A new generation of robots is coming: more flexible, more ‘humanoid’. A research team at the University of Waterloo in Canada has taken a crucial step in that direction with the creation of a super-strong elastomeric material that paves the way for soft robotics. “This study examines the privacy risks in collaborative robotics, focusing on the potential for traffic analysis in encrypted robotic communications,” the scientists explain in the research publication.
“We evaluated the effectiveness of prominent website fingerprinting techniques (e.g., TikTok, RF)”
Robotics known as ‘soft’ robotics has significant implications in medicine, for example. Hospitals employ them as surgical assistants due to their ability to perform fine movements without fatigue, and in factories they have become valuable allies for repetitive or hazardous tasks. Scientists Cheng Tang, Diogo Barradas, Urs Hengartner, and Yue Hu explain how the research began: “We evaluated the effectiveness of prominent website fingerprinting techniques (e.g., TikTok, RF) and their limitations in accurately identifying robotic actions due to their inability to capture detailed temporal relationships.”
The new material has proven to be nine times stronger and more flexible than its predecessors.
In the study ‘On the Feasibility of Fingerprinting Collaborative Robot Network Traffic’, experts explain in detail how the idea arose and how it was developed. “We present a traffic classification approach using signal processing techniques, which demonstrates high accuracy in identifying actions and highlights the vulnerability of encrypted communications to privacy breaches,” the experts explain. The new material has proven to be nine times stronger and more flexible than its predecessors. According to the researchers, its fibers can lift loads 2,000 times their own weight and perform work equivalent to about 24 kilograms (53 pounds). “Our work is based on a closed-world environment, where, although the set of monitored actions is limited, it is expected to be sufficient for an adversary to infer a robot’s activity in specific contexts,” they comment.
The result is a composite that retains its flexibility but offers significantly greater tensile strength than traditional materials used to date
The study combines liquid crystals with liquid crystal elastomers (LCEs), a type of polymer that expands and contracts with heat. The experiment was conducted using a Kinova Gen3 robotic arm, a lightweight robot commonly used in research settings. The result is a composite that retains its flexibility but offers significantly greater tensile strength than traditional materials used to date. The controller executed pre-programmed commands, and communications were protected with TLS encryption. The scientists added specific details: “This workstation runs Ubuntu 20.04 LTS and is configured with a 2.40 GHz Intel Core i7-8700T CPU and 32 GB of RAM.”
“The interaction between the controller and the Kinova arm generates network traffic patterns that can inadvertently reveal operational details”
Privacy is a key issue for scientists, as they themselves explain: “The interaction between the controller and the Kinova arm generates network traffic patterns that can inadvertently reveal operational details,” and they add: “Our findings underscore the need for continued development of practical defenses for robotic privacy and security.” On the other hand, in most tests the system was able to identify the robot’s actions with an accuracy rate of nearly 97%.
This could represent one of the major breakthroughs in recent years
In conclusion, regarding the issue of privacy, the scientists state that “our findings highlight potential vulnerabilities in robotic communication channels, which necessitates the development of enhanced security measures as the deployment of collaborative robots becomes more widespread.” For other sectors, this could represent one of the major breakthroughs in recent years. “While we focus on collaborative robots, our method and findings are likely to be applicable to industrial robotic arms, such as those used in manufacturing,” they conclude from Waterloo.
