Department of Electrical Engineering (ISY)

This article considers a sensor management problem where a number of road bounded vehicles are monitored by an unmanned aerial vehicle (UAV) with a gimballed vision sensor. The problem is to keep track of all discovered targets and simultaneously search for new targets by controlling the pointing direction of the vision sensor and the motion of the UAV. A planner based on a state-machine is proposed with three different modes; target tracking, known target search, and new target search. A high-level decision maker chooses among these sub-tasks to obtain an overall situational awareness. A utility measure for evaluating the combined search and target tracking performance is also proposed. By using this measure it is possible to evaluate and compare the rewards of updating known targets versus searching for new targets in the same framework. The targets are assumed to be road bounded and the road network information is used both to improve the tracking and sensor management performance. The tracking and search are based on flexible target density representations provided by particle mixtures and deterministic grids.

A vessel navigating in a critical environment such as an archipelago, requires very accurate movement estimates. Intentional or unintentional jamming makes gps unreliable as the only source of information and an additional independent navigation system should be used. In this paper we suggest estimating the vessel movements using a sequence of radar images from the preexisting body-fixed radar. Island landmarks in the radar scans are tracked between multiple scans using visual features. This provides information not only about the position of the vessel but also of its course and velocity. We present here a complete navigation framework that requires no additional hardware than the already existing naval radar sensor. Experiments show that visual radar features can be used to accurately estimate the vessel trajectory over an extensive data set.

This work presents a particle filter (PF) method closely related to FastSLAM for solving the simultaneous localization and mapping (SLAM) problem. Using the standard FastSLAM algorithm, only low-dimensional vehicle models can be handled due to computational constraints. In this work an extra factorization of the problem is introduced that makes high-dimensional vehicle models computationally feasible. Results using experimental data from a UAV (helicopter) are presented. The algorithm fuses measurements from on-board inertial sensors (accelerometer and gyro), barometer, and vision in order to solve the SLAM problem.

Many RGB-D sensors, e.g. the Microsoft Kinect, use rolling shutter cameras. Such cameras produce geometrically distorted images when the sensor is moving. To mitigate these rolling shutter distortions we propose a method that uses an attached gyroscope to rectify the depth scans. We also present a simple scheme to calibrate the relative pose and time synchronization between the gyro and a rolling shutter RGB-D sensor. We examine the effectiveness of our rectification scheme by coupling it with the the Kinect Fusion algorithm. By comparing Kinect Fusion models obtained from raw sensor scans and from rectified scans, we demonstrate improvement for three classes of sensor motion: panning motions causes slant distortions, and tilt motions cause vertically elongated or compressed objects. For wobble we also observe a loss of detail, compared to the reconstruction using rectified depth scans. As our method relies on gyroscope readings, the amount of computations required is negligible compared to the cost of running Kinect Fusion.

We present a system that rectifies and stabilizes video sequences on mobile devices with rolling-shutter cameras. The system corrects for rolling-shutter distortions using measurements from accelerometer and gyroscope sensors, and a 3D rotational distortion model. In order to obtain a stabilized video, and at the same time keep most content in view, we propose an adaptive low-pass filter algorithm to obtain the output camera trajectory. The accuracy of the orientation estimates has been evaluated experimentally using ground truth data from a motion capture system. We have conducted a user study, where the output from our system, implemented in iOS, has been compared to that of three other applications, as well as to the uncorrected video. The study shows that users prefer our sensor-based system.

We consider stand still detection for indoor localization based on observations from a foot-mounted inertial measurement unit (IMU). The main contribution is a statistical framework for stand-still detection, which is a fundamental step in zero velocity update (ZUPT) to reduce the drift from cubic to linear in time. First, the observations are transformed to a test statistic having non-central chi-square distribution during zero velocity. Second, a hidden Markovmodel is used to describe the mode switching between stand still, walking, running, crawling and other possible movements. The resulting algorithm computes the probability of being in each mode, and it is easily extendable to a dynamic navigation framework where mapinformation can be included. Results of first mode probability estimation, second map matching without ZUPTand third step length estimation with ZUPT are provided.

Using prior information about the road network will improve the estimation performance for a road constrained target significantly. Several estimation methods have been proposed to handle the multi-modality that arise in a road target tracking application. One popular filter suitable for this kind of non-linear problems is the Particle Filter, but a major drawback is that the Particle filter requires a large amount of particles as the state dimension increases to maintain a good approximation of the filtering distribution. In this paper a Rao-Blackwellized Particle Filter based approach is proposed to reduce the dimension of the state space in road target tracking applications. Furthermore, it is also shown how prior information about the probability of detection can be used to improve the estimation performance further.

