Gustaf Hendeby
Associate Professor and Docent in Automatic Control
(Swe: Universitetslektor och docent i reglerteknik)
- On-Line Appearance:
-
©2026. All rights reserved.
©2026. All rights reserved.
Dr Gustaf Hendeby’s research is centered around statistical sensor fusion - primarily model based methods - with one foot in theory and one foot in more applied research. Sensor fusion provides methods that are used to combine information from many sources in order to produce a coherent view of phenomena observed in the world not obtainable from the separate sources on their own. With recent advances in sensor technology, resulting in smaller and more affordable sensors, a multitude of sensors are available almost everywhere; and much in life that we have learned to take for granted would not be possible without advanced sensor fusion. This knowledge drives his interest in the field.
Below he present his research divided into three main categories: core algorithm understanding, on-body sensor networks, and security and surveillance applications.
From the beginning, Dr Hendeby have been interested in the inner working of the algorithms that make up sensor fusion. By studying the Cramér-Rao lower bound (CRLB), he derived a method using intrinsic accuracy to indicate the potential performance gain from using nonlinear estimation and detection methods. Intrinsic accuracy is a noise property, which is closely related to Fisher information, that can be interpreted as measure of how informative a noise distribution is. In practice it acts like an effective variance term. To analyze the second order statistical properties of algorithms this way is very common, but does also in many situations not capture all aspects of an estimate that are important for the end result. By utilizing the above mentioned results, it is possible to provide insights into how to approach problems, and to beforehand answer the important question: “Is it reasonable to expect the results I need to solve this specific task?” This without having to resort to extensive Monte Carlo simulations.
In the same way, understanding how popular methods relate to each other makes it possible to make better design decisions early in the design process. Therefore, Dr Hendeby has analyzed the popular unscented Kalman filter (UKF) and managed to find clear connections to the classic extended Kalman filter (EKF), which in part explains its behavior. He currently pursue similar questions with Lic Eng Michael Roth. He has also analyzed the Rao-Blackewellized particle filter (RBPF), and was able to make connections between it and the important class of methods that are based on filter banks. Interpreting the RBPF this slightly different way can for instance be utilized to make efficient generalizable implementations.
Dr Hendeby’s interest for the particle filters (PFs) also resulted in the first complete PF implementation on a graphics card (GPU), which is a first step to make the PF realizable on low-cost readily available parallel hardware. A current fucus is to investigate the ensemble Kalman filter (EnKF), another approximate stochastic method that is heavily used in the geo-sciences, but so far not by the signal processing society. It could have the potential to solve extremely high-dimensional problems that would otherwise not be solvable at all.
At DFKI Dr Hendeby’s work was centered around on-body sensor networks consisting of inertial measurement units (IMUs) and in some cases other sensors such as cameras. Here he had the opportunity to work with and contribute to the whole signal processing chain from low-level sensor interaction to high-level fusion, and successfully derived methods to estimate the pose of the wearer of the sensors. In doing so we faced many practical problems, such as how to handle sensor synchronization, calibration of a heterogeneous sensor network, etc.
The sensor network was used in two European projects for: monitoring of physical exercise to ensure proper and safe execution (EU AAL project PAMAP (Physical Activity Monitoring for Aging People); and as input to workflow recognition and monitoring to aid in assembly tasks (EU FP7 COGNITO (Cognitive Workflow Capturing and Rendering with On-Body Sensor Networks)). Dr Hendeby was technical coordinator of COGNITO comprising 7 partners from 4 countries with specialties such as biomechanics, computer vision, workflow recovery, computer graphics, human-machine interaction, and hardware manufacturing.
Stripped down wearable sensor network were also used in the PAMAP project to derive the wearer’s everyday daily activities. This is a problem that can be solved with good results using classifiers based on ensemble learners (Boosting).
At the Swedish Defence Research Agency (FOI), Dr Hendeby work with applied research (algorithm development, analysis, and implementation) in the area of security and surveillance. The focus is target tracking, where he overseea the tracking efforts in the group. He coordinates the design and development of an in-house tracking software. In this position he has spent considerable time with questions regarding decentralized and distributed fusion, and tracking systems in production code.
Recently, the focus of his research has been multi-sensor multi-target tracking to provide situational awareness, which is important aspects of both internal as well as external projects, e.g., the EU FP7 projects ADABTS (Automatic detection of abnormal behavior and threats in crowded spaces) and P5 (Privacy preserving perimeter protection project), he has been involved in. For this purpose he has developed a multi-hypothesis tracker (MHT), which is now used as an off-the-shelf toolbox to solve tracking tasks at FOI. In many cases detections in images are used as input, but the developed framework is not limited to this. Another application is tracking in sonar data, for which a Gaussian-Mixture Probability Hypothesis Density (PHD) filter solution was developed.
A topic of recent work, jointly with FOI and Linköping University, is passive tracking in acoustic networks and how the Doppler shifts in received signals can be used to track targets and also direction of arrival estimation. Another area of cooperation is SLAM.
