Biblio
Deep learning techniques have demonstrated the ability to perform a variety of object recognition tasks using visible imager data; however, deep learning has not been implemented as a means to autonomously detect and assess targets of interest in a physical security system. We demonstrate the use of transfer learning on a convolutional neural network (CNN) to significantly reduce training time while keeping detection accuracy of physical security relevant targets high. Unlike many detection algorithms employed by video analytics within physical security systems, this method does not rely on temporal data to construct a background scene; targets of interest can halt motion indefinitely and still be detected by the implemented CNN. A key advantage of using deep learning is the ability for a network to improve over time. Periodic retraining can lead to better detection and higher confidence rates. We investigate training data size versus CNN test accuracy using physical security video data. Due to the large number of visible imagers, significant volume of data collected daily, and currently deployed human in the loop ground truth data, physical security systems present a unique environment that is well suited for analysis via CNNs. This could lead to the creation of algorithmic element that reduces human burden and decreases human analyzed nuisance alarms.
Cooperative spectrum sensing is often necessary in cognitive radios systems to localize a transmitter by fusing the measurements from multiple sensing radios. However, revealing spectrum sensing information also generally leaks information about the location of the radio that made those measurements. We propose a protocol for performing cooperative spectrum sensing while preserving the privacy of the sensing radios. In this protocol, radios fuse sensing information through a distributed particle filter based on a tree structure. All sensing information is encrypted using public-key cryptography, and one of the radios serves as an anonymizer, whose role is to break the connection between the sensing radios and the public keys they use. We consider a semi-honest (honest-but-curious) adversary model in which there is at most a single adversary that is internal to the sensing network and complies with the specified protocol but wishes to determine information about the other participants. Under this scenario, an adversary may learn the sensing information of some of the radios, but it does not have any way to tie that information to a particular radio's identity. We test the performance of our proposed distributed, tree-based particle filter using physical measurements of FM broadcast stations.
Establishing a secret and reliable wireless communication is a challenging task that is of paramount importance. In this paper, we investigate the physical layer security of a legitimate transmission link between a user that assists an Intrusion Detection System (IDS) in detecting eavesdropping and jamming attacks in the presence of an adversary that is capable of conducting an eavesdropping or a jamming attack. The user is being faced by a challenge of whether to transmit, thus becoming vulnerable to an eavesdropping or a jamming attack, or to keep silent and consequently his/her transmission will be delayed. The adversary is also facing a challenge of whether to conduct an eavesdropping or a jamming attack that will not get him/her to be detected. We model the interactions between the user and the adversary as a two-state stochastic game. Explicit solutions characterize some properties while highlighting some interesting strategies that are being embraced by the user and the adversary. Results show that our proposed system outperform current systems in terms of communication secrecy.
This paper aims to address the security challenges on physical unclonable functions (PUFs) raised by modeling attacks and denial of service (DoS) attacks. We develop a hardware isolation-based secure architecture extension, namely PUFSec, to protect the target PUF from security compromises without modifying the internal PUF design. PUFSec achieves the security protection by physically isolating the PUF hardware and data from the attack surfaces accessible by the adversaries. Furthermore, we deploy strictly enforced security policies within PUFSec, which authenticate the incoming PUF challenges and prevent attackers from collecting sufficient PUF responses to issue modeling attacks or interfering with the PUF workflow to launch DoS attacks. We implement our PUFSec framework on a Xilinx SoC equipped with ARM processor. Our experimental results on the real hardware prove the enhanced security and the low performance and power overhead brought by PUFSec.
One of the specially designated versatile networks, commonly referred to as MANET, performs on the basics that each and every one grouping in nodes totally operate in self-sorting out limits. In any case, performing in a group capacity maximizes quality and different sources. Mobile ad hoc network is a wireless infrastructureless network. Due to its unique features, various challenges are faced under MANET when the role of routing and its security comes into play. The review has demonstrated that the impact of failures during the information transmission has not been considered in the existing research. The majority of strategies for ad hoc networks just determines the path and transmits the data which prompts to packet drop in case of failures, thus resulting in low dependability. The majority of the existing research has neglected the use of the rejoining processing of the root nodes network. Most of the existing techniques are based on detecting the failures but the use of path re-routing has also been neglected in the existing methods. Here, we have proposed a method of path re-routing for managing the authorized nodes and managing the keys for group in ad hoc environment. Securing Schemes, named as 2ACK and the EGSR schemes have been proposed, which may be truly interacted to most of the routing protocol. The path re-routing has the ability to reduce the ratio of dropped packets. The comparative analysis has clearly shown that the proposed technique outperforms the available techniques in terms of various quality metrics.
