze their performance through Monte-Carlo simulations. The results show that the grid based search strategy performs the best but with high information overhead.

A discrete state perspective of manipulator workspaces Vorstellungen von diskreten zuständen von manipulator-arbeitsräumen

The finite resolution of joint drives or sensors imparts a discrete nature to the joints of a manipulator. Because of this an arbitrary point in the workspace cannot be reached without error even in ideal mechanical environment. This paper investigates the effect of this discrete nature of the joints on the accuracy of performance of a manipulator and develops a method to select the joint states to reach a point with least error. It is shown that the configuration leading to least error cannot, in general, be found from configuration space, especially when there is large variation in the link lengths or joint resolutions or both. The anomaly becomes severe when the gross motion of the end-effector approaches the local resolution of the workspace. The paper also shows how to distinguish two workspaces which may be identical so far as the boundary points are concerned, taking the joint resolutions into account. Finally, the concepts have been extended to define continuous space global and local performance indices for general multi degree of freedom manipulators.

Multiple UAV Coalitions for a Search and Prosecute Mission

Unmanned aerial vehicles (UAVs) have the potential to carry resources in support of search and prosecute operations. Often to completely prosecute a target, UAVs may have to simultaneously attack the target with various resources with different capacities. However, the UAVs are capable of carrying only limited resources in small quantities, hence, a group of UAVs (coalition) needs to be assigned that satisfies the target resource requirement. The assigned coalition must be such that it minimizes the target prosecution delay and the size of the coalition. The problem of forming coalitions is computationally intensive due to the combinatorial nature of the problem, but for real-time applications computationally cheap solutions are required. In this paper, we propose decentralized sub-optimal (polynomial time) and decentralized optimal coalition formation algorithms that generate coalitions for a single target with low computational complexity. We compare the performance of the proposed algorithms to that of a global optimal solution for which we need to solve a centralized combinatorial optimization problem. This problem is computationally intensive because the solution has to (a) provide a coalition for each target, (b) design a sequence in which targets need to be prosecuted, and (c) take into account reduction of UAV resources with usage. To solve this problem we use the Particle Swarm Optimization (PSO) technique. Through simulations, we study the performance of the proposed algorithms in terms of mission performance, complexity of the algorithms and the time taken to form the coalition. The simulation results show that the solution provided by the proposed algorithms is close to the global optimal solution and requires far less computational resources.

Active Vibration Suppression of One-dimensional Nonlinear Structures Using Optimal Dynamic Inversion

A flexible robot arm can be modeled as an Euler-Bernoulli beam which are infinite degrees of freedom (DOF) system. Proper control is needed to track the desired motion of a robotic arm. The infinite number of DOF of beams are reduced to finite number for controller implementation, which brings in error (due to their distributed nature). Therefore, to represent reality better distributed parameter systems (DPS) should be controlled using the systems partial differential equation (PDE) directly. In this paper, we propose to use a recently developed optimal dynamic inversion technique to design a controller to suppress nonlinear vibration of a beam. The method used in this paper determines control forces directly from the PDE model of the system. The formulation has better practical significance, because it leads to a closed form solution of the controller (hence avoids computational issues).

Classical screw theory: A fresh look using dual eigenvalue approach

This paper presents a novel algebraic formulation of the central problem of screw theory, namely the determination of the principal screws of a given system. Using the algebra of dual numbers, it shows that the principal screws can be determined via the solution of a generalised eigenproblem of two real, symmetric matrices. This approach allows the study of the principal screws of the general screw systems associated with a manipulator of arbitrary geometry in terms of closed-form expressions of its architecture and configuration parameters. The formulation is illustrated with examples of practical manipulators.

A near-singular, flexure jointed, moment sensitive Stewart platform based fore-torque sensor

A force-torque sensor capable of accurate measurement of the three components of externally applied forces and moments is required for force control in robotic applications involving assembly operations. The goal in this paper is to design a Stewart platform based force torque sensor at a near-singular configuration sensitive to externally applied moments. In such a configuration, we show an enhanced mechanical amplification of leg forces and thereby higher sensitivity for the applied external moments. In other directions, the sensitivity will be that of a normal load sensor determined by the sensitivity of the sensing element and the associated electronic amplification, and all the six components of the force and torque can be sensed. In a sensor application, the friction, backlash and other non-linearities at the passive spherical joints of the Stewart platform will affect the measurements in unpredictable ways. In this sensor, we use flexural hinges at the leg interfaces of the base and platform of the sensor. The design dimensions of the flexure joints in the sensor have been arrived at using FEA. The sensor has been fabricated, assembled and instrumented. It has been calibrated for low level loads and is found to show linearity and marked sensitivity to moments about the three orthogonal X, Y and Z axes. This sensor is compatible for usage as a wrist sensor for a robot under development at ISRO Satellite Centre.

Whole arm manipulation using closed and hybrid planar chains: A kinematic study

This paper studies planar whole arm manipulation of a circular object using closed loop and hybrid manipulators. The manipulation is simple with fewer degrees of actuation than the task space. This is an useful operation if the initial and final positions of the object are on the same surface. Closed loop manipulator is a 4/5 bar mechanism. In hybrid manipulators a open loop manipulator with 3/4 links is attached to the floating link of 4/5 bar mechanism. The mobility analysis is carried out to find the connectivity of the object with reference to frame. The manipulation (forward kinematics) starts from a given configuration of the object and the manipulator. In hybrid manipulators determination of initial configuration involves inverse kinematics of open loop manipulator. The input joint velocities are used to demonstrate the manipulation. Conditions are specified for prehensile manipulation.

Simultaneous source localization and boundary mapping for contaminants

This paper addresses the problem of localizing the sources of contaminants spread in the environment, and mapping the boundary of the affected region using an innovative swarm intelligence based technique. Unlike most work in this area the algorithm is capable of localizing multiple sources simultaneously while also mapping the boundary of the contaminant spread. At the same time the algorithm is suitable for implementation using a mobile robotic sensor network. Two types of agents, called the source localization agents (or S-agents) and boundary mapping agents (or B-agents) are used for this purpose. The paper uses the basic glowworm swarm optimization (GSO) algorithm, which has been used only for multiple signal source localization, and modifies it considerably to make it suitable for both these tasks. This requires the definition of new behaviour patterns for the agents based on their terminal performance as well as interactions between them that helps the swarm to split into subgroups easily and identify contaminant sources as well as spread along the boundary to map its full length. Simulations results are given to demonstrate the efficacy of the algorithm.