Felix Govaers

• PhD
• Deputy Head of Department at Fraunhofer Institute for Communication, Information Processing and Ergonomics

Originally the Accumulated State Density (ASD) has been proposed to provide an exact solution to the out-of-sequence measurement problem. To this end, the posterior of the joint density of all states accumulated over time was derived for a single sensor scenario. An exact solution for T2TF has been published as the Distributed Kalman Filter (DKF)....

A solution to exact Track-to-Track Fusion (T2TF) at arbitrary communication rates has been found under the assumption that all measurement error covariances are known to each of the sensors. The scheme, which is referred to as the “Distributed Kalman Filter” (DKF), produces a fused estimate that is equivalent to a central Kalman filter that process...

Track-to-track fusion (T2TF) aims at combining locally preprocessed information of individual trackers at a fusion center. Particularly, such schemes are obligatory in many applications of distributed sensors because of limited communication resources. If T2TF yields equivalent results compared with a Kalman filter (KF) processing all measurements...

In target tracking applications, the full information on the kinematic target states accumulated over a certain time window up to the present time is contained in the joint probability density function of these state vectors, given the time series of all sensor data. This joint density may also be called an accumulated state density (ASD) and provi...

Globalization contributes to the high expansion of the transport industry. This is particularly evident in air traffic on the smaller airports that lack the costly specialized systems that accommodate nowadays needs of increased travel. One aspect, that needs to be addressed, is a prediction of Time of Arrival and any potential delays and obstacles...

In the research literature there are numerous publications on Track-to-Track fusion (T2TF) and Track-to-Track association (T2TA). These algorithms assume that the problem of target existence is solved on a per sensor basis. More sophisticated detection methods incorporate the full knowledge of a distributed system by making this decision on a globa...

Quantum computing promises significant improvements of computation capabilities in various fields such as machine learning and complex optimization problems. Rapid technological advancements suggest that adiabatic and gate base quantum computing may see practical applications in the near future. In this work, we adopt quantum computing paradigms to...

Random matrices are widely used to estimate the extent of an elliptically contoured object. Usually, it is assumed that the measurements follow a normal distribution, with its standard deviation being proportional to the objects extent. However, the random matrix approach can filter the center of gravity and the covariance matrix of measurements in...

In stochastic estimation problems, we aim to estimate an unknown state of interest given a set of measurements received from noisy sensory devices, such as radar, light detection and ranging, etc. A common model of the measurements’ random error is white Gaussian noise. This noise model is usually used to derive an estimator of the unknown state; a...

Quantum computing promises significant improvements of computation capabilities in various fields such as machine learning and complex optimization problems. Recent technological advancements suggest that the adiabatic quantum computing ansatz may soon see practical applications. In this work, we adopt this computation paradigm to develop a quantum...

This paper introduces BAAS, a new Extended Object Tracking (EOT) and fusion-based label annotation framework for radar detections in the context of autonomous driving. Our framework utilizes Bayesian based tracking, smoothing and eventually fusion methods to provide veritable and precise object trajectories along with shape estimation in order to p...

Many real-world applications of target tracking and state estimation are nonlinear filtering problems and can therefore not be solved byclosed-formanalyticalsolutions.Intherecentpast,tensor-basedapproaches have become increasingly popular due to very effective decomposition algorithms, which allow a compressed representation of discretized, high-di...

Localisation of people that do not carry active tags is needed in security as well as in rescue applications. Ultra-wideband technology is promising due to its high ranging resolution capability, robustness against multipath interference and obstacle penetration among others. In this chapter an approach for detection, localisation and tracking of p...

In this paper we introduce a new method of deciding, if a trajectory is following a pre-defined path. This is achieved by representing hypotheses as trajectories themselves using Accumulated State Densities. Live tracking data is incorporated into the trajectories via out of sequence processing. Through this, we gain two representations of the sens...

The optimal non-linear filter estimates can be obtained by solving the Fokker-Planck equation (FPE) for the time propagation, together with the Bayesian measurement inclusion. An issue faced when solving the FPE is the curse of dimen-sionality. Recently, a tensor based approach has been proposed, which is said to be suitable for high dimensional pr...

We consider the general continuous-discrete nonlin-ear filtering problem. In particular, the prediction step involving the numerical solution of the Fokker-Planck equation for the time evolution of the state probability density is known to be a challenging problem as it suffers from the curse of dimensionality. In this contribution a novel approach...

This paper presents a concept on how a military formation can be jointly protected by linking available single-platform protection systems. To achieve this, a mobile ad hoc network is established between different vehicles carrying heterogeneous sensors, specifically acoustic and ESM sensors, emulating the protection systems. Distributed data fusio...

The task of tracking targets, that generate more than one measurement per scan appears in several applications such as extended object and group tracking. In this case, the target (or group) extent implies that multiple measurements, drawn according to a spatial probability distribution, are measured per sensor-scan. However, applications exist whe...