This paper proposed a feedback shift register structure which can be split, it is based on a research of operating characteristics about 70 kinds of cryptographic algorithms and the research shows that the “different operations similar structure” reconfigurable design is feasible. Under the configuration information, the proposed structure can implement the multiplication in finite field GF(2n), the multiply/divide linear feedback shift register and other operations. Finally, this paper did a logic synthesis based on 55nm CMOS standard-cell library and the results show that the proposed structure gets a hardware resource saving of nearly 32%, the average power consumption saving of nearly 55% without the critical delay increasing significantly. Therefore, the “different operations similar structure” reconfigurable design is a new design method and the proposed feedback shift register structure can be an important processing unit for coarse-grained reconfigurable cryptologic array.
Reliability and robustness of Internet of Things (IoT)-cloud-based communication is an important issue for prospective development of the IoT concept. In this regard, a robust and unique client-to-cloud communication physical layer is required. Physical Unclonable Function (PUF) is regarded as a suitable physics-based random identification hardware, but suffers from reliability problems. In this paper, we propose novel hardware concepts and furthermore an analysis method in CMOS technology to improve the hardware-based robustness of the generated PUF word from its first point of generation to the last cloud-interfacing point in a client. Moreover, we present a spectral analysis for an inexpensive high-yield implementation in a 65nm generation. We also offer robust monitoring concepts for the PUF-interfacing communication physical layer hardware.
Hardware implementations of cryptographic algorithms may leak information through numerous side channels, which can be used to reveal the secret cryptographic keys, and therefore compromise the security of the algorithm. Power Analysis Attacks (PAAs) [1] exploit the information leakage from the device's power consumption (typically measured on the supply and/or ground pins). Digital circuits consume dynamic switching energy when data propagate through the logic in each new calculation (e.g. new clock cycle). The average power dissipation of a design can be expressed by: Ptot(t) = α · (Pd(t) + Ppvt(t)) (1) where α is the activity factor (the probability that the gate will switch) and depends on the probability distribution of the inputs to the combinatorial logic. This induces a linear relationship between the power and the processed data [2]. Pd is the deterministic power dissipated by the switching of the gate, including any parasitic and intrinsic capacitances, and hence can be evaluated prior to manufacturing. Ppvt is the change in expected power consumption due to nondeterministic parameters such as process variations, mismatch, temperature, etc. In this manuscript, we describe the design of logic gates that induce data-independent (constant) α and Pd.
The current state of the internet relies heavily on SSL/TLS and the certificate authority model. This model has systematic problems, both in its design as well as its implementation. There are problems with certificate revocation, certificate authority governance, breaches, poor security practices, single points of failure and with root stores. This paper begins with a general introduction to SSL/TLS and a description of the role of certificates, certificate authorities and root stores in the current model. This paper will then explore problems with the current model and describe work being done to help mitigate these problems.
Security of control systems have become a new and important field of research since malicious attacks on control systems indeed occurred including Stuxnet in 2011 and north eastern electrical grid black out in 2003. Attacks on sensors and/or actuators of control systems cause malfunction, instability, and even system destruction. The impact of attack may differ by which instrumentation (sensors and/or actuators) is being attacked. In particular, for control systems with multiple sensors, attack on each sensor may have different impact, i.e., attack on some sensors leads to a greater damage to the system than those for other sensors. To investigate this, we consider sensor bias injection attacks in linear control systems equipped with anomaly detector, and quantify the maximum impact of attack on sensors while the attack remains undetected. Then, we introduce a notion of sensor security index for linear dynamic systems to quantify the vulnerability under sensor attacks. Method of reducing system vulnerability is also discussed using the notion of sensor security index.
