Friday, March 20, 2020

URSP Student Mitch Martinez Works to Research and Develop a Cyber Reconnaissance Working Dog System

Working dogs often play a niche role in various military operations and investigations. They are are historically recognized for their unique sensory and search capabilities in which humans have vitally depended on. Growing up, working dogs have always been an integral part of my life. My dad was a former explosives canine handler and division specialist. In Fall 2019, I was looking for a course project that I could use to simultaneously further my experience outside of the classroom. So, I decided to look at what I already knew. I discovered that non-line-of-sight (NLOS) canine control was well sought after by handlers and that little research had been conducted involving working dogs in the cyber domain. I initiated a plan of research and began reaching out to canine trainers and relevant researchers. I asked my professor, Dr. Winston, to mentor a project that bridges the gap between working dogs and cybersecurity.

Despite recent developments in artificial intelligence and an increased emphasis on robotics, nothing compares to the to the portability, agility, and trainability of working dogs in mission environments.  By developing a mobile low-power signal and packet gathering sniffer, harness-wearing working dogs would be able to directly contribute to passive reconnaissance operations as the delivery device to areas of interest.  Once the dog reaches the target location and hides, operatives would then be able to remotely execute probing commands and automated scripts utilizing modern hacking software and log analysis tools. However, long distance off leash handler-to-canine communication remains a challenge. 

Nevertheless, based on current canine training practices, this project aiming to solve the dilemma by integrating established lidar, radar, and GPS technologies coupled with wireless signal capturing capabilities. Following a ‘just enough data’ paradigm, fault tolerant NLOS communication between handler and canine may be achieved

By utilizing a smartphone, microcomputer, and software-defined radio, remote communication via audio frequencies and harness vibrations may be established over a peer-to-peer LTE network supported by machine learning detection algorithms and signal engineering techniques. The development of this technology would provide governments and agencies a niche risk averse alternative to unmanned-aerial-vehicles and hardware dead drops. The intended cyber psychical system use-cases are for discrete night operations where human-threatening boundaries are present. Working dogs may be the most reliable and non-invasive weapon for delivering cyber reconnaissance tools in these scenarios.