FSFarnaz Sabahi

Farnaz Sabahi

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Farnaz Sabahi has a total of 5 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Predictive safety framework using Kalman smoothing and LSTMs for robotic exoskeletons

    by Farnaz Sabahi

    In this paper, we develop a predictive safety control framework for robotic exoskeletons that integrates Kalman smoothing (KS) for noise reduction in sensor data, long short-term memory (LSTM)-based force prediction, adaptive impedance control, and dynamic barrier functions. The proposed approach is applicable to safety-critical tasks in uncertain environments and addresses important limitations of prior methods including disjointed perception control pipelines, rigid safety constraints, and lack of guaranteed stability. The main contributions of the proposed approach are: a unique KS-LSTM coupling to reduce interaction force prediction errors by 44.5% while operating in real-time (22.3 ms/cycle); adaptive control that modulates stiffness/damping parameters using LSTM-predicted forces; and context-aware barrier functions with formal globally uniformly ultimately bounded (GUUB) stability guarantees via Lyapunov analysis. Through extensive simulations, the proposed approach demonstrates superior performance over baselines in prediction accuracy, trajectory tracking, and constraint compliance (95.04%), especially in obstacle avoidance and human–robot collaboration. This framework connects learning-based prediction and real-time certifiable control, which helps to ensure safer physical human–robot interaction in rehabilitation, industrial, and assistive settings.

  • Open Access

    Adaptive inverse control for trajectory tracking with dead‐zone nonlinearity under cyberattacks

    by Farnaz Sabahi

    Control systems rely heavily on the accuracy and reliability of sensor data; however, the integrity of these data can be compromised through spoofing attacks, leading to significant modelling errors that can render control impractical. In addition, centralized control poses a significant threat to system security. To address these issues, a distributed framework is suggested for a discrete‐time nonlinear system that encounters unknown dead‐zones at its input. The framework uses the inherent resilience of a decentralized peer‐to‐peer network to secure information exchange, eliminating the need for prior knowledge of system dynamics or potential attacks. The proposed framework performs two complex tasks: identifying the nonlinear system and dealing with the unknown nonlinearity at the input in the form of a dead‐zone. An adaptive dead‐zone inverse is used to handle the unknown nonlinearity at the input in the form of a dead‐zone and integrate blockchain technology to secure communication between components. The blockchain component ensures tamper‐proof data transmission and resistance to cyberattacks, providing both detection and defence mechanisms without prior knowledge of system dynamics or potential attacks. The actuator and plant components are matched and synchronized using a private network with static nodes, ensuring deterministic and well‐coordinated communication. Simulation results demonstrate that the proposed framework both with and without blockchain integration, maintains stability and outperforms traditional methods in terms of robustness and accuracy, even when all parts of the framework are adjusted in response to attacks.


  • Open Access

    Analyzing Causality in Uncertain Domains using Extended Fuzzy Logic (FLe) Applied to Coronary Heart Disease Diagnosis

    by Farnaz Sabahi

    To extend the application of FLe in computing the degree of causality, we employ granular fuzzy causality tactics to determine the direction of causality among related variables when there is only imperfect information. Our approach involves a hierarchy of nested inferences of interval Type-2 fuzzy sets to achieve more approximate reasoning for more workable solutions, in the sense of f-valid philosophy. To deal with intrinsic hard uncertainty in the problem architecture, we leverage expert knowledge about the problem structure. Our method involves three key steps: encoding reasons into interval Type-2 fuzzy, utilizing the interaction of concepts in forward reasoning through qualitative descriptions and allowing a certain level of uncertainty, and determining the direction of valid results based on extended fuzzy logic. Our simulation results demonstrate the reliability of our proposed method when compared with traditional paradigms of precise reasoning. Overall, this work highlights the potential of FLe in causality problems and provides a basic framework for handling causality in uncertain domains.


