AU2021100964A4 - GPS data spoofing and malfunctioning detection system using classifiers - Google Patents

Abstract

GPS data spoofing and malfunctioning detection system using classifiers ABSTRACT: GPS-dependent positioning, navigation, and timing synchronization processes have a crucial influence on each and every day human existence. As a result, such an extensively accessed device is incredibly becoming an appealing victim for counterfeit impoverishment by terrorist attacks and cyber criminals for a variety of reasons. The significance of the Global Positioning System (GPS) and associated communication items tends to grow across a wide variety of ecological engineering and navigating technologies. Therefore, spoofing, and anti-spoofing methodologies have become a valuable exploratory target within the GPS system. Nonmilitary GPS signals are, furthermore, susceptible to radio frequency intervention. Spoofing is a deliberate intrusion which attempts to compel the GPS receiver to obtain and monitor illegitimate navigating information Review of spoofing and accurate signaling characteristics poses variations as phase, intensity, and fictitious attributes of the signal. Early-late phase, delta, and signal level as the three key attributes are taken as from correlation performance of the monitoring network. Utilizing such functions, spoofing identification can be done which use traditional machine learning algorithms like K-Nearest Neighborhood (KNN) and naive Bayesian classifiers. This invention proposes the convolutional Neural Network (CNN) as a learning model is a modern computational tool for accumulating the information needed and evaluating performance variables in problematic platforms. The computational analysis on the GPS receiver technology demonstrated reasonable identification efficiency was extracted from NN with a limited processing period. 1 - I~WflNar m GPS wr*to'r*' Lute S PRNCoda NonUnear 1 It f _ X_~ ~ 4 _ ZM _ J Ganartor Delay biurminato D LL DLL _ spnce 1Non-Unear Phasno Disriminator DU4 PLL Figure 3: GPS Spoofing and malware detection using CNN Off ne No Tracking Loop lock? .... Yes_+ No NNo Yes Algouithm +-Ye Test Success? Spoofr g larmNo Figure 4: Spoofing detection algorithm based on NN 2

Description

- m I~WflNar GPS wr*to'r*' Lute

It f _ X_~ ~ 4 _ ZM _ J S PRNCoda NonUnear 1 Ganartor Delay biurminato

DLL D LL

1Non-Unear

Phasno Disriminator DU4 _ PLL spnce

Figure 3: GPS Spoofing and malware detection using CNN

Off ne

No Tracking Loop lock? ....

Yes_+ No NNo

Yes Algouithm +-Ye Test Success?

Spoofr g larmNo

Figure 4: Spoofing detection algorithm based on NN

GPS data spoofing and malfunctioning detection system using classifiers

Description

Field of the Invention:

The effectiveness of implementing protections to telecommunications and digital networks has enhanced considerably, contributing to a range of signal safety approaches. Global Positioning System (GPS) signals are being secured from threats and GPS spoofing efforts to mislead the recipient by transmitting fraudulent transmissions. Typically, spoofing signals are marginally better than genuine signals. They can be produced by delaying and re-emitting the authoritative allocated GPS signal. The current innovation concerns the Global Navigation Satellite System (GNSS) signals. Rather precisely, the present innovation concerns methodological approaches for evaluating unless the obtained GSP signals come from a reliable source.

Background of the invention:

The accessible feature of GPS coarse acquisition protocols leads the device to possible security threats on location and timing-dependent programs. Such deliberate or malicious activities are known as eavesdropping and impersonation attack. Although blasting threatens to interrupt or weaken GPS networks by disrupting or overwhelming signals, a smarter and more dangerous spoofing assault can mimic GPS signals intended to deceive the specified recipient and contravene on incorrect location and scheduling decisions. The susceptibility to GSP spoofing is a prominent exploration field due to its influence on crucial and ever-growing GSP-dependent application fields.

Ochin et al. observed spoofing by identifying empirical incongruity when evaluating the GPS satellite parameters. The key issue with this strategy is the accessibility of authoritative stimulus knowledge until an offensive is launched. In particular, after such a spoofing attack, the obfuscation pulse parameters are set at the cutoff point permitted. Implementation and analysis of the input is a strategy that derives the features of the GPS signal. In this process, some sudden or major improvement is the phenomenon of a malware attempt.