This contribution aims at unifying two recent trends in applied particle filtering (PF). The first trend is the major impact in simultaneous localization and mapping (SLAM) applications, utilizing the FastSLAM algorithm. Thesecond one is the implications of the marginalized particle filter (MPF) or the Rao-Blackwellized particle filter (RBPF) in positioning and tracking applications. Using the standard FastSLAM algorithm, only low-dimensional vehicle modelsare computationally feasible. In this work, an algorithm is introduced which merges FastSLAM and MPF, and the result is an algorithm for SLAM applications, where state vectors of higher dimensions can be used. Results using experimental data from a UAV (helicopter) are presented. The algorithmfuses measurements from on-board inertial sensors (accelerometer and gyro) and vision in order to solve the SLAM problem, i.e., enable navigation over a long period of time.

Detection of spurious features is instrumental in many computer vision applications. The standard approach is feature based, where extracted features are matched between the image frames. This approach requires only vision, but is computer intensive and not yet suitable for real-time applications. We propose an alternative based on algorithms from the statistical fault detection literature. It is based on image data and an inertial measurement unit (IMU). The principle of analytical redundancy is applied to batches of measurements from a sliding time window. The resulting algorithm is fast and scalable, and requires only feature positions as inputs from the computer vision system. It is also pointed out that the algorithm can be extended to also detect nonstationary features (moving targets for instance). The algorithm is applied to real data from an unmanned aerial vehicle in a navigation application.

Computational complexity is a major concern for practical use of the versatile particle filter (PF) for nonlinear filtering applications. Previous work to mitigate the inherent complexity includes the marginalized particle filter (MPF), with the fastSLAM algorithm as one important case. MPF utilizes a linear Gaussian sub-structure in the problem, where the Kalman filter (KF) can be applied. While this reduces the state dimension in the PF, the present work aims at reducing the sampling rate of the PF. The algorithm is derived for a class of models with linear Gaussian dynamic model and two multirate sensors, with different sampling rates, one slow with a nonlinear and/or non-Gaussian measurement relation and one fast with a linear Gaussian measurement relation. For this case, the KF is used to process the information from the fast sensor and the information from the slow sensor is processed using the PF. The problem formulation covers the important special case of fast dynamics and one slow sensor, which appears in many navigation and tracking problems.

Fault detection based on comparing a batch of data with a model of the system using the generalized likelihood ratio test is considered. Careful treatment of the initial state of the model is quite important, in particular for short batch sizes. There are two standard approaches to this problem. One is based on a parity space, where the influence ofinitial state is removed by projection, and the other on using prior information obtained by Kalman filtering past data. A new idea of anti-causal Kalman filtering in the present data batch is introduced and compared to the previous methods. An efficient parameterization of incipient faults is given. It is shown in simulations of torque disturbances on a DCmotor that efficient fault profile parameterization and using smoothed estimates of the initial state increase performance considerably. (Copyright 2006 IFAC)

The Generalized Likelihood Ratio (GLR) test for fault detection as derived by Willsky and Jones is a recursive method to detect additive changes in linear systems in a Kalman filter framework. Here, we evaluate the GLR test on a sliding window and compare it to stochastic parity space approaches. Robust fault detection defined as being insensitive to faults in the signal space is also studied in the GLR framework.

Centralized control and coordination of the connections in a wireless network is not possible in practice. To keep the delay from measurement instants to actuating the decisions, distributed control is required. This paper focuses on the uplink (from mobiles to base stations) and discusses distributing the decision of when and when not to transmit data (distributed scheduling) to the mobiles. The scheme, uplink transmission timing, utilizes mobile transmitter power control feedback from the base station receiver to determine whether the channel is favorable or not compared to the average channel condition. Thereby, the battery consumption and disturbing power to other connections are reduced. The algorithm can be described as a feedback control system. Some transient behaviors are analyzed using systems theory, and supported by wireless network simulations of a system with a WCDMA (Wideband Code Division Multiple Access) radio interface as in most 3G systems.

The ability to navigate in an unknown environment is an enabler for truly utonomous systems. Such a system must be aware of its relative position to the surroundings using sensor measurements. It is instrumental that these measurements are monitored for disturbances and faults. Having correct measurements, the challenging problem for a robot is to estimate its own position and simultaneously build a map of the environment. This problem is referred to as the Simultaneous Localization and Mapping (SLAM) problem. This thesis studies several topics related to SLAM, on-board sensor processing, exploration and disturbance detection.

The particle filter (PF) solution to the SLAM problem is commonly referred to as FastSLAM and has been used extensively for ground robot applications. Having more complex vehicle models using for example flying robots extends the state dimension of the vehicle model and makes the existing solution computationally infeasible. The factorization of the problem made in this thesis allows for a computationally tractable solution.

