System and Method Configured to Perform Penetration Testing of Virtual Reality Systems using Machine Learning
The system automates VR system penetration testing and vulnerability assessment using machine learning, addressing the lack of efficient methods by identifying and rating cybersecurity risks, including insecure protocols and weak encryption.
Patent Information
- Application Number
- SA1020254908
- Authority / Receiving Office
- SA · SA
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2025-07-08
- Publication Date
- 2026-10-01
AI Technical Summary
Existing VR systems lack efficient and automated methods for penetration testing and vulnerability assessment, requiring manual setup and expertise, and are prone to cybersecurity risks due to insecure network protocols and weak encryption.
A system and method utilizing machine learning to automatically identify VR systems, apply predefined penetration tests, and generate comprehensive assessment reports, including risk ratings, without manual intervention.
Automated penetration testing and vulnerability assessment efficiently identify and rate cybersecurity issues in VR systems, providing a comprehensive risk assessment without manual setup, and detecting insecure protocols and weak encryption.
Smart Images

Figure 48F25D8F-5F99-43A9-B7CD-C760C17188B1 
Figure 4B9350D9-80BC-41C1-B84F-1E81D2D5361E 
Figure 74FF3E58-CA6D-40DB-A9C0-AE3FBD58FA32
Abstract
Description
Models and representative applications that conform to the teachings included in the current disclosure are directed to System 100 and Method 500, designed to perform penetration testing of VR systems using machine learning. With reference to Figure 1, in an application compatible with the invention, System 100 includes an Evaluation System 102 or a subsystem operationally connected to the VR System 104 under test, with Evaluation System 102 designed to generate and output an Evaluation Report 106 indicating the results of the penetration test of the VR System 104 under test. In one application, the VR System 104 includes a Headset 108, a VR Module 110, and VR System 112 variants. In another application, the VR System 104 includes gloves or other accessories that allow the user to operate the VR System 104 with the VR Headset 108. In an application compatible with the invention, the VR Module 110 includes hardware, software, or a combination of hardware or software to implement a VR application, allowing the user to interact with a VR environment. For example, the VR System 104 is implemented using a Headset 108 and a Module 110 compatible with the Android operating system, publicly available from Google. In another example, the VR 104 system is implemented using a VR 108 headset and a VR 110 unit compatible with the publicly available Android operating system from GOOGLE LLC.In another example, the VR 104 system is implemented using the VR 108 headset and the VR 110 module, which are compatible with any publicly available iOS operating system from Apple. In another example, the VR 104 system is implemented using the VR 108 headset and the VR 110 module, which are compatible with any known operating system designed to run VR applications. In one application, the VR 110 module runs VR software written in Virtual Reality Modeling Language or Virtual Reality Markup Language (VRML). In another application, the VR 110 module runs VR software written in any known programming language designed to operate the VR 104 system for user use in a VR environment. In an application compatible with this invention, the VR 104 system includes memory designed to store VR 112 system variables. For example, VR 112 system variables include device drivers, file systems, and Media Access Control (MAC) addresses associated with the VR 104 system and its components 108 and 110. The 114 connection effectively links the 102 evaluation system to the VR 104 system via a known communication device. In an application compatible with this invention, the 114 connection includes a physical wired connection. For example, the 114 connection is a USB-compatible cable. A USB-compatible cable is a USB Type-B cable, a USB Type-C cable, or any other known USB-compatible device. In another example, the 114 connection includes any known physical wired connection between the 102 evaluation system and the VR 104 system. In yet another application, the 114 connection is a wireless connection between the 102 evaluation system and the VR 104 system. In yet another application, the 114 connection is a combination of a wired and a wireless connection between the 102 evaluation system and the VR 104 system. In one application, connection 114 is a network. For example, the network is the internet. In another example, the network is an internal network or an intranet of an organization. In yet another example, the network is a heterogeneous or hybrid network that includes both the internet and an intranet. Connection 114 is linked to connection 116 settings, which define the operating variables and other functions of connection 114. For example, connection 116 settings include network protocols, i.e., encryption algorithms that encrypt the communications carried by connection 114, and network traffic variables. In one application, connection 116 settings are stored in connection 114's memory. The Assessment System 102 includes a hardware-based processor (HPC 118), memory (RAM 12) designed to store instructions and connected to the HPC 118 to provide instructions to the HPC 118, a communication interface (ICI 122), an input / output device (I / O 124), and a set of modules (IC126 and I28) designed