In this paper, we propose to fuse radar measurements with Visual Inertial Odometry (RVIO) or Thermal Inertial Odometry (RTIO). FMCW radar sensor data enables to estimate the 3D ego velocity independent of the visual conditions. Fusion with VIO or TIO heavily improves the robustness in challenging conditions such as darkness, direct sunlight or fog.Specifically, we propose RRxIO: An extension to the state of the art filter based VIO framework Robust Visual Inertial Odometry (ROVIO) to fuse radar sensor data. Due to the drift free 3D radar ego velocity estimates, scale errors are reduced and only a small number of features needs to be tracked to achieve good results. This yields faster runtimes while outperforming state of the art VIO frameworks regarding accuracy. RVIO is able to bridge phases of degraded or even no visual features resulting in a low cost system even for poor visual conditions. RTIO is robust even in environments with small temperature gradients and bridges phases of no thermal images caused by Non-Uniformity Corrections (NUCs).We evaluated our system with various experiments in different environments and visual conditions. Comparison to other state of the art dual domain approaches including radar-inertial, visual-inertial and thermal-inertial proves superior performance of our approach regarding accuracy and processing time.

In this paper we propose an iterative Graph SLAM based approach to create HD maps from series production vehicles fleet data. Only high level sensor measurements provided by advanced driver assistance systems are used. This reduces the required bandwidth and makes this approach scalable to vehicle fleet data. Creating HD maps from fleet data enables up to date HD maps since no dedicated mapping vehicles are required. At first, the data is aligned based on odometry, GNSS and traffic sign measurements. Next, road boundary measurements are included, which results in an optimized lateral alignment. Finally, lane boundary association can be carried out. This results in an HD map containing high accuracy data of traffic signs, road and lane boundaries. The approach was evaluated using series production vehicle fleet data recorded on US highways and on a German autobahn covering a distance of 35.4km. The final HD map was evaluated with a groundtruth HD map and achieved an average error of 0.59m.

Micro Aerial Vehicles (MAVs) lend themselves to perform supporting flights in many cases from consumer-oriented service to Urban Search and Rescue (USAR) scenarios. However, most of today's approaches are restricted to piloted outdoor applications for lack of autonomy in unknown scenarios. In this paper, we present a comprehensive mapping system tailored to the specific needs of independent indoor reconnaissance flights. We provide for both metric and floor plan mapping and exploit the benefits of online extracted semantics in two ways. First, grid-mapping performance directly benefits from semantic feedback loop. Second, the MAV uses a multi-layered topological map established on the fly for autonomous reconnaissance in unknown environment. A new exploration strategy provides for complete coverage in consideration of desired prioritization. We evaluate the proposed enhancements in robust mapping and floor plan generation for MAVs by means of real experimental data acquired in demanding indoor environment. The flight guidance is validated in simulations.

This paper presents a new finite state machine based elevator and escalator classification for pedestrians with foot mounted MEMS sensors. Generally, stance phase detectors are used to reduce the position error drift by applying Zero-Velocity-Updates (ZUPTs) in the navigation filter. However, if the person rides an elevator or escalator ZUPTs should not be processed during the stance phases of the foot. A novel elevator and escalator classifier separating the motion of the moving platform in different sub-states is introduced. Unique is the consideration of foot movement, so that the person can walk during the ride with an elevator or escalator. In addition, powerful motion constraints differing for the various platforms are presented which enable accurate positioning performance. Horizontal ZUPTs and delta ZUPTs are introduced for elevator rides and absolute velocity measurements based on an escalator velocity model are used to limit the position error growth during escalator movement. We evaluated the presented algorithms with recorded sensor data and achieved an average return position error of 0.72% in the vertical axis and an average error of 0.58% in the horizontal plane of the total walked distance of 1162.09m.

Unmanned Air Vehicles (UAVs) lend themselves to provide rescue forces with important information beyond their own field of vision. In this paper, we present a two-part yet synergetic navigation and mapping system for indoor reconnaissance and surveillance through aerial platforms. It faces inherent drawback of common Simultaneous Localization and Mapping (SLAM) approaches: map errors cannot be excluded and will directly afflict the positioning. With a flexible navigation filter and an efficient 2.5D mapping tool as centerpieces, our system facilitates appropriate mutual feedback of supporting information. That is map-based navigation aiding in terms of ego-motion estimation and improved mapping in terms of accuracy, robustness and efficiency. Hereby both modules remain independent in their basic function and are therefore guarded against overlapping errors. Finally, the system is validated by means of real experimental data and a first proof of suitability for application within autonomous exploration is given based on simulated flights.

