COLLABORATIVE PROCESSING OF DISTRIBUTED RECEIVERS OF OPPORTUNITY FOR JAMMING AND SPOOFING MITIGATION

Register for the Final Presentation of the NAVISP EL1 -029 Project "Jamming And Spoofing mitigation by ProcEssing Receivers of opportunity (JASPER)”

The fast evolution of the Internet of Things (IoT), commercial Cloud platforms, and the future 5G standards, are boosting the development of new applications and technologies in the Position Navigation and Timing (PNT) field. More devices are expected to be increasingly connected to the Internet in the next few years, especially with 5G expected to bring very soon higher data rates and lower latencies to those devices than the current 4G Long Term Evolution (LTE). Additionally, some mobile operating systems provide increased access to raw measurements, including GNSS, Wi-Fi and other sensors (as an example, Wi-Fi round-trip time is available in the first developer preview build of Android P). This makes these devices ideal Receivers of Opportunity (RoO) to assist in PNT determination. In this context, the exploitation of GNSS signal snapshots and observables, together with other sensors’ measurements and peer-to-peer communications, seems to constitute a feasible approach to enable new signal processing techniques. In particular, the collaborative processing of distributed RoO in nearby locations can be used to solve some of the limitations faced by current GNSS receivers in the presence of jamming and spoofing attacks.

Considering this, the activities of JASPER focused on the development of collaborative positioning techniques using RoOs for the detection and mitigation of jamming and spoofing attacks.  This had the goal of creating an end-to-end concept simulator to demonstrate the capabilities of collaborative positioning by utilizing GNSS signals from distributed RoOs and the peer-to-peer (P2P) ranges between them. Machine Learning (ML) models were also applied to exploit the signal characteristics and detect “trusted” RoOs that could be used in collaborative positioning. 

The Software-based Concept Demonstrator (SW-CD) developed, implemented, and tested, in the JASPER project, illustrated that the collaborative P2P engine implemented can compute the target receiver position by fusing measurements from the available GNSS signals and ranges from nearby trusted RoOs, while the untrusted RoOs identified based on the machine learning model were discarded. The SW-CD is configurable to operate in several processing modes which include (1) independent solution mode (GNSS-only), (2) an independent solution with height-aiding mode, (3) hybrid solution mode (using P2P ranges and sparse GNSS signals), (4) collaborative positioning mode (hybrid solution with ML). Additionally, the SW-CD includes the ability to operate directly in the positioning domain (DPE) based on GNSS signals, exploiting a coarse estimation of position and time.  The machine learning algorithm used to generate the model for jamming detection was Convolutional Neural Network and Long Short-Term Memory Network (CNN LSTM), whereas, for spoofing detection, the Support Vector Machine (SVM) method was used. The developed machine learning models were able to detect whether a RoO was “impacted” by jammer/s and spoofer/s with an accuracy of about 95%. It's been observed that the collaborative positioning output from the SW-CD is more robust as a result of detecting and removing untrustworthy RoOs and the inclusion of additional ranges to the positioning algorithm. As expected, the accuracy of the DPE approach is better than the independent solution mode, with the trade-off, that the method is computationally expensive.

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

Please check the Agenda and register for the event!