Disturbance detection for magnetometers and detection of spurious features in image sensors must be done before these sensor measurements can be used for estimation. Disturbance detection based on comparing a batch of data with a model of the system using the generalized likelihood ratio test is considered. There are two approaches to this problem. One is based on the traditional parity space method, where the influence of the initial state is removed by projection, and the other on combining prior information with data in the batch. An efficient parameterization of incipient faults is given which is shown to improve the results considerably.

Another common situation in robotics is to have different sampling rates of the sensors. More complex sensors such as cameras often have slower update rate than accelerometers and gyroscopes. An algorithm for this situation is derived for a class of models with linear Gaussian dynamic model and sensors with different sampling rates, one slow with a nonlinear and/or non-Gaussian measurement relation and one fast with a linear Gaussian measurement relation. For this case, the Kalman filter is used to process the information from the fast sensor and the information from the slow sensor is processed using the PF. The problem formulation covers the important special case of fast dynamics and one slow sensor, which appears in many navigation and tracking problems.

Vision based target tracking is another important estimation problem in robotics. Distributed exploration with multi-aircraft flight experiments has demonstrated localization of a stationary target with estimate covariance on the order of meters. Grid-based estimation as well as the PF have been examined.

Indoor positioning in unknown environments is crucial for rescue personnel and future infotainment systems. Dead-reckoning inertial sensor data gives accurate estimate of distance, for instance using zero velocity updates, while the heading estimation problem is inherently more difficult due to the large degree of magnetic disturbances indoors. We propose a Kalman filter bank approach based on supporting a magnetic compass with gyroscope turn rate information, where a hidden Markov model is used to model the presence of magnetic disturbances. In parallel, we suggest to run a robust heading estimation system based on data from a sliding window. The robust estimate is used to detect filter divergence, and to restart the filter when needed. The underlying assumptions and the heading estimation performance are supported in field trials using more than 500 data sets from more than 50 venues in 5 continents.

This work considers a sensor management problem where a number of road bounded vehicles are monitored by a UAV with a gimballed vision sensor. The problem is to keep track of all discovered targets and simultaneously search for new targets by controlling the pointing direction of the vision sensor and the motion of the UAV. A planner based on a state-machine is proposed with three different modes; target tracking, known target search, and new target search. A high-level decision maker chooses among these sub-tasks to obtain an overall situational awareness. A utility measure for evaluating the combined search and target tracking performance is also proposed. By using this measure it is possible to evaluate and compare the rewards of updating known targets versus searching for new targets in the same framework. The targets are assumed to be road bounded and the road network information is used both to improve the tracking and sensor management performance. The tracking and search are based on flexible target density representations provided by particle mixtures and deterministic grids.

This work presents a particle filter (pf) method closely related to Fastslam for solving the simultaneous localization and mapping (SLAM) problem. Using the standard Fastslam algorithm, only low-dimensional vehicle models can be handled due to computational constraints. In this work an extra factorization of the problem is introduced that makes high-dimensional vehicle models computationally feasible. Results using experimental data from a UAV (helicopter) are presented. The proposed algorithm fuses measurements from on-board inertial sensors (accelerometer and gyro), barometer, and vision in order to solve the SLAM problem.

n/a

This contribution aims at unifying two recent trends in applied particle filtering (PF). The first trend is the major impact in simultaneous localization and mapping(SLAM) applications, utilizing the FastSLAM algorithm. Thesecond one is the implications of the marginalized particlefilter (MPF) or the Rao-Blackwellized particle filter (RBPF)in positioning and tracking applications. Using the standardFastSLAM algorithm, only low-dimensional vehicle modelsare computationally feasible. In this work, an algorithm isintroduced which merges FastSLAM and MPF, and the resultis an algorithm for SLAM applications, where state vectorsof higher dimensions can be used. Results using experimentaldata from a UAV (helicopter) are presented. The algorithmfuses measurements from on-board inertial sensors (accelerometerand gyro) and vision in order to solve the SLAM problem, i.e., enable navigation over a long period of time.

This contribution aims at unifying two recent trends in applied particle filtering (PF). The first trend is the major impact in simultaneous localization and mapping (SLAM) applications, utilizing the FastSLAM algorithm. The second one is the implications of the marginalized particle filter (MPF) or the Rao-Blackwellized particle filter (RBPF) in positioning and tracking applications. An algorithm is introduced, which merges FastSLAM and MPF, and the result is an MPF algorithm for SLAM applications, where state vectors of higher dimensions can be used. Results using experimental data from a 3D SLAM development environment, fusing measurements from inertial sensors (accelerometer and gyro) and vision are presented.

n/a

The Generalized Likelihood Ratio (GLR) test for fault detection asderived by Willsky and Jones is a recursive method to detectadditive changes in linear systems in a Kalman filter framework.Here, we evaluate the GLR test on a sliding window and compare it tostochastic parity space approaches. Robust fault detection definedas being insensitive to faults in the signal space is also studiedin the GLR framework.

n/a