  • Optimal bounded policy for nonlinear tracking control of unknown constrained-input systems

    by Farnaz Sabahi

    This paper introduces a novel algorithm that seamlessly integrates type-2 fuzzy systems with a sliding mode controller, aiming to create an optimal bounded control policy for tracking nonlinear problems that are plagued with uncertain or incomplete system dynamics and control input constraints. Proving its efficacy in navigating uncertainties, the proposed approach maintains effectiveness even when such encounters are sporadic or infrequent. The algorithm tactically employs three type-2 fuzzy systems. Among these, the actor and critic fuzzy systems are specifically tasked to resolve the optimal control tracking problem, while the third fuzzy system is designated to approximate the system’s unknown dynamics. The sliding mode controller’s role is instrumental in this setup. It dynamically adjusts to ensure the system’s convergence, enabling precise tracking of the desired trajectory, undeterred by the prevalent uncertainties. We validate the stability of the entire amalgamation, consisting of the actor, critic, identifier and controller. The robustness and efficiency of this innovative method are confirmed through rigorous simulation testing on a nonlinear system. Our results substantiate that the proposed solution excels in optimal tracking control, particularly in situations where system dynamics are uncertain or incomplete and where control input constraints are a critical factor.


  • A Design of mRNA-Based COVID-19 Vaccine Through Fuzzy Neural Network

    by Farnaz Sabahi

    The worldwide effort to develop a vaccine against SARS-CoV-2 has led to a revolution in vaccinology by introducing a completely new class of vaccines messenger RNA (mRNA) vaccine. The mRNA-based vaccine is a singular molecule made in the lab that teaches the cells to produce an antigen to trigger the immune response against the fake infection. However, new variants of SARS-CoV-2 may consist of an unprecedented set of genetic mutations including a sampling of earlier variants in addition to the other unknown mutations on the spike protein that may bind a part of the virus to human cells like a grappling hook. A common paradigm in designing a vaccine is to create a fixed architecture in the hope that it can make connections between the vaccine and mutations. In this paper, we propose a COVID-19 RNA-based vaccine in four modules: SARS-CoV-2 profile, mRNA-based vaccine design, interaction box, and neural codon optimization. We use epitopes' perception to collectively analyze mutations for designing mRNA-based vaccines and optimize the vaccine through neural codon optimization. In the proposed vaccine, the structural variation of the inhibitor is changed with the interaction of COVID-19 variants. To evaluate, the proposed vaccine is applied to the real data set. The results demonstrate that the proposed vaccine can provide high levels of protection against various virus mutations in comparison. Even with the challenging New3 mutation, the proposed vaccine still provided a good 78% protection with two doses of vaccination.


  • Introducing convex BiLSTM-based controller applied to 3-PSP spatial parallel robot manipulator

    by Farnaz Sabahi

    The idea of modeling first and then finding a control approach might have made sense when systems were changing on a vast scale, but not in today’s world where we are in danger of being run over on a highway called the information highway. The feedback phenomenon, which involves a system’s future being influenced by its past, presents a rational conflict in this context. However, we can move beyond these challenges by leveraging the system’s inherent ability to randomly repeat and re-create parts, allowing new features to emerge and the best repetition to become the standard. To address the mentioned challenges, this paper proposes a new structure based on bilateral long short-term memory (BiLSTM) to provide supervising feedback on the nonlinear components comprising friction, Coriolis, and Centrifugal forces in addition to gravity for adapting uncertain variations in the robotic system dynamics. A key characteristic of the proposed BiLSTM-based controller is the incorporation of memory. A new formulation is developed for determining when to recover old memories and to what extent, ensuring the controller remains both adaptive and efficient. Furthermore, given that for system identification, speed is vital; the proposed approach has dealt with the speed category by introducing a convex combination of BiLSTM layers. The stability of the whole structure has also been analyzed, and the updated learning rules have been derived. The simulation results show improved tracking performance and learning time compared to alternative approaches.


  • Open Access

    Power Control of a Grid-Connected Doubly Fed Induction Generator Using H∞ Control and Kalman Filter

    by Aboozar Mohammad Nezhad, Behrouz Tousi, Farnaz Sabahi

    A new control method for a grid-connected doubly fed induction generator (DFIG) is proposed in this paper, which is robust against parametric uncertainty and measurement noise. In general, the DFIG controllers can be divided into two main groups: the rotor side converter (RSC) and the grid side converter (GSC) controllers. The parameters of a DFIG may deviate from their rated values due to the operating conditions. For this parametric uncertainty, a robust H∞ vector control (VC) is employed using the complex sensitivity approach. The design of the RSC controller has been carried out using the vector control strategy, and instead of proportional-integral (PI) controllers, a designed robust controller is used. One of the steps in vector control is the extraction of the measured currents to be used in the control equations. If the currents are polluted with noise, the system control will be impaired. Thus, using a Kalman filter is suggested to solve this problem. The effectiveness of the proposed method is then investigated using extensive simulations under various conditions. The obtained results confirm the efficient performance and robustness of the presented controller with model and measurement uncertainties.


