Advancing Precise Positioning in Challenging Environments

Last Updated: 10/06/2024 10:29     Created at: 10/06/2024 10:21

Final Presentation of NAVISP Project EL1 042 now available:

On Wednesday, June 5th, 2024, ANAVS and Space Applications presented the results of the NAVISP EL1 042 project "Next Generation Motion Sensors for Hybrid GNSS INS Solutions in High Accuracy Machine Control Applications (MSFUSION)". Over 30 people from industry and research institutes followed the interesting presentation and the subsequent interactive Q&A session.

The MSFUSION project focused on developing advanced sensor data fusion techniques to improve position estimates in challenging use cases for applications ranging from pedestrian navigation to autonomous vehicle operation in environments with limited GNSS coverage. MSFUSION aimed to enhance positioning accuracy by integrating data from various sensors, including multi-GNSS receivers, INS, and complementary devices such as odometers, visual cameras, LiDAR, magnetometers and IMUs. The project emphasized using low-cost sensors and leveraging AI/ML techniques to complement traditional synthesis methods. The ultimate goal was to create a robust and cost-effective Elegant Bread Board (EBB) capable of delivering precise positioning in both outdoor and indoor scenarios.

The applications of MSFUSION's technology are vast, covering autonomous cars, robots, ships, and railways. These applications face significant challenges in environments with sporadic GNSS coverage, such as urban areas with multipath effects, ports with metallic obstructions, and mixed indoor-outdoor environments. One of the goals of the project was to improve industrial environments like warehouses, where it is important to know the location of mobile assets, such as Autonomous Guided Vehicles (AGVs), for safety and efficiency reasons. The EBB combines various types of sensors to create a complete positioning system. For instance, the pedestrian EBB includes a backpack equipped with a Jetson Orin NX processor, a Zed Stereo camera, and an Intel L515 LiDAR sensor. The vehicle EBB employs a Velodyne VLP16 LiDAR sensor for high-resolution feature extraction and a multi-IMU setup to enhance data availability.

The selection of sensors for the EBB is driven by the need to balance performance and cost. Key sensors include MEMS IMUs, visual cameras, and ranging sensors. MSFUSION explored various fusion techniques, including central preprocessing combined with EKF and AI extensions, and federated loosely coupled EKF for GNSS PPP, multi-pose estimation, and AI-aided satellite visibility estimation, to integrate data from different sensors effectively and achieve maximum performance. These techniques aim to mitigate errors and provide reliable position estimates even in GNSS-degraded conditions. Extensive testing was conducted in various scenarios, including urban environments, ports, and mixed indoor-outdoor settings. In the urban scenario, the EBB demonstrated robust performance despite GNSS signal degradation caused by obstacles like trees and buildings. The port scenario, characterized by narrow spaces and tall metallic obstructions, highlighted the EBB's ability to maintain accuracy with minimal GNSS coverage. In mixed scenarios, the EBB successfully navigated transitions between indoor and outdoor environments, proving the effectiveness of the federated filter in maintaining accurate positioning.

MSFUSION's development of advanced sensor fusion techniques and low-cost EBB solutions marks a significant advancement in precise positioning technology. The project achieved satisfactory results across various challenging environments, demonstrating the potential for wide application in both industrial and commercial settings.

MSFUSION future possible improvements aim at simplifying the architecture, incorporate SLAM approaches, and increase the robustness of visual localization methods. These enhancements will further solidify the project's contributions to the field of precise positioning and autonomous navigation.

The project was carried out in the scope of NAVISP Element 1, which is dedicated to technology innovation of the European industry in the wide PNT sector.

More detailed information can be found in the slides of the Final Presentation.