to execute the instructions provided to the HPC 118. The set of modules (IC126 and I28) includes a machine learning (ML) module (IC126) designed to implement machine learning algorithms, and a report generation module (IC128) designed to generate the Assessment System 106 report. For example, the Assessment System 102 is implemented on a laptop, allowing it to be portable and easily connected to the VR System 104 under test, as is the case in the field. In another example, a kit is provided that includes a case, such as a hard case or document case, designed and sized to store the laptop, including the 102 rating system, as well as a 112 connection, such as a USB cable.In another example, the 102 evaluation system is implemented on a relatively small single-board computer (SBC) such as the publicly available RASBERRY PI devices from RASBERRY PI LTD. Memory 120 stores predefined test cases specific to the VR 130 resource, designed to evaluate the VR 104 system under test. The communication interface 122 is designed to send and receive data to and from the VR 104 system, respectively, via communication link 114. In an application compatible with the invention, the input / output device 124 includes a display or monitor designed to display the evaluation report 106 and other user information via a graphical user interface (GUI) 132. The machine learning module 126 includes a classification module 134, which will be described in more detail below. In another application, the input / output device 124 includes a paper printer designed to output a printed copy of the evaluation report 106 to the user. Figure 2 shows a schematic of a 200 computing device that includes a 202 processor with code, 204 memory, and a 206 communication interface. Optionally, the 200 computing device can include a 208 user interface, such as an input device, an output device, or an input / output device. The 202 processor, 204 memory, 206 communication interface, and 208 user interface are connected to each other via any known connections, such as the system bus, network, etc. Any component, combination of components, or module from System 100 in Figure 1 can be implemented by the corresponding 200 computing device. For example, the 118 processor, 120 memory, 122 communication interface, 124 input / output device, 126 machine learning module, 128 report generation module, 132 GUI, and 134 classification module shown in Figure 1 can all be implemented by the corresponding 200 computing device shown in Figure 2 and described below. It is understood that a Compute 200 can include different components. Alternatively, it can include additional components. In another implementation, some or all of the functions of a particular component can be performed by one or more different components. A Compute 200 can be implemented by a virtual computing device. Alternatively, a Compute 200 can be implemented by one or more computing resources in a cloud computing environment. Additionally, a Compute 200 can be implemented by a group of known computing devices. The 202 processor can be a hardware-based processor that implements a system, subsystem, or module. It can include one or more general-purpose processors. Alternatively, it can include one or more special-purpose processors. The 202 processor can be fully or partially integrated with memory (204), communication interface (206), and user interface (208). In another application, the 202 processor can be implemented by any known hardware-based processing device, such as a controller, integrated circuit, microchip, central processing unit (CPU), microprocessor, system-on-a-chip (SoC), field-programmable logic gate array (FPGA), or application-specific integrated circuit (ASIC). Additionally, the 202 processor can include an array of processing elements designed to perform parallel processing. In yet another application, the 202 processor can include an array of nodes or artificial neurons designed as an artificial neural network.The 202 processor can be designed to implement any known machine learning (ML)-based hardware, any known artificial intelligence (AI)-based hardware, and any known artificial neural networks, including convolutional neural networks (CNNs). Memory 204 can be implemented as a non-volatile storage medium that is readable by a computer, such as a hard disk drive, solid-state drive, erasable programmable read-only memory (EPROM), universal serial bus (USB) storage device, floppy disk, compact disc-read-only memory (CD-ROM) disk, or digital multipurpose disc (DVD), cloud-based storage, or any other known non-volatile storage. The code of Processor 202 can be stored in the processor's internal memory. This code can be instructions executed on hardware components. Alternatively, it can be instructions executed in programs. These instructions can be machine language instructions that Processor 202 can execute to instruct Computing Device 200 to perform the functions described in this document. Alternatively, the instructions can be text instructions that can be executed by a text interpreter designed to instruct Processor 202 and Computing Device 200 to execute the instructions specified in the text instructions. In another implementation, the instructions can be executed by Processor 202 to instruct Computing Device 200 to execute an artificial neural network. Processor 202 can be implemented using hardware components or software, such as code. Processor 202 can implement a system, a subsystem, or a module, as described in this document. Memory 204 can store data in any known format, such as databases, data structures, data