A5001137240- Target Tracking and Data Fusion in Sensor Networks
- Indoor and Outdoor Localization Technologies
- Robotics and Sensor-Based Localization
- Inertial Sensor and Navigation
- Fault Detection and Control Systems
- Speech and Audio Processing
- Gaussian Processes and Bayesian Inference
- Gait Recognition and Analysis
- Control Systems and Identification
- Distributed Sensor Networks and Detection Algorithms
- Context-Aware Activity Recognition Systems
- Underwater Acoustics Research
- GNSS positioning and interference
- Underwater Vehicles and Communication Systems
- Non-Invasive Vital Sign Monitoring
- Structural Health Monitoring Techniques
- Advanced Adaptive Filtering Techniques
- Video Surveillance and Tracking Methods
- 3D Surveying and Cultural Heritage
- Robotic Path Planning Algorithms
- Bayesian Modeling and Causal Inference
- Anomaly Detection Techniques and Applications
- Guidance and Control Systems
- Distributed Control Multi-Agent Systems
- Human Pose and Action Recognition
Linköping University
2015-2024
Swedish Defence Research Agency
2011-2023
Université Grenoble Alpes
2023
Uppsala University
2023
Institut polytechnique de Grenoble
2023
Grenoble Images Parole Signal Automatique
2023
Grain Inspection, Packers and Stockyards Administration
2023
Centre Inria de l'Université Grenoble Alpes
2023
German Research Centre for Artificial Intelligence
2009-2011
The unscented Kalman filter (UKF) has become a popular alternative to the extended (EKF) during last decade. UKF propagates so called sigma points by function evaluations using transformation (UT), and this is at first glance very different from standard EKF algorithm which based on linearized model. claimed advantages with are that it two moments of posterior distribution does not require gradients system We point out several less known links between in terms conceptually implementations...
Safety and security applications benefit from better situational awareness. Radar micro‐Doppler signatures an observed target carry information about the target's activity, have potential to improve This article describes, compares, discusses two methods classify human activity based on radar data. The first method extracts physically interpretable features time‐velocity domain such as main cycle time properties of envelope spectra use these in classification. second derives its components...
Micro-Doppler radar signatures have great potential for classifying pedestrians and animals, as well their motion pattern, in a variety of surveillance applications. Due to the many degrees freedom involved, real data need be complemented with accurate simulated able successfully design test signal processing algorithms. In cases, ability collect is limited by monetary practical considerations, whereas environment, any desired scenario may generated. Motion capture (MOCAP) has been used...
The particle filter(PF) has during the last decade been proposed for a wide range of localization and tracking applications. There is general need in such embedded system to have platform efficient scalable implementation PF. One graphics processing unit (GPU), originally aimed be used fast rendering graphics. To achieve this, GPUs are equipped with parallel architecture which can exploited general-purpose computing on GPU (GPGPU) as complement central (CPU). In this paper, GPGPU techniques...
Today, the workflows that are involved in industrial assembly and production activities becoming increasingly complex. To efficiently safely perform these is demanding on workers, particular when it comes to infrequent or repetitive tasks. This burden workers can be eased by introducing smart assistance systems. article presents a scalable concept an integrated system demonstrator designed for this purpose. The basic idea learn from observing multiple expert operators then transfer learnt...
A sensor management method for joint multitarget search and track problems is proposed, where a single user-defined parameter allows tradeoff between the two objectives. The density propagated using Poisson multi-Bernoulli mixture filter, which eliminates need separate handling of undiscovered targets provides theoretical foundation unified method. Monte Carlo simulations scenarios are used to evaluate performance proposed
Regular and moderate physical activity practice provides many physiological benefits. It reduces the risk of disease outcomes is basis for proper rehabilitation after a severe disease. Aerobic strength exercises are strongly reco
Classification of motion mode (walking, running, standing still) and device (hand-held, in pocket, backpack) is an enabler personal navigation systems for the purpose saving energy design parameter settings also its own sake. Our main contribution to publish one most extensive datasets this problem, including inertial data from eight users, each performing three pre-defined trajectories carrying four smartphones seventeen measurement units on body. All kind metadata available such as ground...
We study the fundamental problem of fusing one round trip time (RTT) observation associated with a serving base station time-difference arrival (TDOA) to and neighbor localize 2-D mobile (MS). This situation can arise in 3GPP Long Term Evolution (LTE) when number reported cells is reduced minimum order minimize signaling costs support large devices. The studied corresponds geometrically computing intersection circle hyperbola, both measurement uncertainty, which generally has two equally...
In the context of a smart user assistance system for industrial manipulation tasks it is necessary to capture motions upper body and limbs worker in order derive his or her interactions with task space. While such capturing technology already exists, novelty proposed work results from strong requirements application context: The method should be flexible use only on-body sensors, accurately environments that suffer severe magnetic disturbances, enable consistent registration between frame...
The monitoring of physical activities under realistic, everyday life conditions thus while an individual follows his regular daily routine is usually neglected or even completely ignored. Therefore, this paper investigates the development and evaluation robust methods for scenarios,
This paper presents a method for global pose estimation using inertial sensors, monocular vision, and ultra wide band (UWB) sensors. It is demonstrated that the complementary characteristics of these sensors can be exploited to provide improved estimates, without requiring introduction any visible infrastructure, such as fiducial markers. Instead, natural landmarks are jointly estimated with platform simultaneous localization mapping framework, supported by small number easy-to-hide UWB...
Physical activity monitoring has recently become an important topic in wearable computing, motivated by e.g. healthcare applications. However, new benchmark results show that the difficulty of complex classification problems exceeds potential existing classifiers. Therefore, this paper proposes ConfAdaBoost.M1 algorithm. The proposed algorithm is a variant AdaBoost.M1 incorporates well established ideas for confidence based boosting. method compared to most commonly used boosting methods...
The ensemble Kalman filter (EnKF) is a Monte Carlo based implementation of the (KF) for extremely high-dimensional, possibly nonlinear and non-Gaussian state estimation problems. Its ability to handle dimensions in order millions has made EnKF popular algorithm different geoscientific disciplines. Despite similarly vital need scalable algorithms signal processing, e.g., make sense ever increasing amount sensor data, hardly discussed our field. This self-contained review paper aimed at...
A computationally efficient method for online joint state inference and dynamical model learning is presented. The combines an a priori known, physically derived, state-space with radial basis function expansion representing unknown system dynamics inherits properties from both physical data-driven modeling. uses extended Kalman filter approach to jointly estimate the of learn dynamics, via parameters expansion. key contribution computational complexity reduction compared similar globally...
A class of nonlinear transformation-based filters (NLTF) for state estimation is proposed. The transformations that can be used include first (TT1) and second (TT2) order Taylor expansions, the unscented transformation (UT), Monte Carlo (MCT) approximation. Kalman filter (UKF) by construction a special case, but also nonstandard implementations (KF) extended (EKF) are included, where there no explicit Riccati equations. theoretical properties these mappings important performance NLTF. TT2...
For computational efficiency, it is important to utilize model structure in particle filtering. One of the most cases occurs when there exists a linear Gaussian substructure, which can be efficiently handled by Kalman filters. This standard formulation Rao-Blackwellized filter (RBPF). contribution suggests an alternative this well-known result that facilitates reuse filtering components and also suitable for object-oriented programming. Our RBPF seen as bank with stochastic branching pruning.