Model Predictive Control (MPC) has emerged as a pivotal technology for optimizing control tasks in autonomous driving, particularly within Adaptive Cruise Control (ACC) systems. However, the manual tuning of MPC cost function weights and prediction horizons remains a significant challenge. In this paper, we introduce a novel framework that combines Deep Reinforcement Learning (DRL) with MFC to dynamically tune both the weight parameters and prediction horizon in real time. This approach, referred to as the Weights and Prediction Horizon Varying MPC (W-PH-MPC), overcomes traditional MPC limitations by utilizing proximal Policy optimisation and Deep Deterministic Policy Gradient (DDPG) algorithms to adjust control parameters. We evaluate the effectiveness of our approach through simulations in vehicle-tracking scenarios. Simulation results show that the adaptive MPC-RL controller achieves better tracking performance, without compromising power consumption, and lowers longitudinal jerk compared to a fixed-parameter MPC baseline, resulting in smoother and more efficient vehicle behavior.

In 3D object mapping, category-level priors enable efficient object reconstruction and canonical pose estimation, requiring only a single prior per semantic category (e.g., chair, book, laptop, etc.). DeepSDF has been used predominantly as a category-level shape prior, but it struggles to reconstruct sharp geometry and is computationally expensive. In contrast, NeRFs capture fine details but have yet to be effectively integrated with category-level priors in a real-time multi-object mapping framework. To bridge this gap, we introduce PRENOM, a Prior-based Efficient Neural Object Mapper that integrates category-level priors with object-level NeRFs to enhance reconstruction efficiency and enable canonical object pose estimation. PRENOM gets to know objects on a first-name basis by meta-learning on synthetic reconstruction tasks generated from open-source shape datasets. To account for object category variations, it employs a multi-objective genetic algorithm to optimize the NeRF architecture for each category, balancing reconstruction quality and training time. Additionally, prior-based probabilistic ray sampling directs sampling toward expected object regions, accelerating convergence and improving reconstruction quality under constrained resources. Experimental results highlight the ability of PRENOM to achieve high-quality reconstructions while maintaining computational feasibility. Specifically, comparisons with prior-free NeRF-based approaches on a synthetic dataset show a 21% lower Chamfer distance. Furthermore, evaluations against other approaches using shape priors on a noisy real-world dataset indicate a 13% improvement averaged across all reconstruction metrics, and comparable pose and size estimation accuracy, while being trained for 5× less time.Code available at: https://github.com/snt-arg/PRENOM

This paper presents an integrated Reinforcement Learning (RL) and Model Predictive Control (MPC) framework for autonomous satellite docking with a partially filled fuel tank. Traditional docking control faces challenges due to fuel sloshing in microgravity, which induces unpredictable forces affecting stability. To address this, we integrate Proximal Policy Optimization (PPO) and Soft Actor-Critic (SAC) RL algorithms with MPC, leveraging MPC’s predictive capabilities to accelerate RL training and improve control robustness. The proposed approach is validated through Zero-G Lab of SnT experiments for planar stabilization and high-fidelity numerical simulations for 6-DOF docking with fuel sloshing dynamics. Simulation results demonstrate that SAC-MPC achieves superior docking accuracy, higher success rates, and lower control effort, outperforming standalone RL and PPO-MPC methods. This study advances fuel-efficient and disturbance-resilient satellite docking, enhancing the feasibility of on-orbit refueling and servicing missions.