warehouses, or neural network variables. Data can be stored in a table, a separate file, a file system, a heap file, a B+ tree, a hash table, or a hash statement unit. Memory 204 can be implemented using any known memory, including RAM, cache memory, register memory, or any other known memory device designed to store instructions or data for quick access by processor 202, including storing instructions during execution. The 206 interface can be used with any known device designed to perform the functions of the Computing Device 200 interface described in this document. The 206 interface can implement a wired connection between the Computing Device 200 and another entity. Alternatively, the 206 interface can also implement a wireless connection between the Computing Device 200 and another entity. The 206 interface can be implemented via Ethernet, Wi-Fi, Bluetooth, or USB. The 206 interface can also send and receive data across the network and to other devices using any known communication link or protocol. A 208 user interface can be any known device designed to perform user input and output functions. A 208 user interface can be designed to receive user input. Alternatively, a 208 user interface can be designed to output information to the user. A 208 user interface can be a computer monitor, television, speaker, computer headset, or any other known device operationally connected to a 200 computing device and designed to output information to the user. User input can be received through a 208 user interface that implements a keyboard, mouse, or any other known device operationally connected to a 200 computing device for user input. Alternatively, a 208 user interface can be implemented by any known touchscreen. A 200 computing device can include a server, PC, laptop, smartphone, or tablet. With reference to Figures 1 and 3, in an implementation consistent with the invention, the Classification Unit 134 of the Machine Learning Unit 126 incorporates a Neural Network 302 comprising a set of Nodes 304 or artificial neurons arranged in a set of Layers 306–314. The Layers 306–314 set includes an Input Layer 306, at least one Hidden Layer 308–312, and an Output Layer 314. The Neural Network 302 receives VR System 112 variables, including device drivers, file systems, and Media Access Control (MAC) addresses, in the Input Layer 306 and generates output in the Output Layer 314. The resulting output classifies the VR System 104 to identify the VR resource for the VR System 104, the VR Unit 110, and the operating system and installed applications for the VR System 104. For example, installed applications are applications installed by default that are implemented by the VR System 104 to provide a VR environment for the user. In another application, Classification Module 134 includes a Support Vector Machine (SVM) designed to classify and identify the VR vendor for VR 104, the VR 110 module, and the operating system and applications installed on VR 104 through VR 112 system variables, including device drivers, file systems, and MAC addresses. In another application, Classification Module 134 includes any known machine learning processors and algorithms designed to classify and identify the VR vendor for VR 104, the VR 110 module, and the operating system and applications installed on VR 104 through VR 112 system variables, including device drivers, file systems, and MAC addresses. In an application compatible with the invention, the Classification Unit 134 is trained using predefined test cases for VR Resource 130, as illustrated in Figures 1 and 4. Referring to Figure 4, each set of VR Resource 402, 404, and 406 is associated with a set of tests 408–424, such as penetration tests, with varying traffic behaviors and different virtualizations based on the VR Resource to account for different VR environments. Each of the test cases for VR Resource 408–424 is designed for a VR Resource system or device environment. For example, VR Resource 402, 404, and 406 are designated as “VR Resource 1,” “VR Resource 2,” and even “VR Resource N,” where N is an integer greater than one. The VR 402 vendor is associated with tests 408-412, labeled "Test 1," "Test 2," and up to "Test P," respectively, where P is an integer greater than one. The VR 404 vendor is associated with tests 414-418, labeled "Test 1," "Test 2," and up to "Test Q," respectively, where Q is an integer greater than one.The VR 406 vendor is associated with tests 420-424, called "Test 1", "Test 2", and up to "Test R", where R is an integer greater than one. For example, the 416 test associated with a VR 404 resource is designed to identify virtual applications installed on an Android-based VR headset by performing an input validation test that covers a wide range of potential cybersecurity attacks, including the Open Web Application Protection Project (OWASP) list of the top 10 most dangerous mobile and web attacks. The OWASP top 10 most dangerous mobile and web attacks for 2024 include: M1: Improper use of credentials, M2: Inadequate supply chain security, M3: Insecure authentication / authorization, M4: Inadequate input / output validation, M5: Insecure communications, M6: Inadequate privacy controls, M7: Inadequate two-factor authentication, M8: Poor protection design, M9: Insecure data storage, and M10: Inadequate encryption. In an alternative application, the Machine Learning Unit 126, including the Classification Unit 134, is also trained and designed to check for the use of any insecure network protocols in communications over the 112 link. The Machine Learning Unit 126, including the Classification