Raushan Kumar Singh (IIT Ropar), Sudeepta Mishra (IIT Ropar)

Modern technology is advancing on many different levels, and the battlefield is no exception. India has 15000 km of lengthy land borders shared with many other neighboring countries, and only 5 of the 29 states in India do not have any shared international borders or coastlines. Wire fences and conventional sensor-based systems are used to protect terrestrial borders. Wire fences, being the only line of defense against intrusions at most unmanned borders, result in frequent cases of unreported incursion, smuggling, and human trafficking. Typically, intruders cut the fence to gain access to Indian land, and sensor-based systems are prone to false alarms due to animal movements. We propose combining the intelligence of Tiny Machine Learning (TinyML) with the communication capability of IoT to make borders safer and intrusion more challenging. To learn the typical fence movements from natural causes, we use TinyML. Our learning technique is created explicitly to differentiate between regular fence movement and suspicious fence disturbance. The system is efficient enough to detect metal fence cuts and trespassing carefully. With the aid of online learning environments, the sophisticated TinyML microcontroller’s built-in accelerometer can differentiate between different movement patterns. To identify the most effective defense against sensor-level attacks, we conducted tests to gauge the tolerance levels of conventional microcontroller sensor systems against TinyML-powered microcontrollers when exposed to Electromagnetic Pulse (EMP) based sensor hacking attempts. To the best of our knowledge, this is the first research conducted for the Identification of the best suite sensor system for high-precision Internet of Battlefield Things (IoBT) applications. During the real-time model test, the system is found to be 95.4% accurate and readily deployable on TinyML microcontrollers.

View More Papers

Phoenix: Surviving Unpatched Vulnerabilities via Accurate and Efficient Filtering...

Hugo Kermabon-Bobinnec (Concordia University), Yosr Jarraya (Ericsson Security Research), Lingyu Wang (Concordia University), Suryadipta Majumdar (Concordia University), Makan Pourzandi (Ericsson Security Research)

Read More

Differentially Private Dataset Condensation

Tianhang Zheng (University of Missouri-Kansas City), Baochun Li (University of Toronto)

Read More

Low-Quality Training Data Only? A Robust Framework for Detecting...

Yuqi Qing (Tsinghua University), Qilei Yin (Zhongguancun Laboratory), Xinhao Deng (Tsinghua University), Yihao Chen (Tsinghua University), Zhuotao Liu (Tsinghua University), Kun Sun (George Mason University), Ke Xu (Tsinghua University), Jia Zhang (Tsinghua University), Qi Li (Tsinghua University)

Read More

FreqFed: A Frequency Analysis-Based Approach for Mitigating Poisoning Attacks...

Hossein Fereidooni (Technical University of Darmstadt), Alessandro Pegoraro (Technical University of Darmstadt), Phillip Rieger (Technical University of Darmstadt), Alexandra Dmitrienko (University of Wuerzburg), Ahmad-Reza Sadeghi (Technical University of Darmstadt)

Read More