Domain-Agnostic Representation of Side-Channels
Abstract
:1. Introduction
- To collate and systematically review the literature demonstrating SCS across the MDM, HCI, Misc, and CYB categories in order to derive their shared fundamental principles.
- To capture each domain’s approach to SCS under a standardised set of terms that encompass all components of the SCS process, from side-channel leakage to target information acquisition, via extraction techniques. Such defined terms must be applicable for adequately describing SCS across all domains, i.e., be domain-agnostic.
- To formally establish that utilisation of SCS exists within HCI and Misc domains, albeit implicitly.
- Establishment of a domain-agnostic SCS framework that provides unified definitions of the SCS process and demonstrate its applicability to the MDM, HCI, Misc, and CYB domains.
- Discussion of enabled opportunities: the cross- and intra-domain transferability of SCS extraction techniques, and avenues for SCS data structure representation for side-channel leakage detection and side-channel discovery methodologies.
2. SCS Across Domains
2.1. SCS Framework
2.1.1. Target Information
- CYB
- focuses on target information within electronic-based computational devices. These may be secured cryptographically [15,16,17,18,19], but can also include personal information that is leaked unintentionally, such as media viewing [20], web-browsing habits [21,22,23], and information communicated or generated by devices such as printers [24,25,26]. In rare cases, non-electronic target information is included within the CYB-related literature, such as PIN codes stored in a person’s brain being extracted through analysis of EEG readings [6,7].
- MDM
- quantifies physiological or physical parameters within the human body (MDM’s target system). Examples include heart rate [27,28,29,30,31], breathing rate [27,32], chemical concentrations (e.g., oxygen) within the bloodstream [33,34], and tremors [35,36]. Conditions with associated physiological manifestations (including psychological conditions with physiological biomarkers, such as cognitive load via pupillary response [37]) also have the potential to be accessible via side channels.
- HCI
- infers the state or intention of humans (i.e., the user) expressed through physiological parameters. This includes gestures initiated by the eyes [38], jaw [39,40,41], tongue [39], hands [13,42], and fingers [43,44]. Such gestures (e.g., moving the eyes from left to right [38] or scratching a surface in an ‘X’ formation [42]) are sensed (e.g., by a smartphone) and the user’s intention is inferred. Unlike other domains, information is part of a feedback loop in which the response from the device loops back to the human, who can then adapt their actions to manipulate the sensing process. For example, touching the flashlight to reduce the light level detected by the camera to infer pressure [43].
- Misc
- is diverse in its target information; it is no longer just a read-only quantity, but can be injected for communication [45], control [46] or to support other forms of sensing [47,48]. The diversity of application areas allows for categorisation of classes of target information:
- Target information can be digital information (CYB), a signal associated with some physical process (MDM), an abstraction such as the representation of a state (CYB, Misc), or an intention to perform a particular action (HCI). Opportunities exist to extract target information in any of these forms. In certain scenarios, partial recovery of target information may still provide value, perhaps by revealing insight into a particular execution path (CYB) or enabling a diagnosis (MDM).
2.1.2. Target Systems
- CYB
- typically ‘attacks’ physical and electronics-based systems such as smart cards [19,59], FPGAs, microcontrollers, embedded devices, PCs [16,17,60], medical devices [61], smartphones [4,14,62], printers (traditional and 3D) [24,25], TVs [20], displays [6,7], and keyboards [63]. Exploitation of specific sensors rather an entire system can also be a side-channel source through signal transformations and injection of signals/values into the signal measured by the sensor [5]. These devices share fundamental exploitable properties such as power consumption, clock rates, acoustic noise generation, and electromagnetic radiation. Consequently, all systems that share a demonstrated exploitable property are vulnerable to SCS [2].
- MDM
- focuses on the human body: an inherently complex and interconnected system. Consequently, it is a rich source of side channels, and can be viewed as an amalgamation of components (or subsystems) in a number of ways. For example, it can be segmented based on the location of components such as the back, thorax, and abdomen [64]. Each in turn is comprised of subcomponents such as the musculoskeletal components of the back (vertebrae, scapula, vertebral column, and pelvic bone) [64]. An alternative view is of its major systems (e.g., the nervous system) [65]. Spatial coherence supports tracking of the physical paths followed by target information through neighbouring components. A condition’s cause inhabiting one component may manifest entirely elsewhere, creating opportunities for information leakage [1]. The human body’s interconnectivity is a fundamental principle that should be considered when approaching SCS for MDM.
- HCI
- interconnects two distinct target systems, namely, biological systems (e.g., a human’s intentions and intended movements) and electronics-based systems. HCI solutions use the electronic systems to ‘sense’ an appropriate side-channel signal identified within the biological system. The signal is input to a computer which contains an actuation element, such as a display, which feeds a response of the sensed signal back to the human. Unlike MDM, where the human is a passive generator of signals, in HCI signals are actively generated through intentional gestures [13].
- Misc
- may exhibit a selection bias in which it is likely to identify more extreme system structures. Examples range from using generic smartphone sensors to categorise driver behaviour [57], exploiting the existing complexity of a taxi network [49], and more complex interactions. Insights into the system configuration, such as wall position and router placement [56], are parameters used to support signal processing. Compared to previous domains, these target systems place less emphasis on the physical co-location of system components. Existing active elements in the target system are exploited by measuring changes in reflected signals, for instance to detect movement [54,55] or by using potholes to induce pain levels [66]. New elements are added to trigger a side-channel response [48]. Examples in the Misc category demonstrate greater diversity in terms of the system components exploited or added compared to those of the other domains. As illustrated by the potato chip bag microphone in [53], any system component (specifically in addition to electronics or local active elements) can be a relevant part of the target system. The boundary of target systems is also less well defined ( versus, e.g., a smartphone), covering large areas such as a country’s electrical grid [51], or a road network [49].
- A target system consists of its components and their interconnections. System structures associated with side channels have long paths (sequences of components) which increase leakage opportunities [1,9,10]. The interconnections between components allow information to be mixed and distributed to different destinations. Information can be retrieved by manipulating components to inject signals, trigger actions, or generate signals that provide insight into internal operations, such as reflected signals from a WiFi router [56].
