MHMarijan Herceg

Marijan Herceg

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Marijan Herceg has a total of 25 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Multi-Layer Security for Wireless BCIs via Adversarial Chaos Synergy

    by Majid Mobini, Georges Kaddoum, Marijan Herceg

    While we benefit from the capabilities of Denoising Auto-Encoders (DAEs), some security risks also arise with the widespread adoption of these systems. Conventional chaos masking and jamming defenses do not protect against DAE-equipped eavesdroppers. Chaotic masking adds randomness to signals and helps with security, but it cannot stop detection by a DAE. To tackle this problem, we design a secure Brain-Computer Interface (BCI) using Long Short-Term Memories (LSTMs) that combines adversarial training and chaotic signals. We formulate an optimization problem to model the proposed BCI and generate a low-power perturbation vector. The results indicate that when the noise variance is equal to 0.2, the Root Mean Square Error (RMSE) of an illegal DAE exceeds 1.2. At the same time, the non-adversarial performance is approximately 0.5. Moreover, the performances of different Neural Networks are evaluated under the same computational complexity conditions. The best result is provided by the LSTM, with a complexity of about 3 × 104 Real Multiplications per Symbol (RMpS). The proposed Adversarial Chaos Synergy (ACS) framework is compatible with cloud/edge infrastructure. With these features, the ACS approach is suitable for applications such as mobile medical services, lightweight bio-medical instruments, brain-to-brain networks, and smart hospitals.


  • A Multi-Carrier Deep Learning CSK Communication System

    by Majid Mobini, Georges Kaddoum, Marijan Herceg

    A new Multi-Carrier Deep Learning Chaos Shift Keying (MC-DLCSK) modulation is designed in this paper. The proposed receiver includes a Neural Network (NN)-based classifier that recovers MC-DLCSK signals. The results demonstrate that the MC-DLCSK system achieves a considerable Bit Error Rate (BER) gain compared to all existing similar systems under both Additive White Gaussian Noise (AWGN) and Rayleigh fading channels.


  • Open Access

    Design of an M-Ary DLCSK Communication System Using Deep Transfer Learning

    by Majid Mobini, Marijan Herceg, Georges Kaddoum

    Conventional coherent chaos-based communication systems require synchronization of chaotic signals, which is still practically unattainable in a noisy environment. Moreover, in non-coherent schemes, a part of the bit duration is spent sending non-information-bearing reference samples, which deteriorates the Bit Error Rate performance (BER) of these systems. To tackle these problems, this paper designs an M-ary Deep Learning Chaos Shift Keying (M-ary DLCSK) system. The proposed receiver uses a Convolutional Neural Network (CNN)-based classifier that recovers M-ary modulated data. The trained NN model grasps different chaotic maps, estimates channels, and classifies the received signals effectively. Moreover, we consider a Transfer Learning (TL) framework that enhances the noise performance and classification results. Due to the generalization capabilities of TL, the trained NN can work in different Signal-to-Noise Ratio (SNR) conditions without the need for re-training. We compare the BER performance, complexity, and bandwidth efficiency of the M-ary DLCSK receiver with existing receivers. The results demonstrate that the M-ary DLCSK receiver is the first practical system that achieves the theoretical BER performance of the coherent CSK systems under Rayleigh fading channels. Moreover, the proposed system provides a considerable performance advantage compared to the existing DL-based receivers under Rayleigh fading channels. For example, the BER performance of 8-ary DLCSK shows a gain of 0.1 over the Long Short-Term Memory (LSTM)-aided DNN systems at the target Eb/N0=14dB. These features make M-ary DLCSK an attractive candidate for several applications, such as Massive Multiple-Input Multiple Output (MIMO), Vehicle-to-everything (V2X), Quantum, and optical communication systems.


  • Test environment code and test-case generators

    by Dino Brkic, Abraham Kostic, Marijan Herceg, Miroslav Popovic

    In this paper, a new implementation of code and test-case generators within the existing test environment generator (TEG) is described. In particular, source code, CAPL code and test-case generators are implemented using C++ programing language. The execution time of the implemented solution is compared with the existing solution which is written in Python programming language. The comparison results show a significant improvement of the C++ solution in the comparison with the Python version.


  • Performance of spatial digital image filtering on a real Advanced Driver Assistance System platform

    by Kristina Opacak, Marijan Herceg, Jelena Kovacevic, Istvan Papp

    In this paper, the implementation and optimization of different spatial filters on the realistic Advanced Driver Assistance Systems (ADAS) platform is performed. In particular, the moving average, high-pass, median and Gaussian spatial filters are implemented on the ADAS Alpha board. The optimization of the implemented filters is performed in the software and hardware domain on different Alpha board processors to improve the filter's execution times. After optimization, the performance of the implemented filters is measured and compared with the implementation on the personal computer. The given results show comparable execution times of the optimized solution.

  • Open Access

    Design of a SIMO Deep Learning-Based Chaos Shift Keying (DLCSK) Communication System

    by Majid Mobini, Georges Kaddoum, Marijan Herceg

    This paper brings forward a Deep Learning (DL)-based Chaos Shift Keying (DLCSK) demodulation scheme to promote the capabilities of existing chaos-based wireless communication systems. In coherent Chaos Shift Keying (CSK) schemes, we need synchronization of chaotic sequences, which is still practically impossible in a disturbing environment. Moreover, the conventional Differential Chaos Shift Keying (DCSK) scheme has a drawback, that for each bit, half of the bit duration is spent sending non-information bearing reference samples. To deal with this drawback, a Long Short-Term Memory (LSTM)-based receiver is trained offline, using chaotic maps through a finite number of channel realizations, and then used for classifying online modulated signals. We presented that the proposed receiver can learn different chaotic maps and estimate channels implicitly, and then retrieves the transmitted messages without any need for chaos synchronization or reference signal transmissions. Simulation results for both the AWGN and Rayleigh fading channels show a remarkable BER performance improvement compared to the conventional DCSK scheme. The proposed DLCSK system will provide opportunities for a new class of receivers by leveraging the advantages of DL, such as effective serial and parallel connectivity. A Single Input Multiple Output (SIMO) architecture of the DLCSK receiver with excellent reliability is introduced to show its capabilities. The SIMO DLCSK benefits from a DL-based channel estimation approach, which makes this architecture simpler and more efficient for applications where channel estimation is problematic, such as massive MIMO, mmWave, and cloud-based communication systems.