Recent works on SLAM extend their pose graphs with higher-level semantic concepts like Rooms exploiting relationships between them, to provide, not only a richer representation of the situation/environment but also to improve the accuracy of its estimation. Concretely, our previous work, Situational Graphs (S-Graphs+), a pioneer in jointly leveraging semantic relationships in the factor optimization process, relies on semantic entities such as Planes and Rooms, whose relationship is mathematically defined. Nevertheless, there is no unique approach to finding all the hidden patterns in lower-level factor-graphs that correspond to high-level concepts of different natures. It is currently tackled with ad-hoc algorithms, which limits its graph expressiveness.To overcome this limitation, in this work, we propose an algorithm based on Graph Neural Networks for learning high-level semantic-relational concepts that can be inferred from the low-level factor graph. Given a set of mapped Planes our algorithm is capable of inferring Room entities relating to the Planes. Additionally, to demonstrate the versatility of our method, our algorithm can infer an additional semantic-relational concept, i.e. Wall, and its relationship with its Planes. We validate our method in both simulated and real datasets demonstrating improved performance over two baseline approaches. Furthermore, we integrate our method into the S-Graphs+ algorithm providing improved pose and map accuracy compared to the baseline while further enhancing the scene representation.

This paper explores the potential of 5G new radio (NR) Time-of-Arrival (TOA) data for indoor drone localization under different scenarios and conditions when fused with inertial measurement unit (IMU) data. Our approach involves performing graph-based optimization to estimate the drone’s position and orientation from the multiple sensor measurements. Due to the lack of real-world data, we use Matlab 5G toolbox and QuaDRiGa (quasi-deterministic radio channel generator) channel simulator to generate TOA measurements for the EuRoC MAV indoor dataset that provides IMU readings and ground truths 6DoF poses of a flying drone. Hence, we create twelve sequences combining three predefined indoor scenarios setups of QuaDRiGa with 2 to 5 base station antennas. Therefore, experimental results demonstrate that, for a sufficient number of base stations and a high bandwidth 5G configuration, the pose graph optimization approach achieves accurate drone localization, with an average error of less than 15 cm on the overall trajectory. Furthermore, the adopted graph-based optimization algorithm is fast and can be easily implemented for onboard real-time pose tracking on a micro aerial vehicle (MAV).

In this paper, we propose a solution for legged robot localization using architectural plans. Our specific contributions towards this goal are several. Firstly, we develop a method for converting the plan of a building into what we denote as an architectural graph (A-Graph). When the robot starts moving in an environment, we assume it has no knowledge about it, and it estimates an online situational graph representation (S-Graph) of its surroundings. We develop a novel graph-to-graph matching method, in order to relate the S-Graph estimated online from the robot sensors and the A-Graph extracted from the building plans. Note the challenge in this, as the S-Graph may show a partial view of the full A-Graph, their nodes are heterogeneous and their reference frames are different. After the matching, both graphs are aligned and merged, resulting in what we denote as an informed Situational Graph (is-Graph), with which we achieve global robot localization and exploitation of prior knowledge from the building plans. Our experiments show that our pipeline shows a higher robustness and a significantly lower pose error than several LiDAR localization baselines. Paper Video: https://youtu.be/3Pv7y8aOsUY

Fiducial markers can encode rich information about the environment and aid Visual SLAM (VSLAM) approaches in reconstructing maps with practical semantic information. Current marker-based VSLAM approaches mainly utilize markers for improving feature detections in low-feature environments and/or incorporating loop closure constraints, generating only low-level geometric maps of the environment prone to inaccuracies in complex environments. To bridge this gap, this paper presents a VSLAM approach utilizing a monocular camera along with fiducial markers to generate hierarchical representations of the environment while improving the camera pose estimate. The proposed approach detects semantic entities from the surroundings, including walls, corridors, and rooms encoded within markers, and appropriately adds topological constraints among them. Experimental results on a real-world dataset collected with a robot demonstrate that the proposed approach outperforms a marker-based VSLAM baseline in terms of accuracy, given the addition of new constraints while creating enhanced map representations. Furthermore, it shows satisfactory results when comparing the reconstructed map quality to the one rebuilt using a LiDAR SLAM approach.