2.1.3. Side Channels
- CYB
- capitalises on a property of side channels whereby the signal embedded within is transformed and modified during traversal. For example, target information originating from a website could be sent as internet traffic, stored in a text file, and then printed to paper. Target information manifests physically as audio produced by dot-matrix printers [24], vibrations within 3D printers [25], or from observing the magnetic field of a laser printer [26]. Side channels do need to ultimately derive from target information. For example, when exploiting electromagnetic radiation for EDCH encryption algorithm key extraction [16], the side channel results from the mutual information between the internal processing of the EDCH decryption and the electromagnetic radiation emitted by the device. Multiple side channels can be viable for a single target system (e.g., PCs [15,69,70,71]); these may be exploited either individually or concurrently (e.g., sensor fusion) [25].
- MDM
- exploits biomarkers, which are quantifiable signals indicative of normal biological processes [72]. Biomarkers may exist internally or externally; thus, their observability varies. When internal, direct measurement with ubiquitous, cheap, and available sensors becomes non-trivial, with reliance on channels where the target information travels to a sensing site being more accessible. For example, dehydration has an externally available biomarker in the level of pH in sweat [73,74], which is quantifiable via colorimetric strips and can be analysed through a smartphone camera [75]. Established gold standard diagnostic devices rely on proven biomarkers. SCS offers the potential to exploit them in new ways or even discover new ones.
- HCI
- exploits the way in which information ‘leaks’ along side channels, with SCS solutions existing for the human and/or computer components:
- Computer channels: Because channels are not deliberately protected (e.g., encrypted), HCI deals with primary channels used in an expected or non-trivial way or via existing or introduced side-channels (for example, sensing pressure through the interaction between accelerometer and vibration generated by the motor built into a smartphone) [44].
- Human channels: Channels exploited in HCI are those that carry information to more accessible or convenient external locations, such as using electrodes placed within the earlobe to detect eye movement [38].
- Misc
- involves long pathways from the target information to the sensor [51], and includes signal transformations such as video to sound [53] and separation of mixtures of signals [77]. Missing scenarios include those where the side channel is predominantly virtual (information), as physical interactions are favoured. As such, a side channel exists as one of a number of signals mixed into physical signals, such as light flicker in a video [51]. Otherwise, the side channel may be fragmented across multiple signals, requiring sensors for all signal types (such as acceleration and rotation to recover driving behaviour [57], or multiple sensor elements using the same signal to enhance recovery in a rolling shutter with individual image elements to recover a high frequency signal) [53].
- Side channels exist in target systems from any domain, respresenting a key insight expressed and explored in this research. Side channels result from entanglement operations within the target system. They allow one signal to be detected by sensing it when mixed in with another. Correlation is one approach used by all domains for identifying side-channel candidates, and is usable even when the causal relationship is not well understood [1]. CYB treats side channels as containing hidden information that needs to be recovered through cunning (i.e., outmanoeuvring defences or obstacles) [69,70]. In contrast, MDM exploits surprising pathways with embedded target information. HCI side channels are not deliberately hidden, but exist due to difficulties in directly sensing the target information. Feedback is one way of achieving the tuning of system parameters and sensor placement required to achieve SCS [68,76]. While electronic and physical/biological channels represent information differently, analogous operations are possible in both. In certain cases, access to one side channel improves opportunities to extract another related channel.
2.1.4. Side-Channel Properties
- Determinism: There is a reproducible, reliable correlation between internal operations and any quantified signals.
- Multi-Stage Pathways: A signal often traverses multiple stages along its pathway from internal interconnected components to its sensing site. This makes discovery of all paths non-trivial.
- Understanding of side channels via their properties provides insights into how to extract their target information, and even into how to discover them. The key is determinism. A reliable correlation between the internal source and quantified output signals indicates a candidate side channel. This is the case irrespective of whether the mechanism enabling it is understood. Furthermore, all side channels are indirect by definition; therefore, the multi-stage pathway property is present in all candidate side-channels. The following additional side-channel properties reflect how signals are subject to a myriad of modifications during traversal of a side channel:
- Signal Transformations: Signals are transformative between states. In MDM for example, excessive bilirubin buildup in the bloodstream (a by-product of recycling red blood cells) manifests as a yellowish discolouration of the skin and sclera if subject to jaundice [67]. Transformations may be invoked via interference (perhaps intentionally by an observer), signal reflection/refraction, or due to obstructions. Additionally, they may occur both within and beyond the target system’s boundary. Signal transformations are capitalised on by all domains, either via custom sensors or, as is common in MDM and HCI, using existing sensors on a constrained platform (e.g., a smartphone).
- Modulation Proportion: How prominent the target information is within a side channel is determined by its proportion in relation to the signal in which it is embedded. A low proportion when mixed with other signals may make the target information difficult to detect.
- Signal Mixing: Target systems internally consist of interwoven channels in which their signals mix and collide. Target information that is not directly accessible may instead be detected when mixed with another signal. Alternatively, a signal external and alien to the side channel (or even the target system itself) could be injected and mixed into the side channel.
- Multivariate: A single target information source can have multiple associated side channels, increasing the number of available attack vectors.
- Such properties provide insight into the side-channel signal’s structure and embedded target information. The multi-stage nature of side-channels means that side channels emanate and traverse from their source in a myriad of ways, during which time side-channels are mixed, transformed, and vary in modulation (Figure 2). Consequently, multiple side channels may be plausible for the same target information. MDM is an example, with the human body’s interconnected nature promoting an array of side channels. For example, a single instance of target information such as the heart rate is accessible by traditional gold standards (ECG, stethoscope) as well as by “less obvious” side channels: within the ear canal [78], via chest and head movement [79,80,81,82], photoplethysmography by a camera view of the ear canal [83] or face [30,31,74], or monitoring of chest oscillations via WiFi (e.g., electromagnetic radiation) [32].
2.1.5. Information Parameters
2.1.6. Sensors
- CYB
- selects sensors most appropriate for the signal in question (e.g., copper coils for electromagnetic radiation [87]). Measurement considers the reduction of extraneous information sources, the placement of sensors, and concurrent internal processes within a target system [14]. CYB recognises that noise is potentially the source of side channels. Trial and error is common for refining measurements, such as finding the optimal location to sense electromagnetic radiation from a PC [16].
- MDM
- utilises sensors along a spectrum as per its context specific suitability, as described by Spence and Bangay [1]:
- Stand-alone sensors: Large variety, small size, ubiquitous, and easily embedded into devices [88].
