CapriTech was established to undertake both theoretic and applied research, shape the development of complex software, and disseminate knowledge by working hand-in-hand with European large industries, SMEs, research centres and universities. CapriTech's executive team members have a rich track record in developing software, publishing their research in high impact factor venues, attracting research and innovation fund and participating successfully in research and innovation projects. CapriTech's focus areas are:
Cybersecurity is increasingly being recognized as one of the major issues in many industries,
and robotics is certainly not an exception. From robots used on assembly lines, to those used in medical settings,
all the way to assistive and entertainment robots, used in our households,
it is clear that safety, security and privacy of robotic systems is critical.
CapriTech's researchers have developed a specialized software for the detection and
classification of cyber and cyber-physical attacks against robotic systems. CapriTech (henceforth abbreviated
by CPT) participates in ROBORDER
(H2020 SEC-20-BES-2016 | GA no 740593), an EU-funded project which develops a fully-functional autonomous
border surveillance system with unmanned mobile robots including aerial, water surface, underwater
and ground vehicles, capable of functioning both as standalone and in swarms,
which will incorporate multimodal sensors as part of an interoperable network.
For more details visit the projects webpage.
As the 5G era is drawing near, security and privacy are cornerstones in transforming
5G systems in a platform for the Networked. CapriTech by realizing this need, develops specialized tools
aimed at detecting and identifying the new risks and security threats that the 5G networks will be facing.
CapriTech participates in 5G ESSENCE (H2020-ICT-07-2017 | GA no 761592), an EU-funded project which addresses the paradigms
of Edge Cloud computing and Small Cell as a Service by fuelling the drivers and removing the barriers
in the Small Cell market, forecasted to grow at an impressive pace up to 2020 and beyond and to
play a key role in the 5G ecosystem.
For more details visit the projects webpage.
Incident Detection and Classification
Incident detection mechanisms, part of the intrusion detection process, fall into two categories, the signature-based and the anomaly-based. Signature-based detection mechanisms can identify a specific predefined behaviour that was marked as potentially dangerous, such as a transmission of a string pattern or a succession of predefined events. These mechanisms, however, are effective only for known threats. The anomaly-based detection methods compare the activity that is considered normal against the observed one. In this case, machine learning techniques are used in order to train the system to identify potential breaches. CPT's researchers have prior experience in developing software for incident detection and classification. CPT's research has led to the development of the FINDER (Finding INtruDERs using advanced machine learning) software.
Game Theory for Cybersecurity
Cybersecurity is a complex and challenging problem. Game theoretical methods offer new insights into quantitative evaluation of this problem. Taking advantage of these methods, CapriTech develops game theoretic frameworks to enable the precise analysis of the interactions between the Defender (i.e., the network) and the cyber Attacker providing “robust” and mathematically optimal recommendation strategies for the Defender. As the advanced and evolving nature of the cyber attacks leads to high degree of uncertainty on the information collected on one, the direction CapriTech follows is on Bayesian game theory, which is able to account for incomplete information known to each player and beliefs with probability distributions.
Security and Privacy for the Internet of Things (IoT)
CPT specialises in decision-making methodologies for cyber security and privacy in the Internet of Things (IoT). These two IoT research domains are relevant to all industries, as threats and breaches endanger a vast range of devices, applications, and services. In general there are different adversarial models that threaten IoT networks (both cloud-based and decentralised). The adversary might aim to compromise a device in order to get access to private data (e.g. Personally Identifiable Information), or control a device (e.g. mobile device, sensor, router, camera, printer). IoT communications can also be targeted to either intercept sent data, or to even impersonate the device (MITM) over a communication path. According to Tripwire Survey (Feb. 2014), 80% of Amazon's Top 25 best-selling SOHO wireless access router models have security vulnerabilities. As these routers can be part of any IoT network, the cyber security and privacy risks associated, must be mitigated, prior to the network establishment. Apart from this, baseline policy devices, applications, and services must be frequently updated with the latest software to ensure that vulnerabilities are patched. To address the above, best practice cyber security and privacy mechanisms and policies must be deployed. Lastly, given the limited battery and processing power of IoT devices, there is a need to balance efficiency with cyber security and privacy. CPT's researchers also specialise in creating models, comparing aspects of differing IoT network implementations, and propose optimal solutions with respect to users' and service providers' security and privacy requirements.
Big Data Privacy
After recent Edward Snowden's revelations, user privacy has been one of the most discussed topics around the world. For example, given that the majority of users share some personal data with cloud providers there is a need for identifying the most trusted, in terms of privacy preservation that at the same time suit their non-functional requirements (e.g. financial budget). Research at CPT focuses on identifying privacy requirements and determine an optimal set of Privacy Enhancing Technologies that satisfy these requirements in an optimal way. CPT researchers are also investigating privacy-preserving mechanisms to ensure data privacy when querying health databases or sharing anonymized patients' records. Techniques that are studied and extended include k-anonymity to protect the patient's identifiable information, and ε-differential privacy when querying databases. CPT's research aims to minimize the chances of patients' records being identified when applying data mining and analytics.
Cyber Security Risk Management
Having to deal with a vast number of cyber attacks, as reported by the Verizon 2015 Data Breach Investigations Report, there is an urgent requirement for minimization of risks by optimally allocating cyber security resources (e.g. manpower). In addition to that, there is an increasingly important need on determining optimal cyber security investment strategies. Many organisations are asking the question "how much do we need to spend on cyber security in order to minimize risks?". Previous research has shown that although some companies spend less than others, they are more secure. Therefore an apparent question is "do we spend our cyber security budgets correctly?" The first steps towards answering these and other relevant questions are (i) to define security and privacy cost models considering both direct costs (e.g. financial loss) and indirect costs (e.g. morale costs), (ii) to price digital assets, (iii) to undertake an assessment of risks and impact of the most prominent cyber attacks, and (iv) to create the right metrics for profiling different organisations based on their cyber security posture. The UK has published a set of guidelines that organisations, similar to the one in the case study, should comply with in order to reduce the risk of damage from cyber attacks. The document called Cyber Essentials ( here) suggests a number of basic controls that organisations should implement to protect themselves from cyber attacks. CPT's researchers have prior experience in creating innovative models and developing software to support decision-making within the field of cyber security investments. These models can be replicated by any company seeking to make such an investment and they provide the same advice with the one advocated by the UK government, in a formal and quantitative manner, with regard to the requirements for basic technical protection from cyber attacks in SMEs. CPT's research in these fields is based upon game-theoretic and multi-objective optimisation techniques in combination coupled with thorough risk assessment methodologies.