Disaster recovery planning involves the creation of detailed instructions regarding how to respond to disasters affecting technology systems [1]. Intelligent Physical Security Systems plans are devised to respond to incidents like natural disasters, cyber-attacks, power outages, theft, arson, and terrorism. Such disasters can lead to the damage or loss of critical components of an organization’s Information Communication Technology (ICT) infrastructure [1]. Thus, bringing an organization’s operations to a halt. Physical security is closely related to disaster recovery. Firstly, physical security helps prevent the damage or loss of critical ICT infrastructure by minimizing threats [2]. Physical security is also critical to minimizing the amount of damage if a security threat were to arise [3]. For example, an alarm system can help detect and respond to a burglary in progress. The problem with many existing physical security solutions is that they require human supervision to detect threats effectively [4]. For instance, CCTV cameras require human supervision to detect intruders. There is a need for integrated solutions that can intelligently detect threats and alert an organization to minimize the likelihood of disasters. A new intelligent video analytics solution has been devised to address drone threats and human intrusion [5]. Another proposed solution is an intelligent anti-tamper using machine learning algorithms to detect threats [4]. The new measures must be evaluated to determine if they help enhance physical security.

Video Analytics Physical Security Solution

Reference [5] promotes a new video analytics solution designed to enhance physical security by detecting drones and human intrusion. The particular project was designed under a project termed the “DEFENDER” project [5]. Physical security threats have been evolving [6]. The increasing purchase and use of drones by individuals has resulted in a new security challenge to the physical security of an organization’s ICT infrastructure [6]. Attackers can now use drones to spy on physical premises hosting ICT infrastructure for intrusion purposes [5]. Attackers can also use drones to damage physical ICT infrastructure [5]. The threat presented by drones is not adequately addressed by existing physical security solutions like CCTV cameras, security guards, and alarm systems. There is a need for an effective solution to address the drone threat.

The video analytics solution uses drones and sensors to continuously obtain a stream of data from the surrounding environment [5]. The technology also uses neural networks [5]. Using neural networks provides the technology with the intelligence needed to intelligently identify potential threats [7]. Once drones and/or humans are detected, the technology alerts relevant parties regarding an intrusion. The particular version of the technology that focuses on drones is “NVIDIA GeForce GTX 1080 for drone detection” [5]. The version focusing on human intruders is “NVIDIA GeForce RTX 2070 Max-Q Design for detecting human intruders” [5]. The technology provides early-stage threat detection capabilities, which are critical to disaster prevention.

Figure 1. NVIDIA GeForce GTX 1080 for Drone Detection System Layout [5]

Intelligent Physical Security Systems

According to Figure 1 above, the video analytics solution for drone detection has a ground-based station that runs on the GTX 1080 processor. The ground-based system can control a drone used to search for foreign drones. The ground-based station solution receives aerial images taken above the facility it is deployed to secure. The Station then analyzes the image data using neural networks to determine if a drone is above the facility. The technology is quite sophisticated.

Intelligent Anti-Tamper Solution

A new autonomous physical security solution using machine learning is currently being developed [4]. Currently, existing anti-tamper solutions have quite limited capabilities. Such systems help defend against quite limited forms of threats and attacks [4]. The reason why the systems protect against limited attacks is that they have deterministic tamper responses. The Intelligent Physical Security Systems depend on pre-programmed instructions to detect and respond to threats in the environment [4]. Future states of such systems cannot be determined by randomness or learning. The deterministic nature of currently existing anti-tamper systems renders them incapable of dealing with stealthier attacks [4]. There is an urgent need for an intelligent system that can respond to stealthier forms of physical attacks on-premises hosting critical ICT infrastructure.

The proposed intelligent anti-tamper solution is designed to overcome the challenges of existing anti-tamper systems. The proposed system will possess several critical components. One component is sensors [4]. The system will be equipped with sensors like motion, temperature, proximity, and image sensors that will be used to gather data from the surroundings. Another critical system component will be a data analytics and tamper response component [4]. The particular component will receive data from sensors, analyze the data, and respond to tamper. The data analytics and tamper response components have machine learning capabilities [4]. Machine learning is a type of artificial intelligence (AI) that enables systems to predict outcomes even when not programmed to predict such outcomes [8]. Machine learning algorithm analyzes historical data using different techniques and uses the results of such analysis to predict outcomes [8]. A machine-learning-based anti-tamper system will be capable of detecting and reporting many incidents of concern. Other system components include a network interface, power supply, and external memory [4].

Fig 2. Intelligent Anti-Tamper System Layout [4]

According to figure 2 above, the data analytics and tamper response components integrate all of the other components of the system. Sensors input data into the component. The network interface helps facilitate data and information communication into and out of the system. The backup battery is aimed at ensuring that the system continues running in the event of power disruption. The external memory would hold data and information currently being processed by the data analytics and tamper response component [4]. The crystal or external clock input provides key data into the data analytics and tamper response component. Almost all of the system’s components would be protected by a tamper-protected enclosure. The enclosure would protect the system’s key components from being easily disabled by intruders. The intelligent anti-tamper solution can be scaled to protect a large installation housing critical ICT infrastructure.