For mobile robots to operate in an autonomous and safe manner they must be able to adequately perceive their environment despite challenging or unpredictable conditions in their sensory apparatus. Usually, this is addressed through ad-hoc, not easily generalizable Fault Detection and Diagnosis (FDD) approaches. In this work, we leverage Bayesian Networks (BNs) to propose a novel probabilistic inference architecture that provides generality, rigorous inferences and real-time performance for the detection, diagnosis and recovery of diverse and multiple sensory failures in robotic systems. Our proposal achieves all these goals by structuring a BN in a multidimensional setting that up to our knowledge deals coherently and rigorously for the first time with the following issues: modeling of complex interactions among the components of the system, including sensors, anomaly detection and recovery; representation of sensory information and other kinds of knowledge at different levels of cognitive abstraction; and management of the temporal evolution of sensory behavior. Real-time performance is achieved through the compilation of these BNs into feedforward neural networks. Our proposal has been implemented and tested for mobile robot navigation in environments with human presence, a complex task that involves diverse sensor anomalies. The results obtained from both simulated and real experiments prove that our architecture enhances the safety and robustness of robotic operation: among others, the minimum distance to pedestrians, the tracking time and the navigation time all improve statistically in the presence of anomalies, with a diversity of changes in medians ranging from ≃20% to ≃500%.

Robot self-localization is essential for operating autonomously in open environments. When cameras are the main source of information for retrieving the pose, numerous challenges are posed by the presence of dynamic objects, due to occlusion and continuous changes in the appearance. Recent research on global localization methods focused on using a single (or multiple) Convolutional Neural Network (CNN) to estimate the 6 Degrees of Freedom (6-DoF) pose directly from a monocular camera image. In contrast with the classical approaches using engineered feature detector, CNNs are usually more robust to environmental changes in light and to occlusions in outdoor scenarios. This paper contains an attempt to empirically demonstrate the ability of CNNs to ignore dynamic elements, such as pedestrians or cars, through learning. For this purpose, we pre-process a dataset for pose localization with an object segmentation network, masking potentially moving objects. Hence, we compare the pose regression CNN trained and/or tested on the set of masked images and the original one. Experimental results show that the performances of the two training approaches are similar, with a slight reduction of the error when hiding occluding objects from the views.

In this paper we apply deep reinforcement learning techniques on a multicopter for learning a stable hovering task in a continuous state action environment. We present a framework based on OpenAI GYM, Gazebo, Robotic Operating System and RotorS MAV simulator, used for successfully training different agents to perform various tasks. The deep reinforcement learning method used for the training is a model-free, on-policy, actor-critic based algorithm called Trust Region Policy Optimization (TRPO). Two neural networks have been used as nonlinear function approximators. Our experiments show that such learning approach achieves successful results, and facilitates the process of controller design.

This paper proposes an unified and advanced framework for the modeling, stability study and stabilization of a Power Networks subject to an omniscient adversary (i.e. cyber-attack). From the system model developed in [24], based on the well-known sector non-linearity approach and the convex polytopic transformation, the attacked system (descriptor model) is re-written in a more convenient form (Linear Parameter Varying-LPV) with unmeasurable premise variables. The so-called Lyapunov-based methods are applied in order to study the stability and security problems despite the presence of cyber-attacks. The conditions will be given in terms of Linear-Bilinear Matrix Inequality LMI-BMI constraints.

The reliability of aircraft inspection is of paramount importance to safety of flights. Continuing airworthiness of aircraft structures is largely based upon the visual detection of small defects made by trained inspection personnel with expensive, critical and time consuming tasks. At this aim, Unmanned Aerial Vehicles (UAVs) can be used for autonomous inspections, as long as it is possible to localize the target while flying around it and correct the position. This work proposes a solution to detect the airplane pose with regards to the UAVs position while flying autonomously around the airframe at close range for visual inspection tasks. The system works by processing images coming from an RGB camera mounted on board, comparing incoming frames with a database of natural landmarks whose position on the airframe surface is known. The solution has been tested in real UAV flight scenarios, showing its effectiveness in localizing the pose with high precision. The advantages of the proposed methods are of industrial interest since we remove many constraint that are present in the state of the art solutions.