- HCI
- adopts a large array of signal types. This is due to HCI’s affinity for prototyping and its focus on two distinct target systems, one electronic-based and the other biological-based. Emphasis is placed on consideration of the information parameters of the quantified signals and how they can be best exploited. Acoustics propagate along or through materials [42] and over distance through empty space [13], both of which enable detection of user input and movements at distance. Sensors can be placed at varied and multiple locations and quantify signals locally or remotely depending on the information parameters being exploited.
- Misc
- tends to deal with signals originating from physical processes. This can be exploited by deliberately triggering a signal change [45,48]. Sensors range from those already available in a standard smartphone [57] to customised rigs that adapt samples for better sensing (e.g., by staining samples to highlight allergens [94]) or actively injecting signals to trigger side channels [48]. Human detection using WiFi progresses from using customised hardware to using existing installations only [54,55,56].
- Sensing in CYB and HCI is achieved with custom sensors deliberately chosen as per the information parameters in focus. Smartphones are commonly used in MDM and HCI because they are ubiquitous and portable with an array of embedded sensors. These factors allow for extended data collection, replacing brief access to high quality but expensive devices. Interesting sensors result from repurposing existing sensors, such as a lensless microscope produced by projecting holographic interference patterns onto a smartphone camera [95]. Examples exist where the same sensor (particularly on smartphones) is used for diverse purposes. In this way, SCS solutions can achieve the same outcome using very different pathways, for instance, through alternative versions of the same gold standard in MDM [1] or the CYB examples that identify screen content through power consumption, audio signals, or electromagnetic radiation. Use of various standalone sensors often indicates research that is in the prototyping stages, representing an effective strategy when approaching the non-obvious nature of side channels. Notably, sensor placement is as crucial as which sensor is used and how [96,97].
2.1.7. Methods and Extraction Techniques
- Invasive vs. Non-Invasive: The target system is physically modified to provide (better) access to specific signals, or only originally accessible signals are sensed.
- Active vs. Passive: Control over the target system’s operations are exploited, perhaps by repeated execution of internal operations to trigger specific signals, or used to aid in the study of how internal operations work, such as the injection of particular input to measure the output signals and their behaviour.
- Remote vs. Local: The side channel’s information parameters (Section 2.1.5) define whether sensing can (and should) occur at a distance. Local measurements provide clearer signals, although this may not be possible depending on the level of access to the target system or the measurement intention (e.g., covert sensing).
- Profiled vs. Non-Profiled: With unimpeded access to the target system, a large number of measurements are collected to build a model of the determinism between its internal operations and sensed signals. Future measurements of this target system are analysed in the context of this profile.
- Utilising existing third-party data: Third-party collected data may already inadvertently house embedded target information.
- Multivariate: Sensor fusion techniques involve the collation of data from multiple sensed side channels output from the same target system.
- Signal Injection: Bespoke signals are injected into the target system to invoke observable or specific output signals. This also includes the intention to invoke faults (i.e., behaviour different to its original design) in order to create measurable outputs to learn more of the internal operations.
- Software-based approaches to expand hardware: Target systems that are not modifiable physically (per an invasive method) may limit available output signals. However, an array of sensors may already exist within the target system itself (e.g., a smartphone), and a multivariate approach can collate them to provide access to target information.
- Repurposing Sensor: Sensors used as actuators can often ‘sense’ target information beyond their original design intention, for example, the ability to recreate audio from visual recordings via a camera [53].
- Similarly, Spence and Bangay [2] collated SCS extraction techniques abstracted so as to be applicable across domains:
- Power analysis attacks: Exploitation of deterministic correlations between internal operations and the output quantified signal, for example, a computational device’s power consumption or emitted electromagnetic radiation. Sufficient resolution in the quantified signal can reveal executed operations, including the execution sequence, from which target information can be inferred or reconstructed.
- Information-theoretical analysis: Treats signals as noisy, with target information muddled within. Encompasses signal processing techniques from information theory (e.g., Shannon’s entropy, Hamming weights), cryptanalysis, statistical analysis (e.g., maximum likelihood [99,100], correlation, or simple regressions), and transformations (e.g., FFTs).
- Machine learning: When large sensing datasets from side channels can be create or acquired, machine learning offers automated feature extraction stages for identifying how target information can be extracted.
- The crux of our SCS research is that noise in a signal may actually contain meaningful content (i.e., target information), thereby qualifying as a candidate side channel. The lack of a unified SCS framework results in solutions using ad hoc sequences of features, filters, and other individually tuned signal processing stages. Adoption of mindsets from varying domains represents an opportunity to reveal novel SCS extraction techniques; for example, one may ‘attack’ the human body for MDM purposes similarly to how cyber–physical systems are exploited within CYB [1].