  • Energy-Efficient Data Transmission for Capacitive-Coupled Human Body Communication Systems

    by Luka Filipovic, Marijan Herceg, Jelena Vlaovic, Ertugrul Basar

    In this letter, an energy-efficient and low complexity method for capacitive-coupled human body communication (CCHBC) systems is proposed. In the proposed method, called impedance shift keying CC-HBC (ISK-CC-HBC), the change of impedance between the transmitters signal and the ground electrode is used to map information. In particular, the transmitter changes the condition of an electrical field generated by the receiver, by changing the impedance between its signal and ground electrodes. Subsequently, by detecting the changes in the electric field, the receiver can demodulate the information sent by the transmitter. The CC-HBC channel is obtained using the transfer function method, while the performance of the proposed scheme is compared with frequency shift keying (FSK) and on-off keying (OOK) modulations.


  • A Tunable CMOS IR-UWB Pulse Generator Based on Feedback Controlled Oscillator Switching

    by Leon Sneler, Tomislav Matic, Marijan Herceg

    The paper presents a tunable Impulse-Radio Ultra-Wideband (IR-UWB) pulse generator based on feedback oscillator switching. The circuit provides a wide range of pulse amplitude and width tuning, thus enabling the required pulse shaping for a desired pulse repetition rate and/or a wireless transmission range. Feedback controlled switching prevents the potential omission of the UWB pulse generation and provides low UWB pulse energy variations at low bias currents and a wide temperature range, ensuring system robustness with respect to temperature variations. The pulse amplitude is tunable within 0.27 Vpp to 2.05 Vpp, while the pulse width tuning range is 1-6.5 ns. The total energy consumption of the IR-UWB pulse generator at the 50 Mpps pulse repetition rate is equal to 25.4 pJ/pulse for the pulse amplitude equal to 0.27 Vpp, while for the pulse amplitude equal to 2.05 Vpp, it is increased up to 99.8 pJ/pulse. The circuit occupies a core area of 0.56 mm in a 0.18 μm CMOS technology.


  • A Framework for the Lower Bound on the BER of DCSK Systems Over Multi-Path Nakagami-m Fading Channels

    by Mohamed Dawa, Georges Kaddoum, Marijan Herceg

    In this paper, a generalized closed-form approach for the evaluation of the lower bound on the BER of differential chaos-shift-keying (DCSK) systems over multi-path Nakagami-m fading channels is presented. A closed-form lower bound is derived for all configurations of the Nakagami-m channel model which extends to all fading profiles covered by the Nakagami model. In contrast to previous works where very specific models were targeted and only numerical analyses of the DCSK systems were presented, we provide an approach that analytically unifies the performance analysis of DCSK systems over more generalized fading models. A similar analytical development is generalized for Quadrature Chaos Shift Keying (QCSK) and Multi-carrier Differential Chaos Shift Keying (MC-DCSK) systems. A lower bound is applied on the Q-function to extract an exponential expression of the BER. Afterwards, a bound on the exponential function is used to derive a simple polynomial expression which is easily integrable using basic integration techniques for all multipath Nakagami-m fading channels. The final expression of the newly derived lower bound is forwarded in terms of the Gaussian hypergeometric functions. Comparative simulations conducted for the different systems under consideration show a tight gap between the BER of the simulated systems and their respective lower bounds.


  • An Energy Efficient Multi-User Asynchronous Wireless Transmitter for Biomedical Signal Acquisition

    by Tomislav Matic, Leon Sneler, Marijan Herceg

    The paper presents a novel transmitter architecture for short-range asynchronous wireless communication, applicable to simultaneous multi-user wireless acquisition of biological signals. The analog signal, provided from an analog biosensor, is transformed to time information using an Integral Pulse Frequency Modulator (IPFM) as a Time-Encoding Machine (TEM). The IPFM generates a time-encoded unipolar pulse train, maintaining the linear dependence of the output pulse distance on analog input voltage. The system enables continuous acquisition of the signals from multiple sensors in which each transmitter has unique feedback loop delay used for multiuser coding. IPFM pulses trigger the Impulse Radio Ultra-Wideband (IR-UWB) pulse generator directly, providing two Ultra-Wideband (UWB) pulses per each IPFM pulse. Due to the lack of internal clock signal and microprocessor-free multiuser coding, the circuitry satisfies the requirements of multi-user coding energy efficiency and size reduction, which are crucial demands in biomedical applications. The proposed Time-Encoded UWB (TE-UWB) transmitter is implemented in 0.18 μm CMOS technology. Measurement results of the IPFM transfer function for input voltage ranging from 0.15 V to 1.5 V are presented, providing the dependence of the IPFM pulse time distance on analog input voltage and power consumption dependence on the input voltage level. For continuous monitoring operation total power consumption of the transmitter circuitry for the maximum input voltage is 10.8 μW, while for the lowest input voltage it increases to 40.48 μW. The circuit occupies 0.14 mm $^2$ .