Autonomous robots are already being used, for example, as tour guides, receptionists, or office-assistants. The proximity to humans and the possibility to physically interact with them highlights the importance of developing secure robot applications. It is crucial to consider security implications to be an important part of the robot application's development process. Adding security later in the application's life-cycle usually leads to high costs, or is not possible due to earlier design decisions. In this work, we present the Robot Application Security Process (RASP) as a lightweight process that enables the development of secure robot applications. Together with RASP we introduce the role of a Security Engineer (SecEng) as an important stakeholder in any robot application development process. RASP enables the SecEng to verify the completeness of his work and allows him to argue about the application's security with other stakeholders. Furthermore, we demonstrate how the RASP supports the SecEng and also other developers in their daily work.

In the present paper, the problem of attacks on cyber-physical systems via networked control system (NCS) subject to unmeasured disturbances is considered. The geometric approach is used to evaluate the security and vulnerability level of the controlled system. The presented work deals with the so-called false data injection attacks and shows how imperfectly known disturbances can be used to perform undetectable, or at least stealthy, attacks that can make the NCS vulnerable to attacks from malicious outsiders. A numerical example is given to illustrate the approach.

Traditional planning and scheduling techniques still hold important roles in modern smart scheduling systems. Realistic features present in modern manufacturing systems need to be incorporated into these techniques. The real-world problem addressed here is an extension of flexible job shop scheduling problem and is issued from the modern printing and boarding industry. The precedence between operations of each job is given by an arbitrary directed acyclic graph rather than a linear order. In this paper, we extend the traditional FJSP solutions representation to address the parallel operations. We propose an imperialist competitive algorithm for the problem. Several instances are used for the experiments and the results show that, for the considered instances, the proposed algorithm is faster and found better or equal solutions compared to the state-of-the-art algorithms.

Object detection and tracking is a challenging task, especially for unmanned aerial robots in complex environments where both static and dynamic objects are present. It is, however, essential for ensuring safety of the robot during navigation in such environments. In this work we present a practical online approach which is based on a 2D LIDAR. Unlike common approaches in the literature of modeling the environment as 2D or 3D occupancy grids, our approach offers a fast and robust method to represent the objects in the environment in a compact form, which is significantly more efficient in terms of both memory and computation in comparison with the former. Our approach is also capable of classifying objects into categories such as static and dynamic, and tracking dynamic objects as well as estimating their velocities with reasonable accuracy.

This paper is presenting the implementation and experimental validation of the cooperative robot localization framework “Atlas”. For ease of application, Atlas is implemented as a package for the Robot Operating System (ROS). ATLAS is based on dynamic cooperative sensor fusion which optimizes the estimated pose with respect to noise, respective variance. This paper validates the applicability of Atlas by cooperatively localizing multiple real quadrotors using cameras and fiduciary markers.

In this note, we design an invariant observer for a two step (acidogenesis-methanogenesis) mass balance non linear model, in order to estimate simultaneously all bacteria and substrate concentrations found in the anaerobic digestion process. The particularity of the designed observer is the use of only the methane flow rate which is cheap to measure and commonly measured online even at industrial scale.

During the past few years, with the fast development and proliferation of the Internet of Things (IoT), many application areas have started to exploit this new computing paradigm. The number of active computing devices has been growing at a rapid pace in IoT environments around the world. Consequently, a mechanism to deal with this different devices has become necessary. Middleware systems solutions for IoT have been developed in both research and industrial environments to supply this need. However, decision analytics remain a critical challenge. In this work we present the Decision Support IoT Framework composed of COBASEN, an IoT search engine to address the research challenge regarding the discovery and selection of IoT devices when large number of devices with overlapping and sometimes redundant functionality are available in IoT middleware systems, and DMS, a rule-based reasoner engine allowing to set up computational analytics on device data when it is still in motion, extracting valuable information from it for automated decision making. DMS uses Complex Event Processing to analyze and react over streaming data, allowing for example, to trigger an actuator when a specific error or condition appears in the stream. The main goal of this work is to highlight the importance of a decision support system for decision analytics in the IoT paradigm. We developed a system which implements DMS concepts. However, for preliminarily tests, we made a functional evaluation of both systems in terms of performance. Our initial findings suggest that the Decision Support IoT Framework provides important approaches that facilitate the development of IoT applications, and provides a new way to see how the business rules and decision-making will be made towards the Internet of Things.