In this paper we present a novel finite state machine based step detection technique for precise personal navigation solutions with a foot-mounted inertial measurement unit (IMU). Generally, step detection methods are used to improve the navigation solution by applying Zero Velocity Updates (ZUPTs) in the navigation filter. All step detection techniques distort the navigation solution if ZUPTs are utilized at wrong times. Our approach based on a finite state machine is able to detect different stances of the foot with high accuracy. Therefore, Zero Velocity Updates can be applied in time and positively affect the precision of the navigation solution. The functionality of the step detection module in combination with a constraint, stochastic cloning (SC) Kalman filter are analyzed with real sensor data recorded with our pedestrian navigation system. Even with ultra-low cost inertial sensors, this new approach can clearly increase the accuracy of pedestrian navigation systems compared to state-of-the-art approaches.

Indoor navigation using inertial sensors with additional radio-signal support is considered in this paper. The experimental results of data fusion between a navigation system, based on inertial measurement unit (IMU) and impulse-based UWB localization system, are presented. The IMU is additionally supported by a pedestrian step length estimations, barometer and electronic compass. The Ultra-Wideband part consists of receiver, carried by the person, and access points distributed in the scenario. The focus of the paper is put on hardware implementation and choice of the optimal data fusion technique. The presented results indicate the clear benefit of tightly coupled navigation filter, where the time differences of arrival of UWB signals are directly processed, without prior calculation of the localization solution.

In this paper a map matching algorithm for multi floor indoor environments including guidance of a user is presented. Inertial sensor based pedestrian navigation systems are subject to drift. Maps are to eliminate that drift by matching the path to a map. In multi floor scenarios, we propose to use not only flat floor plans but also transitions like staircases as well as ladders and elevators. This is essential for an enhanced matching of a user path to a given 3D floor plan. Especially in industrial applications this is a crucial factor as also ladders and elevators exist. In contrast to other authors, this paper considers all of these objects. 3D position estimations from an Integrated Pedestrian Navigation System (IPNS) including an inertial system as well as a barometer and a magnetometer are used. Maps must be available, either from a prior Simultaneous Localization and Mapping (SLAM) survey or from facility maps. The new map representation is used to match real sensor data from the IPNS system to a map in a real time implementation. Furthermore an online guidance implementation is presented which is also based on the new 3D map representation.

Experimental results from data fusion of an inertial sensors based navigation system aided by a pedestrian step length estimations and an ultra-wideband system (UWB) are presented in this paper. One part of the presented experimental setup is a torso mounted inertial system with IMU, barometer and electronic compass with additional step length updates. The other part is the UWB system: The receiver is mounted on a backpack and measures the signal time of flight differences to several indoor base stations. In the paper the hardware setup, UWB implementation, step length updates as well as the data fusion filter are discussed. Finally results are presented, that show the advantage of a tightly coupled navigation filter, where range measurements are processed directly in the navigation filter.

This paper describes non-coherent Deeply Coupled GPS/INS integration in a pedestrian navigation system to improve position accuracy and availability in weak signal conditions. A pedestrian navigation system consists of several sensors to calculate a position of a person to guide for example rescue missions. The system presented in this paper consists of a torso mounted IMU and is used for step detection and step length and heading estimation. Additionally a barometer, magnetometer and a GPS sensor for absolute positioning are used. Since pedestrian navigation systems often are used in challenging environments like urban canyons or indoors, the use of GPS signals is often restricted. We will show that by using a Deeply Coupled GPS/INS integration system, tracking of GPS signals under weak signal conditions is possible and a seamless transition between Indoor and outdoor situations is achieved. By applying the information of a position displacement between two steps from the step length and heading estimation GPS tracking and position accuracy can be increased. For an optimal performance the system uses a deeply acquisition and re-acquisition routine. Therefore additional satellites can be used which could not have been acquired before, due to low signal to noise ratios. By carefully weighting the GPS measurements accordingly to their C / No and having a larger set of satellites available, position accuracy is increased compared to a non-vector tracking approach. The sensor fusion itself is realized in an error state space kalman filter and the step length update is performed using a state cloning technique preserving realistic position uncertainties in the filter. With this approach tracking and acquisition of GPS signals inside buildings with C / No below 20dBHz is possible. In this paper we will show that using deep integration in GPS signal tracking including step length estimations increases position accuracy of a pedestrian navigation system and availability of GPS position updates.