In accordance with S. Daneshmand the spoofing remediation strategy is suggested in effectively which ameliorates spoofing signals as far as their TSP is dramatically greater than the estimated strength of genuine signals. In certain instances, though, it can inadvertently limit the strength of such genuine signals despite the inherent cone of uncertainty in the double-antenna array antenna. This dilemma can be overcome by using greater antenna arrays since the complexity of the antenna radiating patch declines tremendously as the quantity of array components increases. This spoofing prevention strategy cannot operate well enough in the scenario of numerous spoofing propagations.

P. Y. Montgomery explained the GPS antenna on the roof which has been configured to transmit genuine GPS signals. The obtained signals are intensified and then sent inside from the point source antenna. In this situation, spoofing communication can assume effect indoors where it really doesn't contravene the rules on radio signal. This configuration appears to be sufficient, though it may not reflect actual outdoor spoofing situations, particularly for multi-antenna anti-spoofing approaches. In this situation, all authoritative signals are often emitted from a single antenna. Multipath amplification and absolute spoofing and genuine signal forces are some other problems which should be mentioned when using indoor data transfer.

According to C. E. McDowell, the configuration of the antenna array is designed to track and counteract spoofing signals depending on their spatial similarity. The parallel performance phase estimates for various PRN signals are comparable to the distinct output phase observations obtained from the similar spatial field. This methodology can effectively diagnose spoofing signals and does not require any array configuration or array orientation details This strategy can efficiently differentiate against spoofing circumstances that use a single transmit antenna. In particular, multipath propagation does not reduce the efficiency of this system, because all spoofing signals have the similar signal attenuation properties.

The GSP group applied less consideration to certain a hazard in the previous research until that counterfeit was created and effectively evaluated towards a COTS receiver. This hybrid receiver/spoofer takes advantage of information of the valid GPS signals and perception of their destination related to the offender. The Assault approach catches each receiver network by associating the fraudulent signaling with the real signal for each recognizable satellite.

Psiaki, M.L explained the sophisticated receiver-based malware methodology, known as nullification, is intended to communicate radio inputs to the offender recipient. One would be the vulnerable to malicious of the threat signal and another is the derogatory of the actual signal. In the case of a received signal by the accuser recipient, the actual signal element is suspended and only the spoofing element is left. The hazard of this sporadic attack is tremendous. Even so, owing to the accurate orientation of the pulse width and amplitude aligning, the nulling attack is incredibly troublesome to execute.

ATC is a function offered by ground regulators who guide air vehicles via protected airspace on the surface. For the intention of the prevention of accidents and the organisation of air traffic, aerial vehicles regularly send ADS-B or Flarm signals in order to announce their aircraft activity. These status notifications are obtained by ATC base stations create the so-called ATC info. Remember that because ADS-B and Flarm security mechanisms are open to the community anybody can track and capture this ATC information using state-of-the-art easily specified radio equipment.

The OpenSky Network is implemented as a crowdsourcing project to gather ATC information and make the documents accessible to the public. OpenSky land detectors are the mostly built and run by aviation enthusiasts and volunteer organizations. Participants help to track ATC data, which is then transmitted to the cloud server via the Web. More than 200, 000 notifications a moment are received at peak periods from more than 800 base stations all over the world.

Artificial intelligence methods have been employed in recent decades to monitor a wide variety of technologies. This paper introduces a modem counterfeit detection approach used in the GPS software Defined Receiver. The suggested methodology uses Neural Networks to identify irregular association anomalies for hacking identification. As seen further, by implementing Neural Network, the signal index travels above the permitted saturation point and malicious is detected when the attacker manages to exploit the recipient's correlation limit. The approach presented does not include any excessive equipment and does not enhance the dimensions of the receiver or the manufacturing costs. There is no need for any authoritative signal during learning in our strategy.

Objects of the Invention:

* The main objective is to handle the detection ofjammers as an issue of classification tasks, relying on the features extracted determined by the GPS receiver of the Intermediate Frequency (IF) sample were collected. • The second objective is to recommend the effective Convolutional Neural Network (CNN) based Classifiers with a prediction performance of 95%. • The another is to supply the scientific world with an open-access, extensive archive of safe and scrambled GSP signal harmonic frequencies for a range of Carrier-to-Noise Ratio (C/NO) and Jammer-to-Signal Ratio (JSR) environments.

Summary of the Invention:

The initial data acquisition phase documents the valid signals of the GPS satellites. These transmissions are strengthened and processed at the front-end at 5-7MHz. After converting to Intermediate Frequency (IF), Filtering and Analog-to-Digital Conversion (ADC), the sampled time-discrete signal is transferred to the SDR. The input size of CNN is specified, the input is moving via a sequence of convolutional layers of the similar or varying window sizes. In each convolution sheet, the detector scans the feature vector from left to right and up to down using a 2-pixel pace, which is the number of points that eachfilter moves.