Unmanned Aerial Vehicles are currently investigated as an important sub-domain of robotics, a fast growing and truly multidisciplinary research field. UAVs are increasingly deployed in real-world settings for missions in dangerous environments or in environments which are challenging to access. Combined with autonomous flying capabilities, many new possibilities, but also challenges, open up. To overcome the challenge of early identification of degradation, machine learning based on flight features is a promising direction. Existing approaches build classifiers that consider their features to be correlated. This prevents a fine-grained detection of degradation for the different hardware components. This work presents an approach where the data is considered uncorrelated and, using machine learning techniques, allows the precise identification of UAV's damages.

peer reviewed

In this paper, we propose to control the quantity and quality of the produced biogas from the anaerobic digestion of organic matter, digested in either a continuous stirred tank reactor or a fixed bed digester. This is motivated by the aim of providing the power grid with a stable amount of energy despite fluctuations in the treated waste concentration and composition. Therefore, we apply the linearizing control principe to a two step (acidogenesis-methanogenesis) mass balance non linear model, all with the introduction of two new control inputs reflecting the addition of stimulating substrates (acetate and alkalinity). To show the performance of the synthesized control laws we simulate the process under an organic shock load feeding.

To perform a wide range of tasks service robots need to robustly extract knowledge about the world from the data perceived through the robot's sensors even in the presence of varying context-conditions. This makes the design and development of robot perception architectures a challenging exercise. In this paper we propose a robot perception architecture which enables to select and execute at runtime different perception graphs based on monitored context changes. To achieve this the architecture is structured as a feedback loop and contains a repository of different perception graph configurations suitable for various context conditions.

This article addresses the control problem of robots with unknown dynamics and manipulating flexible rubber objects of unknown elasticity. The manipulated rubber object is considered to be interacting with arbitrarily-switched constraints. Such a kind of robot system is shown to have switched impedance parameters during a task execution that results in an unknown hybrid nonlinear system with arbitrarily switched signal. A Model-Free Robust Adaptive Control (MFRAC) strategy is proposed for such a robot system that is proved to guarantee global stable performance with all closed loop signals are assured to be bounded. The suggested MFRAC strategy relies on the synergy of the Adaptive Fuzzy System (AFS), the Sliding Mode Control (SMC), and the notion of Common Lyapunov Functions (CLF). The AFS relaxes the need for knowing the precise robot dynamics, the SMC adds robustness against the drift of the dynamics parameters, and the CLF accommodates the arbitrary switching of the impedance parameters. The bounds of the impedance parameters are adapted online and incorporated in the MFRAC design such that a convergent performance is achieved. Experiment is conducted on a KUKA Lightweight Robot (LWR) doing flexible rubber peg-in-hole assembly process that falls in the category of systems considered in this article. From the experimental results, excellent tracking performance is reported when using the proposed MFRAC strategy for the considered robotic system despite the dynamics anonymity and the unknown impedance parameters arbitrary switching.

Robot interaction with the environment is normally described as a mass-spring-damping impedance model and the estimation of such interaction impedance parameters requires the computation of the joint (or task) space velocity and acceleration. In many cases, the velocity and acceleration are computed by numerically computing the first and second derivatives of the sensed position signal. The numerical differentiation results in approximation errors when computing the velocity and acceleration signals that would have a direct impact on the estimation of the impedance parameters. This article proposes enhancing the estimation of the impedance parameters by smoothing the velocity and acceleration signals prior to the considered estimation process. Gaussian Smoothing Filter (GSF) is employed in smoothing the considered signals. After the smoothing process, impedance parameters estimation becomes more feasible using the available strategies like the Least Mean Square (LMS) or any other estimation approach. Experiments are conducted on a KUKA Lightweight Robot (LWR) doing the assembly of the air-intake manifold of an automotive powertrain. The impedance parameters are estimated for the smoothed and unsmoothed cases in order to show the enhancement in the estimation process.