Reliance on Information Parameters
2.1.8. Summary
3. Analysis and Discussion
3.1. Transferability of SCS Extraction Techniques
3.2. Representation of Side Channels
3.3. Additional Domains
4. Conclusions
Author Contributions
Funding
Data Availability Statement
Conflicts of Interest
References
- Spence, A.; Bangay, S. Side-Channel Sensing: Exploiting Side-Channels to Extract Information for Medical Diagnostics and Monitoring. IEEE J. Transl. Eng. Health Med. 2020, 8, 1–13. [Google Scholar] [CrossRef] [PubMed]
- Spence, A.; Bangay, S. Security beyond cybersecurity: Side-channel attacks against non-cyber systems and their countermeasures. Int. J. Inf. Secur. 2021, 21, 437–453. [Google Scholar] [CrossRef]
- Standaert, F.X.; Malkin, T.G.; Yung, M. A Unified Framework for the Analysis of Side-Channel Key Recovery Attacks. In Advances in Cryptology-EUROCRYPT 2009: 28th Annual International Conference on the Theory and Applications of Cryptographic Techniques, Cologne, Germany, 26–30 April 2009; Proceedings 28; Springer: Berlin/Heidelberg, Germany, 2009; pp. 443–461. [Google Scholar] [CrossRef]
- Spreitzer, R.; Moonsamy, V.; Korak, T.; Mangard, S. Systematic Classification of Side-Channel Attacks: A Case Study for Mobile Devices. IEEE Commun. Surv. Tutorials 2018, 20, 465–488. [Google Scholar] [CrossRef]
- Giechaskiel, I.; Rasmussen, K. Taxonomy and Challenges of Out-of-Band Signal Injection Attacks and Defenses. IEEE Commun. Surv. Tutorials 2020, 22, 645–670. [Google Scholar] [CrossRef]
- Lange, J.; Massart, C.; Mouraux, A.; Standaert, F.X. Side-Channel Attacks Against the Human Brain: The PIN Code Case Study. In Constructive Side-Channel Analysis and Secure Design: 8th International Workshop, COSADE 2017, Paris, France, 13–14 April 2017; Revised Selected Papers 8; Guilley, S., Ed.; Springer: Cham, Switzerland, 2017; pp. 171–189. [Google Scholar]
- Martinovic, I.; Davies, D.; Frank, M.; Perito, D.; Ros, T.; Song, D. On the Feasibility of Side-Channel Attacks with Brain-Computer Interfaces. In Proceedings of the USENIX Security Symposium, Bellevue, WA, USA, 8–10 August 2012; pp. 143–158. [Google Scholar]
- Weber, D.; Ibrahim, A.; Nemati, H.; Schwarz, M.; Rossow, C. Osiris: Automated Discovery of Microarchitectural Side Channels. arXiv 2021, arXiv:2106.03470. [Google Scholar]
- Kemmerer, R.A.; Porras, P.A. Covert Flow Trees: A Visual Approach to Analyzing Covert Storage Channels. IEEE Trans. Softw. Eng. 1991, 17, 1166–1185. [Google Scholar] [CrossRef]
- Rodrigues, B.; Quintão Pereira, F.M.; Aranha, D.F. Sparse Representation of Implicit Flows with Applications to Side-Channel Detection. In Proceedings of the 25th International Conference on Compiler Construction, Barcelona, Spain, 17–18 March 2016; pp. 110–120. [Google Scholar] [CrossRef]
- Ferraiuolo, A.; Xu, R.; Zhang, D.; Myers, A.C.; Suh, G.E. Verification of a Practical Hardware Security Architecture Through Static Information Flow Analysis. In Proceedings of the Twenty-Second International Conference on Architectural Support for Programming Languages and Operating Systems, Xi’an, China, 8–12 April 2017; pp. 555–568. [Google Scholar] [CrossRef]
- Spence, A. Side-Channel Sensing: Systematic Discovery of Side-Channels. Ph.D. Dissertation, Deakin University, Geelong, Australia, 2023. [Google Scholar]
- Lee, J.s.; Yeo, W.S. Sonicstrument: A Musical Interface with Stereotypical Acoustic Transducers. In Proceedings of the International Conference on New Interfaces for Musical Expression, Oslo, Norway, 30 May–1 June 2011; pp. 24–27. [Google Scholar]
- Standaert, F.X. Introduction to Side-Channel Attacks; Springer: Boston, MA, USA, 2010; pp. 27–42. [Google Scholar] [CrossRef]
- Kocher, P.; Jaffe, J.; Jun, B. Differential Power Analysis; Springer: Berlin, Germany, 1999; pp. 388–397. [Google Scholar] [CrossRef]
- Genkin, D.; Pachmanov, L.; Pipman, I.; Tromer, E. ECDH key-extraction via low-bandwidth electromagnetic attacks on PCs. In Topics in Cryptology-CT-RSA 2016: The Cryptographers’ Track at the RSA Conference 2016, San Francisco, CA, USA, 29 February–4 March 2016; Springer: Berlin/Heidelberg, Germany, 2016; pp. 219–235. [Google Scholar]
- Genkin, D.; Pipman, I.; Tromer, E. Get your hands off my laptop: Physical side-channel key-extraction attacks on PCs. J. Cryptogr. Eng. 2015, 5, 95–112. [Google Scholar] [CrossRef]
- Zhou, Y.; Feng, D. Side-Channel Attacks: Ten Years After Its Publication and the Impacts on Cryptographic Module Security Testing. IACR Cryptol. EPrint Arch. 2005, 2005, 388. [Google Scholar]
- Tunstall, M. Smart Card Security. In Smart Cards, Tokens, Security and Applications; Mayes, K., Markantonakis, K., Eds.; Springer International Publishing: Cham, Switzerland, 2017; pp. 217–251. [Google Scholar] [CrossRef]
- Greveler, U.; Justus, B.; Loehr, D. Multimedia content identification through smart meter power usage profiles. In Proceedings of the International Conference on Information and Knowledge Engineering (IKE); The Steering Committee of The World Congress in Computer Science, Computer Engineering and Applied Computing (WorldComp): Athens, Greece, 2012; p. 1. [Google Scholar]
- Zhang, K.; Li, Z.; Wang, R.; Wang, X.; Chen, S. Sidebuster: Automated detection and quantification of side-channel leaks in web application development. In Proceedings of the 17th ACM Conference on Computer and Communications Security 2010, Chicago, IL, USA, 4–8 October 2010. [Google Scholar] [CrossRef]