Side-channel collision attacks have been one of the most powerful attack techniques, combining advantages of traditional side-channel attack and mathematical cryptanalysis. In this paper, we propose a novel multiple-bits side-channel collision attack based on double distance voting detection, which can find all 120 relations among 16 key bytes with only 32 averaged power traces when applied to AES (Advanced Encryption Standard) algorithm. Practical attack experiments are performed successfully on a hardware implementation of AES on FPGA board. Results show that the necessary number of traces for our method is about 50% less than correlation-enhanced collision attack and 76% less than binary voting test with 90% success rate.
Detecting software security vulnerabilities and distinguishing vulnerable from non-vulnerable code is anything but simple. Most of the time, vulnerabilities remain undisclosed until they are exposed, for instance, by an attack during the software operational phase. Software metrics are widely-used indicators of software quality, but the question is whether they can be used to distinguish vulnerable software units from the non-vulnerable ones during development. In this paper, we perform an exploratory study on software metrics, their interdependency, and their relation with security vulnerabilities. We aim at understanding: i) the correlation between software architectural characteristics, represented in the form of software metrics, and the number of vulnerabilities; and ii) which are the most informative and discriminative metrics that allow identifying vulnerable units of code. To achieve these goals, we use, respectively, correlation coefficients and heuristic search techniques. Our analysis is carried out on a dataset that includes software metrics and reported security vulnerabilities, exposed by security attacks, for all functions, classes, and files of five widely used projects. Results show: i) a strong correlation between several project-level metrics and the number of vulnerabilities, ii) the possibility of using a group of metrics, at both file and function levels, to distinguish vulnerable and non-vulnerable code with a high level of accuracy.
The modular multilevel converter with series and parallel connectivity was shown to provide advantages in several industrial applications. Its reliability largely depends on the absence of failures in the power semiconductors. We propose and analyze a fault-diagnosis technique to identify shorted switches based on features generated through wavelet transform of the converter output and subsequent classification in support vector machines. The multi-class support vector machine is trained with multiple recordings of the output of each fault condition as well as the converter under normal operation. Simulation results reveal that the proposed method has high classification latency and high robustness. Except for the monitoring of the output, which is required for the converter control in any case, this method does not require additional module sensors.
This paper presents a true random number generator that exploits the subthreshold properties of jitter of events propagating in a self-timed ring and jitter of events propagating in an inverter based ring oscillator. Design was implemented in 180nm CMOS flash process. Devices provide high quality random bit sequences passing FIPS 140-2 and NIST SP 800-22 statistical tests which guaranty uniform distribution and unpredictability thanks to the physics based entropy source.
The Department of Homeland Security Cyber Security Division (CSD) chose Moving Target Defense as one of the fourteen primary Technical Topic Areas pertinent to securing federal networks and the larger Internet. Moving Target Defense over IPv6 (MT6D) employs an obscuration technique offering keyed access to hosts at a network level without altering existing network infrastructure. This is accomplished through cryptographic dynamic addressing, whereby a new network address is bound to an interface every few seconds in a coordinated manner. The goal of this research is to produce a Register Transfer Level (RTL) network security processor implementation to enable the production of an Application Specific Integrated Circuit (ASIC) variant of MT6D processor for wide deployment. RTL development is challenging in that it must provide system level functions that are normally provided by the Operating System's kernel and supported libraries. This paper presents the architectural design of a hardware engine for MT6D (HE-MT6D) and is complete in simulation. Unique contributions are an inline stream-based network packet processor with a Complex Instruction Set Computer (CISC) architecture, Network Time Protocol listener, and theoretical increased performance over previous software implementations.