In all the processes of Wind Energy (WE) utilization, the Wind Power (WP) assessment is critical stage for all the Wind Farms (WFs). This paper is focused on the WE systems in Luxembourg. It describes the overview of the wind resources in all the WFs and presents an Unified Cooperation Wake Model (UCWM) and Coordination and Optimization Control (CnOC) for WFs. Based on WP assessment of WFs, the statistical method is used to model the distribution of wind speed and Wind Direction (WD). Some simulation figures about the wind rose andWeibull distribution demonstrate the detailed description and assessment of WP. These assessments are expected to enhance the effectiveness of WP exploitation and utilization in WFs of Luxembourg.

This paper presents a distributed world model that is able to adapt to changes in the Quality of Service (QoS) of the communication layer by online reconfiguration of perception algorithms. The approach consists of (a) a mechanism for storage, exchange and processing of world model data and (b) a feedback loop that incorporates reasoning techniques to adapt to QoS changes immediately. The latter introduces a Level of Detail (LoD) metric based on a spatial resolution in order to infer an upper bound for the amount of data that can be transmitted without violating an application specific transmission delay. Experiments have been performed with Octree-based subsampling techniques applied to data originating from a RGB-D camera using simulated and real-world data sets for timevarying bandwidth values as employed QoS measure.

In this paper an H∞decentralized observer-based control is proposed for large-scale uncertain nonlinear systems. These systems are coupled by N interconnected subsystems where the interconnections satisfy the quadratic constraints. The proposed control is based on a new form of dynamic observer (DO) which generalizes the existing results on the proportional observer (PO) and the proportional integral observer (PIO). The design approach is derived from the solution of matrix inequality and based on the algebraic constraints obtained from the analysis of the estimation error. A numerical example is provided to show the effectiveness of the proposed control.

In this paper a safety-oriented model based software architecture for robotic solutions is proposed. The main focus herein is to consider aspects such as real-time, heterogeneity, deployment, modeling and analysis of emerging effects as well as functional safety and to combine all aspects into an overall development approach. The architecture shall capture the complexity caused by the autonomy and mobility of the robot and support the developer with a suitable chain of evidence especially suited for the safety relevant functions. A use case comprising a lightweight robotic manipulator which will be integrated in a mobile service robot underlines the feasibility of this approach.

In this paper, a safety oriented model based software engineering process for autonomous robots is proposed. Herein, the main focus is on the modeling of the safety case based on the standard ISO/DIS 13482. Combined with a safe multilayer robot software architecture it allows to trace the safety requirements and to model safety relevant properties on the early design stages in order to build a reliable chain of evidence. The introduced engineering processes consist of the Domain Engineering, which is dealing with the development of a set of interlinked formalized safety cases and software components. Finally, the proposed engineering process is demonstrated on the example of the assembly assistant robot and ROS (Robot Operating System).

In microproduction, i.e. in the production and assembly of micro-scale components and products, fully automated systems hardly exist so far. Besides the requirements of handling small parts with extreme precision, small batch sizes of highly customized products are among the main challenges. Therefore, economic microproduction requires very flexible production systems with a high level of automation. This contribution proposes a new concept of such a system that provides two main innovations. First, the proposed concept integrates stationary production machines and mobile transport robots in order to configure rapidly changing production processes in real-time. Besides this distributed flexible system structure, also the overall automation system consisting of the Manufacturing Execution System (MES) and the shop floor control is designed in a distributed form as a multiagent system. This distributed automation system is especially suited to automate flexible production scheduling and resource allocation processes but also integrates the multi-robot transport system.

