Thomas Topp

Zusammenfassung

INS toolkit for integrated navigation concepts and training (INSTINCT) is an open-source positioning, navigation and timing (PNT) framework for global navigation satellite system (GNSS) navigation and sensor fusion written in C++. It uses flow-based programming to encapsulate functionality, enforce clean interfaces and promote reusability. Not only multi-constellation, multi-frequency single point positioning (SPP) and real-time kinematic positioning (RTK) algorithms are available, but also inertial navigation system (INS)/GNSS sensor fusion. Moreover, innovative concepts like multi inertial measurement unit (IMU) arrays and factor graph optimization are featured. Furthermore, most file formats common in the PNT field can be read with the software and converted between them. Also, simulation of trajectories and IMU data with different error models is possible. A graphical user interface allows the user to directly set parameters and analyze results in plots, which enables rapid prototyping and testing. A developer can easily extend the functionality with own algorithms and sensor interfaces building upon the existing modules. In order to evaluate the performance of the algorithms two experiments were performed. Analysis of a static dataset shows that the position accuracy of the RTK algorithm of INSTINCT is comparable to RTKLIB. Additionally, a dynamic dataset was generated using a Spirent GNSS simulator and INSTINCT’s IMU simulation capabilities. In-depth assessment confirms the high accuracy of the results and demonstrates that the INS/GNSS loosely coupled Kalman filter can compensate for GNSS outages.

Zusammenfassung

Constrained by tight cost calculations and relatively small payload capabilities, we propose an integrated navigation system for urban air mobility flight operation using low-cost hardware. Our concept features a WiFi-based Positioning System (WPS), which complements GNSS, a barometer and an Inertial Navigation System (INS), which makes use of multiple Inertial Measurement Units (IMUs). These four sensor systems are fused using a loosely coupled Kalman filter. Compared to basic GNSS/INS-fusion, our approach has an improved robustness and accuracy. The WPS consists of an off-the-shelf WiFi router mounted to an Unmanned Aerial Vehicle (UAV) and a set of stationary WiFi routers. If the positions of the stationary routers are known, Round Trip Times (RTTs) to the UAV mounted router can be used to calculate the position of the UAV (see figure 1). Flight tests have shown that the WPS’s position accuracy is comparable to GNSS code-phase positioning. The WiFi routers are also used as a bidirectional data link, providing the UAV with measured RTTs and telemetry and enabling down-streaming of the navigation solution for monitoring purposes. The INS makes use of a multi-IMU array, in which specific forces and angular rates of an arbitrary number of Single IMUs (SIMUs) are fused consistently in a B-spline Kalman filter. The B-spline fusion model acts as a low-pass filter which allows to separate actual dynamic motion from sensor noise. In general, this concept provides redundancy and has a better precision than a SIMU (see figure 2). The proposed integrated navigation system is implemented in the open-source software framework INSTINCT (Institute of Navigation, Stuttgart, Toolkit for Integrated Navigation Concepts and Training). The potential of this approach is validated by simulations and flight tests and the results will be summarized in our presentation.

Zusammenfassung

In order to process the data with INSTINCT a certain file structure is needed : - The INSTINCT executable should be placed at the top level, - Files with extension .flow need to be placed inside a folder called flow, - All data files need to be placed inside a folder called data/GPS-Solutions. To run a .flow file: - Open INSTINCT, - On top in the menu bar press File > Open Flow and select the wanted flow, - If all nodes are green, run the flow from the menu with Run > Run Flow.

Zusammenfassung

Navigating during GNSS outages is becoming more important due to a significant increase of jamming and spoofing attacks and motivated by unmanned operations in urban environments. Inertial Measurement Units (IMUs) are usually the choice to mitigate the GNSS signal losses, but not all applications allow for IMUs that are precise and stable enough to rely solely on inertial navigation over longer periods of time. Payload capacity – either mass or volume – can impose strict limitations on the usage of IMUs and the demand for low-cost solutions can define additional boundary conditions. In this study, we show that fusing multiple IMUs with B-splines can improve the navigation solution considerably. The developed algorithm can fuse angular rates and accelerations from an arbitrary number of Single IMUs (SIMUs) using a Kalman filter, which outputs one combined stream of Virtual IMU (VIMU) measurements. Expressing the angular rates and the accelerations therein as the sum of B-spline base functions makes it possible to resolve temporal changes at a user-defined resolution, while only a few parameters have to be chosen. The B-spline approach does not only allow a dynamic and continuous representation of the angular rates and the accelerations which occur along a given trajectory, but also inherently avoids the problem of high-frequency noise. Several high data rate IMUs can be fused, even if sampling rates or the quality of the measurements are different. The concept of the fusion is also inherently redundant: If one SIMU fails, the fusion keeps on working with the remaining SIMUs. The gain in precision of the navigation solution is directly dependent on the number of SIMUs. The formal error could theoretically be reduced by the square root of the number of SIMUs, if averaging independent measurements. The IMU fusion based on B-splines is implemented in our open source software framework INSTINCT (INS Toolkit for Integrated Navigation Concepts and Training). Its performance will be shown in a Monte Carlo simulation of a pure inertial navigation solution in comparison to a SIMU. A dynamic simulation will show the applicability to GNSS outages, where the VIMU’s combined measurement is used in GNSS/INS-fusion. This setup is validated by drive and flight tests, utilizing an IMU sensor array that was constructed in-house and that consists of five low-cost SIMUs.

