New chaos-based algorithm improves surveillance network
Thu 9 Feb 2017
Researchers have created a new algorithm that successfully prolongs the battery life and efficiency of video sensor surveillance networks using a chaos-based approach.
Testing shows that using a chaos-based algorithm to define sleep and active mode changes in the surveillance system can increase the whole network lifetime, improve network efficiency, and help the system to withstand malicious attacks from smart intruders with knowledge of the network.
Video sensor networks are a cost-effective solution for physical surveillance, offering flexibility and improved monitoring of obstacle-rich environments over traditional fixed-camera systems. However, video sensor systems are limited by weak computational and battery power. To date, this problem has been addressed by taking advantage of system redundancies, and setting sensors to sleep mode when coverage is not required.
However, an intelligent or knowledgeable intruder may detect the pattern in sensor modes and thereby have a better chance of evading detection by the security system.
Creating an algorithm that switches parts of the network from active to sleep mode on an unpredictable schedule is intended to increase the efficiency of the network as a whole and extend battery life of sensors, while at the same time improving the effectiveness of the system to detect intruders.
To that end, researchers from the University of Franche-Comte created an algorithm using a chaos-based approach, where every node in a video sensor system is changed from active to sleep mode based on last detection, combined with a hash value and a random number to preserve the unpredictability of the system.
In testing simulations, the team was able to prove that the lifetime of the network was extended without losing any effectiveness in intruder detection. Simulations were conducted based on two types of events: the first, where a random party attempts to evade the security system without any knowledge of the system other than that a surveillance system exists; and the second, where a malicious intruder attempts to evade the security system with knowledge and foresight using applied observations.
The researchers found that the algorithm successfully protected against both types of intruders. In fact, they found that using the new algorithm increased the system’s ability to withstand malicious attacks due to the unpredictability of the system. Because each sensor is set to sleep and active modes based on a chaotic algorithm, even an informed intruder is unable to predict which nodes are active at a certain time.