A nonlinear problem formulation of an energy-saving model-predictive eco-cruise controller for electric vehicles is presented. With regard to the intended application in real-world tests, the model has to include the specific properties of a serial electric vehicle such as energy-recovery and a discontinuous accelerator input giving rise to a binary control variable. These specific features and the nonlinearity of the system dynamics make it a challenging task to formulate the optimisation problem in a way that allows a fast computation in real-time application. The challenges are addressed by using a model of the vehicle dynamics that is formulated in terms of the vehicle position instead of time and by considering the kinetic energy instead of the velocity. Furthermore, various constraints on the input and state variables are introduced for a realistic representation of the vehicle characteristics. A special focus is put on the treatment of a binary input variable in the optimisation. Here, in order to avoid a mixed-integer formulation of the problem, a continuous variable is introduced which is forced to take only discrete values by a penalty term. Finally, first simulation results underline the feasibility of this control approach.

Future autonomous service robots are intended to operate in open and complex environments. This in turn implies complications ensuring safe operation. The tenor of few available investigations is the need for dynamically assessing operational risks. Furthermore, a new kind of hazards being implicated by the robot's capability to manipulate the environment occurs: hazardous environmental object interactions. One of the open questions in safety research is integrating safety knowledge into robotic systems, enabling these systems behaving safety-conscious in hazardous situations. In this paper a safety procedure is described, in which learning of safety knowledge from human demonstration is considered. Within the procedure, a task is demonstrated to the robot, which observes object-to-object relations and labels situational data as commanded by the human. Based on this data, several supervised learning techniques are evaluated used for finally extracting safety knowledge. Results indicate that Decision Trees allow interesting opportunities.

Small, low-altitude unmanned aerial vehicles (UAV)s can be very useful in many ecological applications as a personal remote sensing platform. However, in many cases it is difficult to produce a single georeferenced mosaic from the many small images taken from the UAV. This is due to the lack of features in the images and the inherent errors from the inexpensive navigation sensors. This paper focuses on improving the orthorectification accuracy by finding these errors and calibrating the navigation sensors. This is done by inverse-orthorectifying a set of images collected during flight using ground targets and General Procrustes Analysis. By comparing the calculated data from the inverse-orthorectification and the measured data from the navigation sensors, different sources of errors can be found and characterized, such as GPS computational delay, logging delay, and biases. With this method, the orthorectification errors are reduced from less than 60m to less than 1.5m.

Autonomous systems are often needed to perform tasks in complex and dynamic environments. For this class of systems, traditional safety assuring methods are not satisfying due to the unknown effects of the interacting system with an open environment. Briefly speaking: What is not known during the development phase, can not be adequately considered. In order to realize a more flexible safety analysis, the internal representation of the outside world to be learned by an autonomous Cognitive Technical System, is used to identify hazardous situations. The so-called safety principles represent the hazard knowledge. These can be added to the system prior to operating time without losing the possibility of adjusting or expanding this hazard knowledge during operating time. This contribution details a new method for safety assurance and therefore proposes the introduction of so-called safety principles. Furthermore, the Cognitive Technical System provides anticipation capabilities, so that is becomes possible to expand the planning process in order to take hazard information into account. Finally, a simulation example demonstrates how the autonomous system determines possible future actions, evaluating them with regard to hazards in order to provide a plan with acceptable risk. Nevertheless, the approach can also be implemented to real world applications since typical real world phenomena as uncertainty and faults can also be considered in the chosen virtual world figuratively.

In an industrial nonwoven production process, data related to spun-web quality including firmness, thickness, and weight are measured online. However, these measurements provide insufficient information about the spun-web quality. As a result, other parameters that give a better insight into the quality of the product have been employed. One of these parameters reveals the minimum force necessary to pull apart a given size of piece of spun-web. It is measured in a quality control laboratory by stretching the ends of a sample piece of spun-web to a high degree of tension. Since the laboratory measurements are infrequent and off-line, an on-line monitoring of product quality can hardly be carried out. Neural network-based virtual sensors are developed based on historical data that incorporate both online and off-line measurements. The virtual sensors provide estimates of key quality parameters at the measurement frequency of the on-line measurements. They are implemented at a production facility of Freudenberg Nonwovens KG in Germany and have been successfully employed for on-line quality monitoring.