In the research literature there are numerous publications on Track-to-Track fusion and Track-to-Track association. All these algorithms assume that the problem of target existence is solved on a per sensor basis. In this paper the distributed computation of the track existence by means of the likelihood ratio score is presented. The approach is ba...

For linear-Gaussian non-deterministic dynamics, that is, systems with non-zero process noise, it is well known that tracklet fusion based on equivalent measurement is optimal only for full communication rate, i.e., if the local posterior probabilities or estimates are communicated and fused after each observation and update time. Despite this const...

In tracking and sensor data fusion applications, the full information on kinematic object properties accumulated over a certain discrete time window up to the present time is contained in the conditional joint probability density function of the kinematic state vectors referring to each time step in this window. This density is conditioned by the t...

In multiple target tracking target occlusion or shadowing is a common occurrence. A target may be occluded by an existing structure, or in many cases, by another moving target in the environment. In this paper we consider a UWB-based range-only person tracking system. Occlusion regions induced by moving targets in the scenario are defined followed...

The Ensemble Kalman Filter (EnKF) is a Kalman based particle filter which was introduced to solve large scale data assimilation problems where the state space is of very large dimensionality. It also achieves good results when applied to a target tracking problem, however, due to its Gaussian assumption for the prior density, the performance can be...

This paper focuses on determining the optimum placement of a given number of sensors for estimating the position of a moving target using range-difference measurements. We define a region of interest and generate several random trajectories with the dynamic white noise acceleration model. After obtaining those trajectories that populate the area we...

In this paper a new interpretation of linear estimation in the context of classical mechanics is presented. In this context, Accumulated State Densities can be interpreted as the Lagrange function of a "least action" principle that provides the expectation vectors for filtering and retrodiction as a solution. The superposition principle, which stat...

When localizing multiple tag-free targets using ultra-wideband sensors, targets closer to the sensor occlude targets further from the sensor and in turn these targets are not detected. In this paper we first model the occlusion region of each target. Based on this model, targets unresolved in sensor data and occluded by other targets are updated. E...

The task of tracking targets, that generate more than one measurement per scan appears in several applications such as extended object and group tracking. There, the target (or group) extend implies that muliple measurements, drawn according to a spatial probability distribution, are measured per sensor-scan. However, applications exist where targe...

In this paper, a novel approach for distributed bearings-only tracking is presented. In the past years, the literature on tracking has focused more and more on the distributed Kalman filter, which yields the optimal state estimate, given that the sensor model of all sensors in the system is known to each local processor. Since this condition is har...

In track fusion, the measurements of individual sensors for each target are processed to generate local state estimates, which are then fused to obtain the global state estimate for the target. When there is no process noise or the fusion rate equals the sensor observation rate, the standard tracklet fusion or equivalent measurement fusion algorith...

The passive non-cooperative localization and tracking of mobile terminals in urban scenarios, called blind mobile localization (BML), is a highly demanding task which occurs for instance in safety, emergency and security applications with non-subscribed phone user locations. Due to the urban environment and physical propagation effects multiple sig...

For the purpose of tracking maneuverable targets and estimating the maneuver mode, a multiple model filter similar to the interacting multiple model (IMM) filter is used. The dependency of the mode transition probabilities on the state is taken into account. By the use of the information about the mode that is contained in the target state, a faste...

The distributed Kalman filter requires the measurement covariances of remote radar nodes to be known at all radar nodes. This is not possible for a radar network, as the true measurement covariances depend on the radar-target geometry and the fluctuating signal-to-noise ratio. This paper tackles this problem using the double debiased distributed Ka...

In this paper, a direct connection between the covariance debiasing methodology for the distributed Kalman (DKF) filter in [1] and the federated Kalman filter is shown. In particular, it can be seen that for a unique choice of the information gain hypothesis of the DKF, the covariance debiasing becomes equivalent to the federated Kalman filter. As...

In this paper we describe a method for localization of multiple persons using a distributed network of autonomous ultra-wideband sensor nodes. The persons do not carry any devices or tags to aid their detection, but are instead detected by using the time variations they impose on the measured channel impulse response between a transmitter and a rec...

A multiple model filter similar to the IMM filter is developed for tracking of maneuvering targets. The mode transition probabilities are modeled as dependent on the state. This allows using information about the mode of a target that is contained in the state. Thus, better estimates of the mode can be obtained. Convergence of the mode estimates oc...

A Gaussian Mixture (GM) target tracking solution is a natural consequence of the multi-target tracking in clutter, given a linear target trajectory propagation and a linear target measurement equation. We examine and compare two prominent GM target trackers: the Multi Hypothesis Tracking (MHT) and the Integrated Track Splitting. Both incorporate th...

In this paper, the target existence probability for a single target in clutter is derived. More specifically, the paper considers target existence in the distributed Kalman filter. First, a conceptual solution is derived explicitly for a two-sensor case, and second a moment-matching approximation is performed, which enables computational tractabili...