Use of digital token - which certifies the bearer's rights to some kind of products or services - is quite common nowadays for its convenience, ease of use and cost-effectiveness. Many of such digital tokens, however, are produced with software alone, making them vulnerable to forgery, including alteration and duplication. For a more secure safeguard for both token owner's right and service provider's accountability, digital tokens should be tamper-resistant as much as possible in order for them to withstand physical attacks as well. In this paper, we present a rights management system that leverages tamper-resistant digital tokens created by hardware-software collaboration in our eTRON architecture. The system features the complete life cycle of a digital token from generation to storage and redemption. Additionally, it provides a secure mechanism for transfer of rights in a peer-to-peer manner over the Internet. The proposed system specifies protocols for permissible manipulation on digital tokens, and subsequently provides a set of APIs for seamless application development. Access privileges to the tokens are strictly defined and state-of-the-art asymmetric cryptography is used for ensuring their confidentiality. Apart from the digital tokens being physically tamper-resistant, the protocols involved in the system are proven to be secure against attacks. Furthermore, an authentication mechanism is implemented that invariably precedes any operation involving the digital token in question. The proposed system presents clear security gains compared to existing systems that do not take tamper-resistance into account, and schemes that use symmetric key cryptography.
Security Evaluation and Management (SEM) is considerably important process to protect the Embedded System (ES) from various kinds of security's exploits. In general, SEM's processes have some challenges, which limited its efficiency. Some of these challenges are system-based challenges like the hetero-geneity among system's components and system's size. Some other challenges are expert-based challenges like mis-evaluation possibility and experts non-continuous availability. Many of these challenges were addressed by the Multi Metric (MM) framework, which depends on experts' or subjective evaluation for basic evaluations. Despite of its productivity, subjective evaluation has some drawbacks (e.g. expert misevaluation) foster the need for considering objective evaluations in the MM framework. In addition, the MM framework is system centric framework, thus, by modelling complex and huge system using the MM framework a guide is needed indicating changes toward desirable security's requirements. This paper proposes extensions for the MM framework consider the usage of objective evaluations and work as guide for needed changes to satisfy desirable security requirements.
We have proposed a method of designing embedded clock-cycle-sensitive Hardware Trojans (HTs) to manipulate finite state machine (FSM). By using pipeline to choose and customize critical path, the Trojans can facilitate a series of attack and need no redundant circuits. One cannot detect any malicious architecture through logic analysis because the proposed circuitry is the part of FSM. Furthermore, this kind of HTs alerts the trusted systems designers to the importance of clock tree structure. The attackers may utilize modified clock to bypass certain security model or change the circuit behavior.
This paper presents a wireless intrusion prevention tool for distributed denial of service attacks DDoS. This tool, called Wireless Distributed IPS WIDIP, uses a different collection of data to identify attackers from inside a private network. WIDIP blocks attackers and also propagates its information to other wireless routers that run the IPS. This communication behavior provides higher fault tolerance and stops attacks from different network endpoints. WIDIP also block network attackers at its first hop and thus reduce the malicious traffic near its source. Comparative tests of WIDIP with other two tools demonstrated that our tool reduce the delay of target response after attacks in application servers by 11%. In addition to reducing response time, WIDIP comparatively reduces the number of control messages on the network when compared to IREMAC.
Gatys et al. recently introduced a neural algorithm that renders a content image in the style of another image, achieving so-called style transfer. However, their framework requires a slow iterative optimization process, which limits its practical application. Fast approximations with feed-forward neural networks have been proposed to speed up neural style transfer. Unfortunately, the speed improvement comes at a cost: the network is usually tied to a fixed set of styles and cannot adapt to arbitrary new styles. In this paper, we present a simple yet effective approach that for the first time enables arbitrary style transfer in real-time. At the heart of our method is a novel adaptive instance normalization (AdaIN) layer that aligns the mean and variance of the content features with those of the style features. Our method achieves speed comparable to the fastest existing approach, without the restriction to a pre-defined set of styles. In addition, our approach allows flexible user controls such as content-style trade-off, style interpolation, color & spatial controls, all using a single feed-forward neural network.
Recent progress in style transfer on images has focused on improving the quality of stylized images and speed of methods. However, real-time methods are highly unstable resulting in visible flickering when applied to videos. In this work we characterize the instability of these methods by examining the solution set of the style transfer objective. We show that the trace of the Gram matrix representing style is inversely related to the stability of the method. Then, we present a recurrent convolutional network for real-time video style transfer which incorporates a temporal consistency loss and overcomes the instability of prior methods. Our networks can be applied at any resolution, do not require optical flow at test time, and produce high quality, temporally consistent stylized videos in real-time.