Unit 134, is also trained and designed to check for and analyze the weak encryption algorithms used by the VR 104 system or the 114 link. In addition, the Machine Learning Unit 126, including the Classification Unit 134, is also trained to analyze VR network traffic to and from the VR System 104 for monitoring external cloud traffic based on traffic behavior and the VR Unit 110. Furthermore, the Machine Learning Unit 126, including the Classification Unit 134, is also trained to perform validation checks on applications executed by the VR System 104. Moreover, the Machine Learning Unit 126, including the Classification Unit 134, is trained or designed to determine whether VR network traffic is considered normal behavior or whether VR network traffic poses a cybersecurity risk. In one application, the Classification Unit 134 is trained from predefined test cases specific to the VR 130 resource using supervised learning. In another application, training is performed using any known machine learning technique. Once the Classification Unit 134 is trained, the Communication Interface 122 of the Assessment System 102 is operationally connected to the tested VR 104 system using Communication Link 114. For example, Communication Link 114 is physically connected to the VR 104 system port. In response to communication signals from Communication Interface 122, Processor 118 detects the VR 104 system's connection to the Assessment System 102 using a known detection technique. For example, Processor 118 is designed to detect a connection to the VR 104 system using either the handshake or acknowledgment communication protocol. The Assessment System 102 then identifies the VR resource for VR System 104, the VR Module 110, and the operating system and applications installed on VR System 104. After identification, the Assessment System 102 automatically applies tests specific to the VR resource associated with the identified VR System 104, the VR Module 110, and the operating system and applications installed on VR System 104. For example, when it identifies that VR System 104 is associated with the VR 404 resource shown in Figure 4, the Assessment System 102 applies tests 414, 416, and 418 to the identified VR System 104 and obtains the penetration test results associated with the identified VR System 104. In another application, the Evaluation System 102 receives Link 116 settings associated with Link 114 between the Evaluation System 102 and the VR System 104 under test. For example, these Link 116 settings define the operational variables and other functions of Link 114, such as network protocols (i.e., encryption algorithms that encrypt communications carried by Link 114) and network traffic variables. The Machine Learning Module 126, including the Trained Classification Module 134, processes these Link 116 settings to identify vulnerabilities in the VR System 104 associated with Link 114. Once the Machine Learning Unit 126, including the trained Classification Unit 134, identifies penetration test results related to the VR 104 system and its applications, and optionally, vulnerabilities in the VR 104 system related to the 114 connection, the Reporting Unit 128 generates Assessment Report 106 for all detected cybersecurity issues, along with risk ratings for each issue. In an application, the Assessment Report 106 is formatted by the Reporting Unit 128 for viewing or printing by the Input / Output Device 124. For example, the Input / Output Device 124 generates and outputs the Assessment Report 106 to GUI 132 so that the user can view and manipulate it, such as emailing the Assessment Report 106 to someone, storing the Assessment Report 106, annotating the Assessment Report 106, or printing the Assessment Report 106.In another example, Input / Output Device 124 includes a web browser, and Report Generation Module 128 formats Assessment Report 106 in Hypertext Markup Language (HTML) for display on the browser, so that the user can view and process Assessment Report 106, such as emailing Assessment Report 106 to someone, storing Assessment Report 106, adding annotations to Assessment Report 106, or printing Assessment Report 106. In another example, the Assessment Report 106 includes an alert, message, or notification to the user, such as a system administrator or cybersecurity expert, regarding cybersecurity issues and risks of the VR 104 system, any insecure network protocols, any weak encryption algorithms used by the VR 104 system or connection 114, and whether VR network traffic is considered normal behavior or whether VR network traffic will pose a cybersecurity risk. Referring to Figures 5A-5B, the computer-based Method 500 involves receiving predefined test cases for the VR 130 resource in the Machine Learning Unit 126 in step 502, and training the Machine Learning Unit 126 using predefined test cases for the VR 130 resource in step 504. Method 500 then discovers the 114 connection of the Assessment System 102 to the VR 104 system under test in step 506, receives the VR 112 system variables for the VR 104 system under test in step 508, and determines the characteristics of the VR 104 system under test using the trained Machine Learning Unit 126 in step 510. Method 500 then performs predefined penetration tests for the VR resource designed for the selected VR 104 system in step 512, and obtains penetration results for the VR 104 system in step 514. Finally, Method 500 generates and outputs an Assessment Report 106 on the VR 104 system from the penetration test results in step 516. In another implementation, Method 500 receives the Link 116 settings for