  • Adaptive Emotion-Centered Fuzzy Tracking Controller With Robust Outer Feedback

    by Farnaz Sabahi

    We develop a novel robust self-organizing construction based on two powerful patterns--the elegance of cognitive (brain emotional learning) and the influence of decision making (fuzzy logic). Early in the design of the proposed controller, the friction, coriolis, centrifugal, and gravity parts of the robot model are considered as they are uncertain, and then they are compensated by the proposed structure. Unlike the existing approaches, the proposed self-organizing fuzzy emotional (SFE) approach achieves not only optimum parameters but also optimum constructions. In structure learning, a new index based on the measured error, the number of rules, and the max sensory input is considered to generate new rules. The parameter learning condition depends on the stability. Furthermore, we formulate a controller based on SFE for a two-link robot and 3-PSP spatial parallel manipulator. To deal with the increasing and decreasing weight of Amygdala, the robust outer feedback is augmented to the whole controller. Simulation results show the priority of the proposed approach in comparison to alternative approaches especially in improving the ability to deal with uncertainty.


  • Risk assessment of COVID-19 based on a new structure of neural fuzzy probabilistic model

    by Farnaz Sabahi

    The risk assessment of the COVID-19 infection can save so many lives, reduce treatment costs, and increase public health. The unknown nature of the COVID-19 infection, the high impreciseness of available information, and not simply recognizing the relevant factors and their effectiveness may cause overestimating and underestimating of factors. This paper puts forward a development of a model with fewer limitations that are more consistent with progressive knowledge about COVID-19. Dealing with the situation of updating the statistical dataset daily, the proposed approach can effectively use the subjectivity inherent in the fuzzy probability interpretation of risk factors using expert knowledge in addition to the statistical dataset. Second, to this uncertainty handling improvement, a specificity-based parameter learning based on the learning network is also added to deal with the complexity aspect of the COVID-19 infection. The learning process helps the proposed structure better adjust the effectiveness of factors. From the achieved results, it is verified that people with advanced age, those with chronic obstructive pulmonary disease, lung cancer, and those having cancer treatments are at higher risk of death if they are infected by COVID-19. Undoubtedly, for vaccination, these three groups should be considered in order to prevent death situations.


  • FLeAC: A Human-Centered Associative Classifier Using the Validity Concept

    by Mahnaz Kadkhoda, Mohammad-R. Akbarzadeh-T., Farnaz Sabahi

    Fuzzy associative classifiers (FACs) have recently received considerable attention in the data mining community due to their ability to address the imprecision and graduality of truth. Similar to their more traditional statistical peers, these classifiers, however, have remained largely data driven, not leveraging human knowledge to their advantage. This is while human expert opinion and intuition should be a unique vantage point for such systems. We introduce here, for the first time, a human-centered framework (FLeAC) for FACs based on extended fuzzy logic and f-transformation that uses experts' opinions and preferences along with statistical data to solve subjective real-world problems. In FLeAC, experts take part in both constructing and reasoning of the classifier by assigning linguistic validity to each item. These validities are then aggregated using collective intelligence that determines final item validity. To examine the proposed framework, we extend an efficient and well-known FAC, CFAR, and present an extended f-CFAR algorithm. Also, several variations of f-CFAR are implemented to examine the effect of rule validity and different f-transformation operators. We then run various nonparametric statistical tests, including Friedman, Nemenyi posthoc, and ROC tests on an actual medical dataset of burn patients from Ahwaz, Iran, to compare f-CFAR performance with those of the original and nine other rule-based classifiers. Statistical analysis shows that f-CFAR not only has a better overall diagnostic performance than CFAR but also it outperforms CFAR and the other rule-based classifiers in terms of the number of rules, the number of conditions, and the execution time, leading to a more compact and comprehensible classifier with comparable accuracy.