- Mather, L.; Oswald, E. Pinpointing side-channel information leaks in web applications. J. Cryptogr. Eng. 2012, 2, 161–177. [Google Scholar] [CrossRef]
- Chen, S.; Wang, R.; Wang, X.; Zhang, K. Side-Channel Leaks in Web Applications: A Reality Today, a Challenge Tomorrow. In Proceedings of the 2010 IEEE Symposium on Security and Privacy, Oakland, CA, USA, 16–19 May 2010; pp. 191–206. [Google Scholar] [CrossRef]
- Backes, M.; Dürmuth, M.; Gerling, S.; Pinkal, M.; Sporleder, C. Acoustic Side-Channel Attacks on Printers. In Proceedings of the USENIX Security Symposium, Washington, DC, USA, 11–13 August 2010; pp. 307–322. [Google Scholar]
- Chhetri, S.R.; Faruque, M.A.A. Side-Channels of Cyber-Physical Systems: Case Study in Additive Manufacturing. IEEE Des. Test 2017, 34, 18–25. [Google Scholar] [CrossRef]
- Tosaka, T.; Taira, K.; Yamanaka, Y.; Nishikata, A.; Hattori, M. Feasibility study for reconstruction of information from near field observations of the magnetic field of laser printer. In Proceedings of the 2006 17th International Zurich Symposium on Electromagnetic Compatibility, Singapore, 27 February–3 March 2006; pp. 630–633. [Google Scholar] [CrossRef]
- Sanyal, S.; Nundy, K.K. Algorithms for Monitoring Heart Rate and Respiratory Rate From the Video of a User’s Face. IEEE J. Transl. Eng. Health Med. 2018, 6, 1–11. [Google Scholar] [CrossRef]
- Grimaldi, D.; Kurylyak, Y.; Lamonaca, F.; Nastro, A. Photoplethysmography detection by smartphone’s videocamera. In Proceedings of the 6th IEEE International Conference on Intelligent Data Acquisition and Advanced Computing Systems, Prague, Czech Republic, 15–17 September 2011; Volume 1, pp. 488–491. [Google Scholar] [CrossRef]
- Lee, J.; Reyes, B.A.; McManus, D.D.; Maitas, O.; Chon, K.H. Atrial Fibrillation Detection Using an iPhone 4S. IEEE Trans. Biomed. Eng. 2013, 60, 203–206. [Google Scholar] [CrossRef] [PubMed]
- Ming-Zher, P.; McDuff, D.J.; Picard, R.W. Advancements in Noncontact, Multiparameter Physiological Measurements Using a Webcam. Biomed. Eng. IEEE Trans. 2011, 58, 7–11. [Google Scholar] [CrossRef]
- Poh, M.Z.; McDuff, D.J.; Picard, R.W. Non-contact, automated cardiac pulse measurements using video imaging and blind source separation. Opt. Express 2010, 18, 10762–10774. [Google Scholar] [CrossRef] [PubMed]
- Adib, F.; Mao, H.; Kabelac, Z.; Katabi, D.; Miller, R.C. Smart Homes that Monitor Breathing and Heart Rate. In Proceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems—CHI’15, Seoul, Republic of Korea, 18–23 April 2015; pp. 837–846. [Google Scholar] [CrossRef]
- Larson, E.C.; Goel, M.; Boriello, G.; Heltshe, S.; Rosenfeld, M.; Patel, S.N. SpiroSmart: Using a microphone to measure lung function on a mobile phone. In Proceedings of the 2012 ACM Conference on Ubiquitous Computing—UbiComp ’12, Pittsburgh, PA, USA, 5–8 September 2012; pp. 280–289. [Google Scholar] [CrossRef]
- Ding, X.; Nassehi, D.; Larson, E.C. Measuring Oxygen Saturation With Smartphone Cameras Using Convolutional Neural Networks. IEEE J. Biomed. Health Informatics 2019, 23, 2603–2610. [Google Scholar] [CrossRef] [PubMed]
- Kang, S.J.; Choi, J.H.; Kim, Y.J.; Ma, H.I.; Lee, U. Development of an acquisition and visualization of forearm tremors and pronation/supination motor activities in a smartphone based environment for an early diagnosis of Parkinson’s disease. Adv. Sci. Technol. Lett. 2015, 116, 209–212. [Google Scholar]
- LeMoyne, R.; Mastroianni, T.; Cozza, M.; Coroian, C.; Grundfest, W. Implementation of an iPhone for characterizing Parkinson’s disease tremor through a wireless accelerometer application. In Proceedings of the 2010 Annual International Conference of the IEEE Engineering in Medicine and Biology, Buenos Aires, Argentina, 31 August–4 September 2010; pp. 4954–4958. [Google Scholar] [CrossRef]
- Wangwiwattana, C.; Ding, X.; Larson, E.C. PupilNet, Measuring Task Evoked Pupillary Response using Commodity RGB Tablet Cameras. Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. 2018, 1, 1–26. [Google Scholar] [CrossRef]
- Manabe, H.; Fukumoto, M.; Yagi, T. Conductive rubber electrodes for earphone-based eye gesture input interface. Pers. Ubiquitous Comput. 2014, 19, 143–154. [Google Scholar] [CrossRef]
- Sahni, H.; Bedri, A.; Reyes, G.; Thukral, P.; Guo, Z.; Starner, T.; Ghovanloo, M. The tongue and ear interface: A wearable system for silent speech recognition. In Proceedings of the 2014 ACM Conference on Ubiquitous Computing, Seattle, WA, USA, 13–17 September 2014. [Google Scholar] [CrossRef]
- Bedri, A.; Byrd, D.; Presti, P.; Sahni, H.; Gue, Z.; Starner, T. Stick it in your ear: Building an in-ear jaw movement sensor. In Proceedings of the 2015 ACM International Joint Conference on Pervasive and Ubiquitous Computing, Osaka, Japan, 7–11 September 2015. [Google Scholar] [CrossRef]
- Bedri, A.; Li, R.; Haynes, M.; Kosaraju, R.P.; Grover, I.; Prioleau, T.; Beh, M.Y.; Goel, M.; Starner, T.; Abowd, G. EarBit: Using Wearable Sensors to Detect Eating Episodes in Unconstrained Environments. Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. 2017, 1, 37. [Google Scholar] [CrossRef]
- Harrison, C.; Hudson, S.E. Scratch input: Creating large, inexpensive, unpowered and mobile finger input surfaces. In Proceedings of the Annual ACM Symposium on User Interface Software and Technology, Monterey, CA, USA, 19–22 October 2008; pp. 205–208. [Google Scholar]