Connection 114 in step 518, identifies vulnerabilities in VR 104 associated with Connection 114 in step 520, and generates and outputs another assessment report on VR 104 from the vulnerabilities identified in step 522. In an alternative implementation, the vulnerabilities identified for VR 104 associated with Connection 114 are included with the Assessment Report 106 generated from the penetration results of VR 104. In an application compatible with the invention, a computer-readable, non-caching storage medium stores processor-executable instructions up to 102. These instructions comprise steps 502–522 of Method 500, as illustrated in Figures 5a and 5b. It is understood that System 100 and Method 500 perform such penetration testing and vulnerability assessment not only on VR systems, but also on augmented reality (AR), mixed reality (MR), extended reality, and other known computer-based reality systems. Therefore, by automating the process of identifying the VR 104 system under test and automatically applying appropriate penetration tests specifically designed for that VR 104 system, System 100 and Method 500 efficiently and accurately identify all cybersecurity issues of the VR 104 system under test, along with their risk ratings, without requiring a cybersecurity expert or the tedious setup and installation of the VR 104 system. Furthermore, automated penetration testing and vulnerability assessments are performed on the applications installed on the VR 104 system, and System 100 and Method 500 then generate Assessment Report 106, which presents a comprehensive risk assessment of the applications deployed on the VR 104 system. Parts of the methods described in this document can be implemented by software or firmware in a machine-readable form on a physical or non-temporary storage medium. For example, the software or firmware could be a computer program containing computer program code modified to make the system perform the various actions described in this document when the program is run on a suitable computer or device, and where the program can be executed on a computer-readable medium. Examples of physical storage media include computer storage devices with computer-readable media such as disks, flash drives, flash memory, and the like, which do not contain diffuse signals. Diffuse signals can be present on a physical storage medium. The software can be suitable for execution on a parallel or serial processor so that the various actions described in this document can be performed in any suitable order, or simultaneously. It should also be understood that similar or identical numbers in drawings represent similar or identical elements through different shapes, and that not all components or steps described and illustrated by reference to shapes are required for all models, implementations, or arrangements. The terminology used in this document is for the purpose of describing specific applications only and is not intended to restrict the invention. As used in this document, singular forms include plural forms unless the context clearly indicates otherwise. It shall also be understood that the terms "contains," "includes," "involves," "comprises," and / or "includes," and their variations, when used in this specification, indicate the presence of specific attributes, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more attributes, integers, steps, operations, elements, components, and / or combinations thereof. The directives used in this document are for convention and reference purposes only and should not be interpreted as restrictive. However, it is understood that these terms can be used to refer to an operator or user. Therefore, no restrictions are implied or inferred. Furthermore, the use of ordinal numbers (e.g., first, second, third) is for differentiation, not enumeration. For example, the use of the word "third" does not imply the existence of a corresponding "first" or "second." The phrases and terms used in this document are for descriptive purposes and should not be considered restrictive. The use of the phrases "including," "incorporating," "with," "contains," "includes," and their various forms in this document is intended to include the items listed thereafter and their equivalents, as well as additional items. While the disclosure may describe several representative applications, it will be understood by those skilled in the art that various modifications can be made, and equivalents can be substituted for elements thereof, without straying from the substance and scope of the invention. Furthermore, many modifications will be appreciated by those skilled in the art to adapt a particular tool, setting, or material to the applications of the disclosure without straying from its essential scope. Therefore, the invention is not intended to be limited to the specific applications disclosed, or to the best-foreseeable method of implementing the invention, but rather to encompass all applications falling within the scope of the attached protective elements. The subject matter described above is provided for illustrative purposes only and should not be interpreted as restrictive. Various modifications and changes may be made to the subject matter described in this document without following the representative models, implementations, and applications illustrated and described, and without straying from the true substance and scope of the invention covered by the present disclosure, which is identified by the set of citations in the following protection elements, structures, functions, or steps equivalent to these citations.