Zusammenfassung

Designing navigation software can be a very time-consuming task. Novice navigation engineers often have to start implementing everything anew and even professionals spend a lot of time adapting algorithms from other projects. This prolongs development cycles and shortens the time available for actual research. Furthermore, when projects shall be realized on test platforms, algorithms often need to be executed on low-cost hardware, like a Raspberry Pi, which can be challenging as the processing capabilities are limited. To solve these problems, the Position, Navigation, and Timing software framework INSTINCT (INS Toolkit for Integrated Navigation Concepts and Training) was developed. It is based on the flow-based programming paradigm, which abstracts functionality, like reading a file or preprocessing data, into separate modules, so-called nodes, to make the implementation more self-contained and reusable. Data flows between nodes over links and triggers the calculation of algorithms, which enables parallelization of tasks and improves the performance on multi-core processor systems. This abstraction also enables algorithm design in post-processing with data files or sensor simulators, and deployment of the algorithm onto test platforms for real-time applications by simply exchanging the input node from file reader to sensor. Besides the implementation of various file formats (e.g. RINEX), raw data-streams (e.g UBX) and sensor protocols from different manufacturers, algorithms like Single Point Positioning (SPP), inertial sensor data fusion and multi-constellation/multi-frequency Real-Time Kinematic (RTK) are available, providing a stable foundation for research or navigation product development. Furthermore, INSTINCT provides tools such as a GNSS analyzer that can be used to calculate signal combinations, detect cycle-slips and assess the quality of GNSS data. Because selecting and configuring such variety of algorithms can become quite complicated through config files, INSTINCT provides an intuitive graphical user interface (GUI), where one can select the necessary nodes, link them together and adjust algorithm parameters. Additionally, results can be directly visualized and evaluated within the software, removing the need to write evaluation scripts entirely and leading to faster iteration times. To share this software and enable potential collaboration, INSTINCT was released under an open-source license on GitHub (https://github.com/UniStuttgart-INS/INSTINCT). The presentation will cover the basic principles of flow-based programming and illustrate its advantages over existing solutions. It will be demonstrated how simple and straightforward RTK processing can be realized with INSTINCT while making use of parallel computation capabilities of the software. It will also be shown how easy results can be combined with other sensors either by loosely- or tightly coupling. The presentation will also provide an outlook on upcoming features of the software and we are going to discuss how the software can be adapted to your own R&D project or being used for teaching activities.

Zusammenfassung

Low-cost Inertial Measurement Units (IMUs) are typically subject to biases and noise. Within this study we aim to reduce these effects by combining measurements in a novel way. An integrated random walk is proposed to model acceleration and angular rates from IMU sensors which are attached to the same platform. Thereby, a Kalman Filter (KF) is used to merge the observations from the different sensors in a single state model that considers also relative biases of individual sensors, with respect to one reference sensor. Such a KF has been tested on an array of IMUs that consists of five identical devices. Figure 1 shows the Allan Deviation (ADEV) of the fused Virtual IMU (VIMU) in comparison to each single IMU (here the vertical component of the angular rate, the other components are similar). For small averaging periods τ, the ADEV is reduced considerably. Specifically, the impact of angle random walk and the bias instability are reduced. With rising τ, the VIMU’s ADEV converges towards the reference sensor’s, due to long term drifts which are not covered by the state space of the KF. In general, it can be stated that the VIMU’s ADEV is lower than any of the single devices which indicates that such an approach has the potential to improve the accuracy of low-cost inertial sensors by fusing the output from an array of such sensors. In our presentation we demonstrate the concept of fusing multiple IMUs within the flow-based software framework INSTINCT (INS Toolkit for Integrated Navigation Concepts and Training), with synthetic data and measurements from an array of IMUs. We discuss the potential of this approach and show that the output of the fusion algorithm, which contains acceleration and angular rates, can be fed into subsequent processing like in classical INS/GNSS-fusion.

Zusammenfassung

Flow-based programming (FBP) splits software functionality into modules which are triggered by data elements flowing from one module to the next. Thus, modules, which are not directly dependent on data from each other, can run independently and applications can be parallelized on a basic level what is a major improvement for the performance of multi-sensor data fusion algorithms. This paper discusses the flow-based software INSTINCT, which is a solid framework for implementing PNT algorithms and helps to reduce development times by making algorithms reusable. It provides an intuitive graphical user interface (GUI) that makes it usable in research and teaching. Inside the software, a variety of file formats and sensors for IMU und GNSS data processing are implemented. The realization of an INS/GNSS loosely-coupled Kalman filter flow is discussed and the results are compared to a flow representing a single point positioning (SPP) solution. Finally, a performance study of the algorithm on a 4th generation Intel CPU is presented, demonstrating that real-time capabilities can be ensured even on older systems.