SeCIF Research Objectives

General context

Secure communication in the context of the industry of the future (“IoF”, or so called Industry 4.0 in the German terminology) calls for the design of transmission mechanisms that meet new prescribed trade-offs between high efficiency, high reliability, and very low latency on a large scale basis.

Beyond the problem of broadband connectivity to the Internet for human-controlled terminals, the issues posed by machine connectivity must be addressed, where machines in the context of the industry 4.0 range from low-capability sensors, to cars, to intelligent robots such as autonomous UAVs, much akin to those used in the context of IoT.

In parallel to the communications challenges, the challenge posed by the security threats in the IoT and IoF contexts is crucial. This goes both in terms of securing privacy-preserving communication links as protecting the radio devices themselves from compromising attacks

Therefore the questions as to “how to achieve the high performance communications needs of future industrial IoT” as well as “Are communication devices protected enough from external attacks” to “how can IoT radio devices be made more secure” lie at the heart of this project.

Efficient communications in the Internet of Things (IoT) context

Machine learning approaches for communication systems

SeCIF investigates the problem of decentralized IoT device coordination as well as drone-aided broadband communications. The objective is to demonstrate the challenges associated with and the feasibility of robot-aided broadband communications, where a flying robot serves as relay to serve out-of-range (industrial) IoT devices. Machine learning tools can be utilized to infer information about optimal drone-relay position based on random initial measurements.

We are also interested in using Q-learning methods to identify an optimal path for the drone in a model-free scenario. The use of machine learning tools is also envisioned in SeCIF for non UAV scenarios such as channel estimation and MIMO precoding etc.

Joint sensing and Communications

Communication systems often benefit from sensed information which can be exploited in order to adapt the communication protocol to a given environment. For instance in the context of car-to-car communications it was recently argued that radar-type information, revealing the distance and velocity of a car in front can be used to set up the communication protocol. Or in a UAV aided networks it is beneficial from a network operation point of view to estimate the UAV localization at all time, which can in principle be achieved by a radar. Such radar systems operate on resources which are orthogonal to these used for payload transmission. In SeCIF a more efficient approach is undertaken in which sensing and communication resources are jointly exploited to provide simultaneous sensing and communication capabilities.

Security in the Internet of Things (IoT) context

The team of security experts under the SeCIF project aim to leverage their expertise within the general domain of device security so as to tackle this problem in the more specific context of IoT. The team focuses on the analysis of already existing and already deployed devices. While the internet of the future needs to be secure, billions of devices have already been deployed and will form the core of the infrastructure for several years to come. Building the industrial internet of the future also means we need to know how the current IoT networks are deployed and how to monitor their security. We need to learn from the security problems of the current systems to build the ground for secure, future systems. Furthermore, the systems which are already deployed will co-exist with the future Industrial IoT systems and overlooking their security and deployments problems opens the risk for jeopardizing the security of the whole ecosystems. To be effective such measurements needs to be performed at scale. To this end, we will continue the TUM-EURECOM collaboration on binary analysis of embedded systems firmware images and extend it to the discovery and measurements of real world deployments. Many such devices use standardized wireless protocols such as IEEE 802.15.4. However, effectively discovering such devices is currently an open issue. This project therefore aims to contribute to the discovery of real world devices, their identification and the evaluation of the security of such deployments.