- Low, S.; Sugiura, Y.; Lo, D.; Inami, M. Pressure detection on mobile phone by camera and flash. In Proceedings of the 5th Augmented Human International Conference, Kobe, Japan, 7–8 March 2014. [Google Scholar] [CrossRef]
- Hwang, S.; Bianchi, A.; Wohn, K.Y. VibPress: Estimating pressure input using vibration absorption on mobile devices. In Proceedings of the 15th International Conference on Human-Computer Interaction with Mobile Devices and Services, Munich, Germany, 27–30 August 2013. [Google Scholar] [CrossRef]
- Sullenberger, R.M.; Kaushik, S.; Wynn, C.M. Photoacoustic communications: Delivering audible signals via absorption of light by atmospheric H2O. Opt. Lett. 2019, 44, 622–625. [Google Scholar] [CrossRef]
- Qin, Y.; Carlini, N.; Cottrell, G.; Goodfellow, I.; Raffel, C. Imperceptible, Robust, and Targeted Adversarial Examples for Automatic Speech Recognition. In Proceedings of the 36th International Conference on Machine Learning, Long Beach, CA, USA, 9–15 June 2019; Chaudhuri, K., Salakhutdinov, R., Eds.; Volume 97, pp. 5231–5240. [Google Scholar]
- Han, C.; O’Sullivan, J.; Luo, Y.; Herrero, J.; Mehta, A.D.; Mesgarani, N. Speaker-independent auditory attention decoding without access to clean speech sources. Sci. Adv. 2019, 5, aav6134. [Google Scholar] [CrossRef] [PubMed]
- Rahman, T.; Adams, A.T.; Schein, P.; Jain, A.; Erickson, D.; Choudhury, T. Nutrilyzer: A Mobile System for Characterizing Liquid Food with Photoacoustic Effect. In Proceedings of the 14th ACM Conference on Embedded Network Sensor Systems CD-ROM, Stanford, CA, USA, 14–16 November 2016; pp. 123–136. [Google Scholar] [CrossRef]
- Eriksson, J.; Girod, L.; Hull, B.; Newton, R.; Madden, S.; Balakrishnan, H. The pothole patrol: Using a mobile sensor network for road surface monitoring. In Proceedings of the 6th International Conference on Mobile Systems, Applications, and Services, Breckenridge, CO, USA, 17–20 June 2008; pp. 29–39. [Google Scholar]
- Wang, W.; He, S.; Sun, L.; Jiang, T.; Zhang, Q. Cross-Technology Communications for Heterogeneous IoT Devices Through Artificial Doppler Shifts. IEEE Trans. Wirel. Commun. 2019, 18, 796–806. [Google Scholar] [CrossRef]
- Garg, R.; Hajj-Ahmad, A.; Wu, M. Geo-location estimation from Electrical Network Frequency signals. In Proceedings of the 2013 IEEE International Conference on Acoustics, Speech and Signal Processing, Vancouver, BC, Canada, 26–31 May 2013; pp. 2862–2866. [Google Scholar] [CrossRef]
- Genkin, D.; Pattani, M.; Schuster, R.; Tromer, E. Synesthesia: Detecting Screen Content via Remote Acoustic Side Channels. In Proceedings of the 2019 IEEE Symposium on Security and Privacy (SP), San Francisco, CA, USA, 19–23 May 2019. [Google Scholar]
- Davis, A.; Rubinstein, M.; Wadhwa, N.; Mysore, G.J.; Durand, F.; Freeman, W.T. The Visual Microphone: Passive Recovery of Sound from Video. ACM Trans. Graph. 2014, 33, 79:1–79:10. [Google Scholar] [CrossRef]
- Zhu, Y.; Xiao, Z.; Chen, Y.; Li, Z.; Liu, M.; Zhao, B.Y.; Zheng, H. Adversarial WiFi Sensing. arXiv 2018, arXiv:1810.10109. [Google Scholar]
- Yang, J.; Zou, H.; Jiang, H.; Xie, L. Device-Free Occupant Activity Sensing Using WiFi-Enabled IoT Devices for Smart Homes. IEEE Internet Things J. 2018, 5, 3991–4002. [Google Scholar] [CrossRef]
- Adib, F.; Katabi, D. See through walls with WiFi! ACM SIGCOMM Comput. Commun. Rev. 2013, 43, 75–86. [Google Scholar] [CrossRef]
- Johnson, D.; Trivedi, M.M. Driving style recognition using a smartphone as a sensor platform. In Proceedings of the 2011 14th International IEEE Conference on Intelligent Transportation Systems (ITSC), Washington, DC, USA, 5–7 October 2011; pp. 1609–1615. [Google Scholar]
- Lemke-Rust, K.; Samarin, P. Exploring Embedded Software with Side Channels and Fault Analysis. In Proceedings of the 2018 12th European Workshop on Microelectronics Education (EWME), Braunschweig, Germany, 24–26 September 2018; pp. 67–70. [Google Scholar]
- Chari, S.; Rao, J.R.; Rohatgi, P. Template Attacks. In Cryptographic Hardware and Embedded Systems-CHES 2002: 4th International Workshop Redwood Shores, CA, USA, 13–15 August 2002; Springer: Berlin/Heidelberg, Germany, 2003; pp. 13–28. [Google Scholar] [CrossRef]
- Kuhn, M.G. Electromagnetic eavesdropping risks of flat-panel displays. In Privacy Enhancing Technologies; Springer: Berlin/Heidelberg, Germany, 2005; Volume 3424, pp. 88–107. [Google Scholar]
- Yaqoob, T.; Abbas, H.; Atiquzzaman, M. Security Vulnerabilities, Attacks, Countermeasures, and Regulations of Networked Medical Devices–A Review. IEEE Commun. Surv. Tutorials 2019, 21, 3723–3768. [Google Scholar] [CrossRef]
- Yang, Q.; Gasti, P.; Zhou, G.; Farajidavar, A.; Balagani, K.S. On inferring browsing activity on smartphones via USB power analysis side-channel. IEEE Trans. Inf. Forensics Secur. 2016, 12, 1056–1066. [Google Scholar] [CrossRef]
- Vuagnoux, M.; Pasini, S. An improved technique to discover compromising electromagnetic emanations. In Proceedings of the 2010 IEEE International Symposium on Electromagnetic Compatibility, Fort Lauderdale, FL, USA, 25–30 July 2010; pp. 121–126. [Google Scholar] [CrossRef]
- Drake, R.; Vogl, A.W.; Mitchell, A.W. Gray’s Anatomy for Students E-Book; Elsevier Health Sciences: Amsterdam, The Netherlands, 2009. [Google Scholar]