The passive non-cooperative localization and tracking of mobile terminals in urban scenarios, called blind mobile localization (BML), is a highly demanding task which occurs for instance in safety, emergency and security applications with non-subscribed phone user locations. In this paper it is demonstrated how Sequential Monte Carlo (SMC) implemen...

In realistic scenarios, distributed tracking is hindered due to limitations on the communication capacity between sensor nodes. In this paper, solutions for measurement fusion and track-to-track fusion in communication constrained distributed sensor networks are summarized. A comparison of the communication required and the target state estimate er...

In this paper, various techniques for information fusion in distributed sensor applications are presented. In the considered scenarios a number of challenges exist due to limitations on the communication between sensor nodes. Firstly, the challenge of delayed data processing is addressed in order to present solutions for optimal state estimation wh...

Radar networks require techniques for information fusion, such as track-to-track fusion. However, existing fusion schemes in the literature seldom apply radar-like models, which can have profound effects on performance. Hence, this paper describes and analyses state of the art algorithms for track-to-track fusion assuming range-Doppler radar measur...

The so-called lack of memory is an inherent challenge of the probability hypothesis density (PHD) filter and leads to the fact that only targets which rely on a currently available measurement can securely be reported as present in the respective iteration. Yet there is no method presented that enables the sequential Monte Carlo (SMC) version of th...

In the recent past, a solution to exact T2TF at arbitrary communication rates was found under the assumption of perfect detection and data assignment. This paper extends this approach by considering the practically relevant problem of measurement origin uncertainty and non-detections. As the local decorrelation of tracks requires the global covaria...

Augmented Reality focuses on the enrichment of the user's natural field of view by consistent integration of text, symbols and interactive three-dimensional objects in real time. Placing virtual objects directly into the user's view in a natural context empowers highly dynamic applications. On the other hand, this necessitates deliberate choice of...

Botnets are and are likely to remain the main vehicle for online crime for the foreseeable future. To protect their business models, botnet operators constantly improve their protocols and applications to harden them against detection, analysis and takedown efforts. Our analysis suggest that future botnets will use proper encryption for their proto...

The increasing trend towards multi-sensor systems is driving a requirement for distributed tracking algorithms. In such systems, communication links often suffer from varying delays, which leads to timely disordered data at the fusion center. The challenge of online processing delayed data is generally known as the Out-of-Sequence (OoS) problem. An...

In this article, we propose a new extension to a Dynamic Programming Algorithm (DPA) approach for Track-before-Detect challenges. This extension enables the DPA to process time-delayed sensor data directly. Such delay might appear because of delays in communication networks. The extended DPA is identical to the recursive standard DPA in case of all...

This paper presents a method for indoor pedestrian tracking using inertial sensing and a laser scanner (Light Detection and Ranging LIDAR). The zero velocity updating technique(2), which is used to enhance the performances of an inertial sensing sensor mounted on the foot, cannot observe heading, resulting in an horizontal position drift. A Lidar m...

The development of modern tracking system in the recent past is towards distributed sensor scenarios. As communication links often suffer from varying delays, timely disordered data appears. An exact solution to the general Out-of-Sequence problem is given by the Accumulated State Density (ASD) filter. In this paper, we derive the Information filte...

Tracking applications in distributed sensor scenarios often suffer from small bandwidths and time delayed transmis- sions. In order to save communication capacity, it is common to transmit locally preprocessed estimates on the target to a central fusion center. A global track is obtained by Track-to-Track fusion. If sensor links are only establishe...

This paper presents a novel method for indoor pedestrian tracking using inertial sensing and sonar sensors. The zero velocity updating technique, which is used to enhance the performances of inertial sensing, cannot observe heading, resulting in a horizontal position drift. Sonar sensors are used as complementary technique to correct heading. The m...

In target tracking applications, the full information on the kinematic target states accumulated over a certain time window up to the present time is contained in the joint probability density function of these state vectors, given the time series of all sensor data. In the structure of this Accumulated State Density (ASD) has been revealed. Furthe...

Track-to-track fusion aims at combining locally preprocessed information of individual sensors optimally, i.e. in a way that is equivalent to fusing all measurements of all sensors directly. It is well known that this can be achieved if the local sensor tracks produced at all individual scan times are available in the fusion center. Full-rate commu...

Command and control applications are important especially in tactical scenarios. For such scenarios wireless multi-hop networks may be deployed as they can be used even when there is no infrastructure left. Due to the specific characteristics of these networks the problem of Out-of-Sequence (OoS) measurements for tracking applications arises. In th...

In many GPS-Sensor based tracking applications, the obtained measurements suffer from time a correlated bias. This is due to shadowing and multipath scattering of the wireless GPS signals. In this paper, we applicate a Schmidt-Kalman Filter (SKF) in order to improve the tracking process of ground vehicles on roads. We investigate possibilities to i...

Nowadays, the necessity of safeguarded environments is stronger than ever. The defence of public areas against terroristic threats requires intelligent security assistance systems that comprise state-of-the-art surveillance technology to localize persons with hazardous materials. The recent progress in the detection of hazardous materials by a new...