Low Power Consumption Data Security Implementation for Artificial Intelligence Based Smart Systems


Creative Commons License

Mumcu M. C., İkinci B., Öztürk E., Ataş K., Güngör A.

Engineering Science and Technology, Durres, Arnavutluk, 4 - 06 Eylül 2024, ss.3, (Özet Bildiri)

  • Yayın Türü: Bildiri / Özet Bildiri
  • Basıldığı Şehir: Durres
  • Basıldığı Ülke: Arnavutluk
  • Sayfa Sayıları: ss.3
  • Açık Arşiv Koleksiyonu: AVESİS Açık Erişim Koleksiyonu
  • İstanbul Gelişim Üniversitesi Adresli: Evet

Özet

Power consumption optimisation is becoming more crucial in the design of artificialintelligence (AI)-powered smart systems. Data security is becoming a major issue in an era where the useof these systems is growing, particularly in applications that demand low power consumption. Due totheir high sensitivity and complexity, artificial intelligence applications typically handle massive volumesof data, making data security crucial. Nonetheless, it seems that low-power AI systems are pushing theboundaries of conventional security techniques and algorithms. In this context, researchers are workingon realistic approaches for ensuring data security and privacy in artificial intelligence-based systems thatrequire less power. One strategy for improving energy efficiency and ensuring data security is to employlightweight cryptography (LWC) algorithms, which are suitable for microprocessors with limited memoryand power in various architectures used in smart system design. Lightweight cryptography techniques onthe Internet of Things (IoT) platform allow a wide range of low-resource and resource-constraineddevices to compute, make decisions, and communicate securely across networks. Numerous advantageshave been achieved through the design and optimization of various structured algorithms. This studyinvestigated the integration and optimization of lightweight cryptography algorithms with a focus on datasecurity in the application field, as well as the wide application field of artificial intelligence-basedsystems that require low power consumption. Furthermore, a RISC microprocessor was used to modelimage processing, and data security and confidentiality were evaluated by integrating an algorithm chosenfrom a list of lightweight cryptography algorithms that complied with the model. Performance metricssuch as memory usage area and power consumption were identified, and the analysis's results arepresented.