- Boron, W.F.; Boulpaep, E.L. Medical Physiology, 2e Updated Edition E-Book; Elsevier Health Sciences: Amsterdam, The Netherlands, 2012. [Google Scholar]
- Ashdown, H.F.; D’Souza, N.; Karim, D.; Stevens, R.J.; Huang, A.; Harnden, A. Pain over speed bumps in diagnosis of acute appendicitis: Diagnostic accuracy study. Bmj 2012, 345, e8012. [Google Scholar] [CrossRef]
- de Greef, L.; Goel, M.; Seo, M.J.; Larson, E.C.; Stout, J.W.; Taylor, J.A.; Patel, S.N. Bilicam: Using Mobile Phones to Monitor Newborn Jaundice. In Proceedings of the 2014 ACM International Joint Conference on Pervasive and Ubiquitous Computing—UbiComp ’14, Seattle, WA, USA, 13–17 September 2014; pp. 331–342. [Google Scholar] [CrossRef]
- Lux, E.; Adam, M.; Dorner, V.; Helming, S.; Knierim, M.; Weinhardt, C. Live Biofeedback as a User Interface Design Element: A Review of the Literature. Commun. Assoc. Inf. Syst. 2018, 43, 257–296. [Google Scholar] [CrossRef]
- Biswas, A.K.; Ghosal, D.; Nagaraja, S. A Survey of Timing Channels and Countermeasures. ACM Comput. Surv. 2017, 50, 1–39. [Google Scholar] [CrossRef]
- Carrara, B.; Adams, C. Out-of-Band Covert Channels—A Survey. ACM Comput. Surv. 2016, 49, 1–36. [Google Scholar] [CrossRef]
- Guri, M.; Solewicz, Y.; Elovici, Y. MOSQUITO: Covert Ultrasonic Transmissions Between Two Air-Gapped Computers Using Speaker-to-Speaker Communication. In Proceedings of the 2018 IEEE Conference on Dependable and Secure Computing (DSC), Kaohsiung, Taiwan, 10–13 December 2018; pp. 1–8. [Google Scholar] [CrossRef]
- Biomarkers Definitions Working Group. Biomarkers and surrogate endpoints: Preferred definitions and conceptual framework. Clin. Pharmacol. Ther. 2001, 69, 89–95. [Google Scholar] [CrossRef] [PubMed]
- Morgan, R.M.; Patterson, M.J.; Nimmo, M.A. Acute effects of dehydration on sweat composition in men during prolonged exercise in the heat. Acta Physiol. Scand. 2004, 182, 37–43. [Google Scholar] [CrossRef] [PubMed]
- Scully, C.G.; Lee, J.; Meyer, J.; Gorbach, A.M.; Granquist-Fraser, D.; Mendelson, Y.; Chon, K.H. Physiological Parameter Monitoring from Optical Recordings With a Mobile Phone. IEEE Trans. Biomed. Eng. 2012, 59, 303–306. [Google Scholar] [CrossRef]
- Oncescu, V.; O’Dell, D.; Erickson, D. Smartphone based health accessory for colorimetric detection of biomarkers in sweat and saliva. Lab Chip 2013, 13, 3232. [Google Scholar] [CrossRef] [PubMed]
- Ritter, W. Benefits of Subliminal Feedback Loops in Human-Computer Interaction. Adv. Hum.-Comput. Interact. 2011, 2011, 346492:1–346492:11. [Google Scholar] [CrossRef]
- Vincent, E.; Bertin, N.; Gribonval, R.; Bimbot, F. From blind to guided audio source separation: How models and side information can improve the separation of sound. IEEE Signal Process. Mag. 2014, 31, 107–115. [Google Scholar] [CrossRef]
- Nirjon, S.; Zhao, F.; Dickerson, R.F.; Li, Q.; Asare, P.; Stankovic, J.A.; Hong, D.; Zhang, B.; Jiang, X.; Shen, G. MusicalHeart: A hearty way of listening to music. In Proceedings of the 10th ACM Conference on Embedded Network Sensor Systems—SenSys’12, Toronto, ON, Canada, 6–9 November 2012; pp. 43–56. [Google Scholar] [CrossRef]
- Jaakkola, J.; Jaakkola, S.; Lahdenoja, O.; Hurnanen, T.; Koivisto, T.; Pankaala, M.; Knuutila, T.; Kiviniemi, T.O.; Vasankari, T.; Airaksinen, K.J. Mobile Phone Detection of Atrial Fibrillation With Mechanocardiography. Circulation 2018, 137, 1524–1527. [Google Scholar] [CrossRef]
- Lee, J.; Reyes, B.A.; McManus, D.D.; Mathias, O.; Chon, K.H. Atrial fibrillation detection using a smart phone. In Proceedings of the 2012 Annual International Conference of the IEEE Engineering in Medicine and Biology Society, San Diego, CA, USA, 28 August–1 September 2012; pp. 1177–1180. [Google Scholar] [CrossRef]
- McManus, D.D.; Lee, J.; Maitas, O.; Esa, N.; Pidikiti, R.; Carlucci, A.; Harrington, J.; Mick, E.; Chon, K.H. A novel application for the detection of an irregular pulse using an iPhone 4S in patients with atrial fibrillation. Heart Rhythm 2013, 10, 315–319. [Google Scholar] [CrossRef]
- Balakrishnan, G.; Durand, F.; Guttag, J. Detecting Pulse from Head Motions in Video. In Proceedings of the 2013 IEEE Conference on Computer Vision and Pattern Recognition, Portland, OR, USA, 23–28 June 2013; pp. 3430–3437. [Google Scholar] [CrossRef]
- Poh, M.Z.; Kim, K.; Goessling, A.D.; Swenson, N.C.; Picard, R.W. Heartphones: Sensor Earphones and Mobile Application for Non-obtrusive Health Monitoring. In Proceedings of the 2009 International Symposium on Wearable Computers, Linz, Austria, 4–7 September 2009; pp. 153–154. [Google Scholar] [CrossRef]
- Shannon, C.E. A mathematical theory of communication. Bell Syst. Tech. J. 1948, 27, 379–423. [Google Scholar] [CrossRef]
- Demme, J.; Martin, R.; Waksman, A.; Sethumadhavan, S. Side-channel vulnerability factor: A metric for measuring information leakage. In Proceedings of the 2012 39th Annual International Symposium on Computer Architecture (ISCA), Portland, OR, USA, 9–13 June 2012; pp. 106–117. [Google Scholar] [CrossRef]
- Hayashi, Y.; Homma, N.; Miura, M.; Aoki, T.; Sone, H. A Threat for Tablet PCs in Public Space: Remote Visualization of Screen Images Using EM Emanation. In Proceedings of the 2014 ACM SIGSAC Conference on Computer and Communications Security—CCS’14, Scottsdale, Arizona, USA, 3–7 November 2014; pp. 954–965. [Google Scholar] [CrossRef]