The deployment is the first step in the activity of the Global Navigation Satellite System (GNSS) receiver. This method helps in the assessment of whether or not the satellite signal is embedded in the received signal and gives preliminary measurements of the dimensions including the code delay and the Doppler change of the signal emitted by the satellite. Both GPS receivers enforce certain an extraction mechanism by assessing the so-called cross-ambiguity feature (CAF), typically in a discreet period. The NN is equipped off-line and then validated and prepared to be used in the training methodology. The examination of the derived aspects is carried out simultaneously recognition of the malicious signal triggers a warning for the recipient. When spoofing is in progress, the procedure is replicated. The specification of the correct design and learning methodology is of considerable significance for using MLP NN. Appropriate layout involves the choosing of the optimum number of layers, the number of neurons in each layer and the correct activation mechanism for each neuron. The optimum architecture is chosen by a trial and-error development process.

Detailed Description of the Invention:

Figure 1: High level GPS attack

Figure 2: Data Acquisition

Figure 3: GPS Spoofing and malware detection using CNN

Figure 4: Spoofing detection algorithm based on NN

Detailed Description of the Invention:

Figure 1 shows the various attacks in GPS navigation system.

Figure 2 demonstrates the data collection process. The initial data acquisition phase documents the valid signals of the GPS satellites. These transmissions are strengthened and processed at the front end at 5-7MHz. After converting to Intermediate Frequency (IF), Filtering and Analog-to-Digital Conversion (ADC), the sampled time-discrete signal is transferred to the SDR.

Figure 3 suggested, the input size of CNN is specified, the input is moving via a sequence of convolutional layers of the similar or varying window sizes. In each convolution sheet, the detector scans the feature vector from left to right and up to down using a 2-pixel pace, which is the number of points that each filter moves. At the end, the convolution layers are preceded by the Totally Connected (FC) layers and the ultimate softmax layer employed for supervised classification.

Figure 4 explained the SDR observing loop where the suggested detection algorithm dependent on Multi (MLP) NNs is introduced. The characteristics are derived from the interaction outcomes after the loop locks are tracked. The NN is equipped off-line and then validated and prepared to be used in the training methodology. The examination of the derived aspects is carried out simultaneously recognition of the malicious signal triggers a warning for the recipient. When spoofing is in progress, the procedure is replicated. The specification of the correct design and learning methodology is of considerable significance for using MLP NN. Appropriate layout involves the choosing of the optimum number of layers, the number of neurons in each layer and the correct activation mechanism for each neuron. The optimum architecture is chosen by a trial and-error development process.

Claims (7)

GPS data spoofing and malfunctioning detection system using classifiers CLAIMS:

1. GPS data spoofing and malfunctioning detection system consists of Signals gathering Measure spatial correlation Movable antenna Data storage Electronic devices Convolutional neural network like MLP

2. From claim 1, the signal gathering process acquires the initial data acquisition phase documents the valid signals of the GPS satellites. These transmissions are strengthened and processed at the front-end at 5-7MHz.

3. According to claim 1, measure of spatial correlation is evaluated among samples gathered. Evaluate if the aforementioned signals come from a unified source on the basis of the said calculation of the spatial similarity among the claimed observations. Assessing the frequency of rise or decline of said ordered collection said rate of rise or decline is said to be the indicator of spatial interaction

4. The claim 1 also includes the movable antenna. The acquired signals are accessed by using multi path movable antenna. Movable antenna is space-translated when collecting said input signals to construct a simulated antenna array.

5. Electronic devices like computers, laptops with high-speed network. Machine recognizable media possessing embedded on the computer understandable commands which, when performed, execute a procedure for measuring whether the input signals come from a particular source.

6. The cloud storage services like amazon services, IBM are used to store the gathered signals.

7. The NN is equipped off-line and then validated and prepared to be used in the training methodology. The examination of the derived aspects is carried out simultaneously recognition of the malicious signal triggers a warning for the recipient.The specification of the correct design and learning methodology is of considerable significance for using MLP NN.

GPS data spoofing and malfunctioning detection system using classifiers

Drawings: 2021100964

Figure 1: High level GPS attack

Figure 2: Data Acquisition

Figure 3: GPS Spoofing and malware detection using CNN

Figure 4: Spoofing detection algorithm based on NN