The video analytics physical security solution and the intelligent anti-tamper solution share two key similarities. One similarity is that both technologies use AI to achieve their intended goals. Video analytics uses neural networks, a type of AI [5]. The intelligent anti-tamper is designed to use machine learning, also a type of AI [4]. The second similarity between the two systems is that they have largely automated the work of detecting security threats. Both systems can scan the surroundings and determine if an object or people pose a threat to the infrastructure that the systems are employed to protect. Despite the two key similarities, the systems share several differences.

TABLE 1: Differences between the Two Solutions

No. Video Analytics Solution Intelligent Anti-Tamper Solution
1. Requires some degree of human control Can operate autonomously
2. Designed to detect drones and humans Designed to detect a large number of threats
3. Has very few sensors Designed to use many sensors
4. A different system for each threat A single system for all threats


Table 1 above lists the differences between the video analytics and the intelligent anti-tamper solution. One key difference between the two physical security solutions is that the video analytics solution requires some level of human control while the other solution can operate without human supervision. An individual is needed to fly the drone integrated with the video analytics solutions [5]. However, the intelligent anti-tamper solutions’ sensors can collect data autonomously and send the data analytics and tamper response component, which can determine if there is a threat to physical infrastructure or otherwise [4]. The second difference between the two is that the video analytics system is limited to detecting drones and humans. However, the intelligent anti-tamper solution is designed to detect numerous threats. The solution can detect threats like intruders, fire, water, and cooling system failure, among others [4].

The third difference is that the video analytics system has few sensors. The particular system only has sensors that can detect drones or human intruders [5]. However, the intelligent anti-tamper solution can be fitted with as many sensors as possible [4]. The ability to integrate many sensors into the intelligent anti-tamper system is what makes it capable of detecting numerous threats. The fourth difference is that a particular video analytics system handles one threat while a single intelligent anti-tamper system can handle multiple threats simultaneously. There is a single video analytics system for detecting drones and another for detecting intruders [5].

After comparing the two Intelligent Physical Security Systems, it is clear that the intelligent anti-tamper solution is the best solution for automating physical security surveillance. The system can detect a wide variety of threats without human supervision. It is possible to further integrate drone detection capabilities into the intelligent anti-tamper system. The system would provide more value for money to its customers. Organizations should use the system in conjunction with their human security personnel. Such use would help detect threats before they arise. Thus helping prevent disasters. Using the intelligent anti-tamper system in conjunction with security personnel would help quickly detect disasters and respond to the same. Thus reducing damage to a lot of critical ICT infrastructure.

  • Zhang, L., Fang, X., Chen, Y., Song, Y., & Qian, S. (2021, March). Research on business continuity rating model in cloud environment. In 2021 IEEE 5th Advanced Information Technology, Electronic and Automation Control Conference (IAEAC)(Vol. 5, pp. 1702-1705). IEEE. https://doi.org/10.1109/IAEAC50856.2021.9390871
  • Yang, X., Shu, L., Liu, Y., Hancke, G. P., Ferrag, M. A., & Huang, K. (2022). Physical Security and Safety of IoT Equipment: A Survey of Recent Advances and Opportunities. IEEE Transactions on Industrial Informatics18(7), 4319-4330. https://doi.org/10.1109/TII.2022.3141408
  • Sun, X., Chen, J., Zhao, H., Zhang, W., & Zhang, Y. (2022). Sequential Disaster Recovery Strategy for Resilient Distribution Network Based on Cyber-Physical Collaborative Optimization. IEEE Transactions on Smart Grid. https://doi.org/10.1109/TSG.2022.3198696
  • Halak, B., Hall, C., Fathir, S., Kit, N., Raymonde, R., Gimson, M., … & Vincent, H. (2022). Towards Autonomous Physical Security Defenses using Machine Learning. IEEE Access. https://doi.org/10.1109/ACCESS.2022.3175615
  • Zhang, X., Chandramouli, K., Gabrijelcic, D., Zahariadis, T., & Giunta, G. (2020, July). Physical security detectors for critical infrastructures against new-age threat of drones and human intrusion. In 2020 IEEE International Conference on Multimedia & Expo Workshops (ICMEW)(pp. 1-4). IEEE. https://doi.org/10.1109/ICMEW46912.2020.9106043
  • Luňáček, O. (2021, June). Analysis of UAV usage in the physical security area. In 2021 International Conference on Military Technologies (ICMT) (pp. 1-5). IEEE. https://doi.org/10.1109/ICMT52455.2021.9502813
  • Djavanshir, G. R., Chen, X., & Yang, W. (2021). A Review of Artificial Intelligence’s Neural Networks (Deep Learning) Applications in Medical Diagnosis and Prediction. IT Professional23(3), 58-62. https://doi.org/10.1109/MITP.2021.3073665
  • Chellappa, R., Theodoridis, S., & van Schaik, A. (2021). Advances in machine learning and deep neural networks. Proceedings of the IEEE109(5), 607-611. https://doi.org/10.1109/JPROC.2021.3072172

Have you ever wanted to learn how to write academic articles for publishing or end of semester term papers? This Astute Scholars will walk you step by step through the process of carrying out research and writing quality articles.