In this paper a tightly coupled UWB/INS system for pedestrian indoor applications is presented. In indoor environments wireless signal outage and severe multipath propagation very often lead to ranging errors and make pure UWB-based localization barely possible. On the other hand inertial sensor based systems (INS) are known for their drift with time. The integration of those two systems will allow to profit from their advantages. For the sensor data fusion the tightly coupled approach and integrity monitoring are crucial factors for a robust implementation. One possibility is the Time of Arrival (ToA) approach where the user clock bias and drift are modeled and estimated in the navigation filter. The other possibility is the Time Difference of Arrival (TDoA) approach, where the range measurements are differenced totally eliminating the clock error. If user/transmitter clocks do not follow the clock drift model very well, it is better to apply the TDoA approach. For time of arrival applications, innovation based integrity monitoring is standard. However for the use of Time Difference of Arrival measurements (TDoA) in this paper an Innovation Based Integrity Monitoring (IBIM) method is presented. This is done to determine and omit all TDoA measurement combinations including the incorrect UWB receivers. This novel approach is verified in a simulation environment. A user trajectory and inertial data are provided by a custom developed pedestrian walk generator, based on real measurement data. The UWB data is generated by a wave propagation simulator for the given indoor trajectory.

In this paper an Indoor Navigation System with map-matching capabilities in real-time on a smart phone is presented. The basis of the system is an in-house development of an Integrated Pedestrian Navigation System, based on 2 low-cost IMUs, an electronic compass and an altimeter with a drifting navigation solution. Combining this system with an additional laser ranger and SLAM algorithms, we are able to build accurate maps of office buildings for already visited rooms in post processing. This paper presents a map matching algorithm based on a new reduced particle filter in order to use these maps later for real-time applications without an expensive laser ranger but relying only on the dual inertial system. It can be used with both, preprocessed SLAM maps or with already available maps. Finally to smooth the resulting trajectory after particle filtering we propose the use of a new “balanced bubble band smoother” allowing the trajectory to optimally match to both, map and recorded IMU data. This new approach makes it possible to do map matching online on a smart phone.

This contribution considers the accuracy limits of ultra-wideband localization systems dedicated to large-scale applications. Industrial environments are known for being extremely difficult for wireless communication, mainly because of the high concentration of large metal objects, such as containers or machines; leading to strong multipath propagation. In this work a detailed 3D model of the warehouse is used for wave propagation simulations. In the simulation, the transmitter is moving along the predefined path and the receiver infrastructure consists of eight base stations. The aim is to model the tracking system and asses its accuracy - first, the upper bound of the accuracy is determined under ideal conditions and second, the precision estimation with the inclusion of effects of hardware non-idealities is done. Effects such as the influence of pulse detection methods, pulse shape and geometrical configuration of base-stations are investigated. The position calculation of the mobile beacon is realized using a Time Difference of Arrival approach. Eventually the accuracy of the obtained results and the theoretical limit are considered based on the dilution of precision evaluations.

Robots are emerging from industrial plants toward every people's daily life. Thus, navigation in and understanding of human related environments becomes a prerequisite for the systems of tomorrow. Most such environments can be efficiently described using line segments. However, incorporation of extent information is often difficult, as line segments are seldom observed completely and erroneous data-association may corrupt the information associated to a certain segment. To reduce such problems this paper proposes a statistically driven description of line segments. The corresponding parameters are decomposed into line and corner features, which are separately tracked through an extended Kalman filter (EKF). Information about the extent of the lines is encoded statistically. Therefore, we use a method to recursively incorporate the information gained through a time-series of measurements. Thus, the covariance matrix belonging to a line segment grows as new regions of the corresponding line are discovered. Experimental results obtained by implementation on the mobile platform ITrike show the validity of our algorithm.