- Lomné, V.; Prouff, E.; Roche, T. Behind the Scene of Side Channel Attacks. In Advances in Cryptology—ASIACRYPT 2013: 19th International Conference on the Theory and Application of Cryptology and Information Security, Bengaluru, India, 1–5 December 2013; Proceedings, Part I; Sako, K., Sarkar, P., Eds.; Springer: Berlin/Heidelberg, Germany, 2013; pp. 506–525. [Google Scholar] [CrossRef]
- Larson, E.C.; Lee, T.; Liu, S.; Rosenfeld, M.; Patel, S.N. Accurate and privacy preserving cough sensing using a low-cost microphone. In Proceedings of the 2011 ACM Conference on Ubiquitous Computing, Beijing, China, 17–21 September 2011; pp. 375–384. [Google Scholar] [CrossRef]
- Yumak, Z.; Pu, P. Survey of Sensor-Based Personal Wellness Management Systems. BioNanoScience 2013, 3, 254–269. [Google Scholar] [CrossRef]
- Rodgers, M.M.; Pai, V.M.; Conroy, R.S. Recent Advances in Wearable Sensors for Health Monitoring. IEEE Sensors J. 2015, 15, 3119–3126. [Google Scholar] [CrossRef]
- Shrestha, P.; Saxena, N. An Offensive and Defensive Exposition of Wearable Computing. ACM Comput. Surv. 2018, 50, 1–39. [Google Scholar] [CrossRef]
- Triantafyllidis, A.K.; Velardo, C.; Salvi, D.; Shah, S.A.; Koutkias, V.G.; Tarassenko, L. A Survey of Mobile Phone Sensing, Self-Reporting, and Social Sharing for Pervasive Healthcare. IEEE J. Biomed. Health Informatics 2017, 21, 218–227. [Google Scholar] [CrossRef] [PubMed]
- Im, H.; Castro, C.M.; Shao, H.; Liong, M.; Song, J.; Pathania, D.; Fexon, L.; Min, C.; Avila-Wallace, M.; Zurkiya, O.; et al. Digital diffraction analysis enables low-cost molecular diagnostics on a smartphone. Proc. Natl. Acad. Sci. USA 2015, 112, 5613–5618. [Google Scholar] [CrossRef]
- Coskun, A.F.; Wong, J.; Khodadadi, D.; Nagi, R.; Tey, A.; Ozcan, A. A personalized food allergen testing platform on a cellphone. Lab Chip 2013, 13, 636–640. [Google Scholar] [CrossRef] [PubMed]
- Tseng, D.; Mudanyali, O.; Oztoprak, C.; Isikman, S.O.; Sencan, I.; Yaglidere, O.; Ozcan, A. Lensfree microscopy on a cellphone. Lab Chip 2010, 10, 1787–1792. [Google Scholar] [CrossRef]
- Jablonsky, N.; McKenzie, S.; Bangay, S.; Wilkin, T. Evaluating sensor placement and modality for activity recognition in active games. In Proceedings of the Australasian Computer Science Week Multiconference, Geelong, Australia, 30 January–3 February 2017; pp. 61:1–61:8. [Google Scholar] [CrossRef]
- Vermeulen, J.; Willard, S.; Aguiar, B.; De Witte, L.P. Validity of a Smartphone-Based Fall Detection Application on Different Phones Worn on a Belt or in a Trouser Pocket. Assist. Technol. 2014, 27, 18–23. [Google Scholar] [CrossRef] [PubMed]
- Stemple, C.C.; Angus, S.V.; Park, T.S.; Yoon, J.Y. Smartphone-Based Optofluidic Lab-on-a-Chip for Detecting Pathogens from Blood. J. Lab. Autom. 2014, 19, 35–41. [Google Scholar] [CrossRef]
- Wang, C.; Wang, X.; Long, Z.; Yuan, J.; Qian, Y.; Li, J. Estimation of temporal gait parameters using a wearable microphone-sensor-based system. Sensors 2016, 16, 2167. [Google Scholar] [CrossRef]
- Le, T.H.; Canovas, C.; Clédiere, J. An overview of side channel analysis attacks. In Proceedings of the Asia CCS’08 ACM Symposium on Information, Computer and Communications Security, Tokyo, Japan, 18–20 March 2008; pp. 33–43. [Google Scholar]
- Trippel, C.; Lustig, D.; Martonosi, M. Checkmate: Automated synthesis of hardware exploits and security litmus tests. In Proceedings of the 2018 51st Annual IEEE/ACM International Symposium on Microarchitecture (MICRO), Fukuoka, Japan, 20–24 October 2018; pp. 947–960. [Google Scholar]
- Gruss, D.; Spreitzer, R.; Mangard, S. Cache template attacks: Automating attacks on inclusive last-level caches. In Proceedings of the 24th USENIX Conference on Security Symposium, Washington, DC, USA, 12–14 August 2015; pp. 897–912. [Google Scholar]
Existing Classifications Related to Side-Channel Use | Classifications Used in This Research |
---|---|
Side-channel attacks (relating to side-channel exploitation within CYB only) [14], side-channel sensing (relating to side-channel exploitation within CYB as well as other domains) [2]. | Side-channel sensing |
Target information (used within MDM and CYB [2] but also applicable to all domains). | Target information |
Cryptographic devices (attacks against cyber–physical devices) [14], target systems (within context of MDM and CYB) [1,2]. | Target system |
Biomarkers (objectively measured characteristics that lead to a diagnosis), side channel (a pathway in which target information traverses along within any context) [1,2,14]. | Side-channel |
Side-channel properties [2]. | Side-channel properties |
Modalities (a description of the target information and the signal within which it is embedded) [1]. | Information parameters |
Measurement setup [14], sensors [2]. | Sensors |
Leakage models (understanding of the signal within a side channel in a CYB context) [14], techniques [1,15], physical attacks [14], side-channel attack techniques [2]. | Methods and Extraction techniques |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Spence, A.; Bangay, S. Domain-Agnostic Representation of Side-Channels. Entropy 2024, 26, 684. https://doi.org/10.3390/e26080684
Spence A, Bangay S. Domain-Agnostic Representation of Side-Channels. Entropy. 2024; 26(8):684. https://doi.org/10.3390/e26080684
Chicago/Turabian StyleSpence, Aaron, and Shaun Bangay. 2024. "Domain-Agnostic Representation of Side-Channels" Entropy 26, no. 8: 684. https://doi.org/10.3390/e26080684