Claims
1. A computer-based system designed to perform penetration testing on a virtual reality (VR) system, comprising: a hardware-based processor; memory designed to store instructions and connect to the hardware-based processor to provide instructions to the processor; and a set of modules designed to execute the instructions provided to the processor. This set of modules includes: a machine learning module designed to receive VR system variables, determine the VR system's characteristics from these variables, identify the VR system, perform a penetration test on the VR resource that matches the identified characteristics, and generate penetration test results associated with the VR system; and a reporting module designed to generate and output an evaluation report of the VR system using the penetration test results.
2. A computer-based system according to security element 1, where memory stores a set of predefined test cases specific to the VR resource, and where the machine learning module is trained to determine the VR system's characteristics from the VR system variables using a set of predefined test cases specific to the VR resource. 3.The computer-based system, according to Security Element 1, includes a machine learning module designed to automatically determine the VR system's characteristics from VR system variables and automatically perform a penetration test on the VR resource corresponding to those characteristics.
4. The computer-based system, according to Security Element 1, also includes a communication interface and a connection linking the communication interface to the VR system. The processor is designed to detect the communication link of the VR system's communication interface. The machine learning module, responding to the detection of the communication link, determines the VR system variables, defines its characteristics, and performs a penetration test on the VR resource.
5. The computer-based system, according to Security Element 3, includes a physical wired connection.
6. The computer-based system, according to Security Element 3, includes a communication link associated with a set of link settings. The machine learning module, responding to the set of link settings, determines the VR system vulnerabilities associated with the communication link, and the assessment report includes the identified vulnerabilities.
7. A computer-based system, as defined by Security Element 1, where the machine learning module includes a neural network with a set of nodes designed in a layered structure, designed to classify the VR system by its VR system variables through the identification of VR system properties.
8. A computer-based system, as defined by Security Element 1, where the VR system variables specify at least one of the device driver, file system, and MAC address, and where the identified properties specify at least one of the VR resource, VR module, operating system, and installed application associated with the VR system.
9. A computer-based system, as defined by Security Element 1, also includes an output device with a graphical user interface (GUI) designed to display the evaluation report.
10. A computer-based approach, which includes: discovering a communication link between an assessment system and a virtual reality (VR) system; receiving VR system variables in the assessment system from the VR system via the communication link; determining the VR system's characteristics using a machine learning module, and thus identifying the VR system from these characteristics; performing a predefined penetration test specific to the VR resource on the identified VR system; generating the penetration test results; and generating and outputting an assessment report on the VR system from the penetration test results.
11. A computer-based approach, according to Security Element 10, also includes: storing a set of predefined test cases specific to the VR resource in memory; and training the machine learning module to determine the VR system's characteristics from the VR system variables using a set of predefined test cases specific to the VR resource.
12. A computer-based approach, according to Security Element 10, where the machine learning module is designed to automatically determine the VR system's characteristics from the VR system variables and automatically perform a penetration test on the VR resource based on the defined characteristics.
13. The computer-based approach, as defined by Security Element 10, also includes: connecting a VR system connection; discovering the VR system connection; and, in response to connection discovery, performing the steps of receiving VR system variables, identifying properties, and performing a penetration test of the VR resource.
14. The computer-based approach, as defined by Security Element 13, also includes: receiving a set of connection settings associated with the connection; and identifying VR system vulnerabilities associated with the connection using a machine learning module, where the generation and output of an assessment report may include the identified vulnerabilities.
15. The computer-based approach, as defined by Security Element 10, involves identifying at least one VR system variable from the device driver, file system, and MAC address, and identifying at least one VR resource, VR module, operating system, and installed application associated with the VR system.