Dr Hendeby is also in the Wild Life Security initiative and its Vinnova financed project “Smarta Savanner” intended to develop solutions to help protect and document endangered spices. In this project he coordinates and has technical lead for many of the different efforts to develop cheap, portable, and efficient methods and technical solutions to protect endangered wildlife. This allows him to work with both rather low-level sensor data as well as with higher level of abstraction in algorithm and demonstrator development.
Autonomous robots require robust perception and navigation in adverse environments where sensors like LiDAR, RADAR, and RGB-D cameras often fail or degrade. We propose MOUSE, a novel adaptive multimodal fusion framework that learns the complementary strengths of each sensor and dynamically selects optimal configurations for changing environments. Unlike existing methods, MOUSE jointly learns multiple downstream perception tasks within a unified framework to enhance situational awareness. Furthermore, MOUSE leverages learned semantic information to enhance measurement associations and object classification, enabling enhanced tracking, localization, and navigation in challenging and dynamic environmental conditions.
State estimation has been highly successful in many applications, such as mobile navigation and situational awareness. Typically, current solutions rely on relatively small-scale systems where all sensor data is available in a central node. However, the growing availability of sensory data introduces the challenge of managing large-scale sensor networks, such as fleets of vehicles, groups of drones, or numerous environmental sensors.
Current state-of-the-art methods struggle with the sheer scale, requiring excessive communication and high computational power in a single node, and are sensitive to node failures and communication issues. Additionally, as the network size increases, the risk of faulty data from failing or malicious nodes rises, and uncertain sensor quality complicates data utilization.
This project aims to develop distributed estimation methods to enhance scalability, robustness, and efficiency in large sensor networks. Key research areas include information compression, detecting misbehaving nodes, and utilizing data with uncertain quality, ultimately improving situational awareness and data utilization for various applications.
Research will focus on methods for multi-target tracking based on audiovisual (AV) data. While sound can support and complement visual object tracking, the combination remains highly underutilized. The probabilistic framework of random finite sets (RFS) offers an appealing approach for inference when both data modes involve a random number of measurements per frame. RFS also provides a flexible foundation with strong data compression, which is crucial for real-time and collaborative settings, enabling decision support while meeting communication requirements. The research will involve developing novel RFS estimators for AV data, leveraging state-of-the-art detectors and classifiers from each domain. Achieving efficient and accurate tracking necessitates investigating fundamentals such as Cramér-Rao lower bounds, optimal sub-pattern assignment, and covariance intersection. These findings will be integrated into estimators like the Poisson multi-Bernoulli mixture for both single and collaborative multi-target tracking. The results will be disseminated and made available to a wider audience, including an open benchmark dataset with associated software.
The goal to provide superior situational awareness (SA) at large scale. To achieve this, the focus of the project is to derive large scale SA using the state-of-the-art Poison multi-Bernoulli mixture (PMBM) filter, first in a centralized setting and then progress to a distributed setting. To succeed, methods from simultaneous localization and mapping (SLAM) and distributed estimation such as covariance intersection (CI) generalizations will be explored, and combined with recent developments of PMBM filtering and data driven modeling.
Modern society relies heavily on GNSS for precise localization and timing. However, recent events have shown that GNSS is susceptible to jamming and spoofing, highlighting the need for alternative solutions. One promising approach is terrain-based localization, using visible landmarks, elevation maps, or Earth’s magnetic field irregularities – similar to techniques used by explorers long before GNSS.
While terrain-aided navigation is not new, it typically requires a high-quality inertial navigation system. This project explores how similar techniques can be used to localize a swarm or group of collaborating agents. By leveraging the geographical spread of the agents, a dynamic sensor array can be created, reducing the stringent requirements on the inertial system. This approach offers new opportunities but relies on well-known relative positions of individual agents, raising questions about the quality of relative positioning within the group.
Methods for distributed learning of augmented state-space models will be researched. This class of models enables domain knowledge to be incorporated in the learning process to guarantee a minimum of perfor- mance and enable an efficient learning process; still they are flexible enough to permit learning of partially unknown model dynamics and inputs. Thus, the targeted methods are foreseen to play an important role to realize large-scale sensing systems. Focus of the research will be on learning of state-space models augmented with sparse Gaussian process models. To enable distributed and efficient learning of the mod- els, techniques for distributed Gaussian process regression, such as Bayesian commit machines and inner product quantization, will be integrated with distributed filter algorithms, such as covariance intersection, consensus, and diffusion Kalman filters. The use of information-gain based active learning strategies to streamline the learning process, both on local and global scale, will be investigated.
Funding from WASP for one PhD student, Sebastian Karlsson.
For more information:
Peng Liu (Co-supervisor)
Gustav Zetterqvist (Co-supervisor)
Chuan Huang (Co-supervisor)
Jakob Åslund (Co-supervisor)
Ashwani Koul (Co-supervisor)
Eric Sevonius (industrial: Saab Dyamics)
Louise Lennartsson (industrial: Saab Aeronautics)
Xiaojing He (Co-supervisor)
For more information:
For more information:
Performance and Implementation Aspects of Nonlinear Filtering
Fundamental Estimation and Detection Limits in Linear Non-Gaussian Systems