In the context of systems diagnosis, communication systems, and embedded signal processing, Intelligent Signal Processing (ISP) is used to develop intelligent systems that can diagnose faults or anomalies in complex systems, such as industrial systems. ISP is used to optimize the performance of communication systems by analyzing signal quali-ty, optimizing network resources, performing channel estimation and predicting traffic patterns. Ultimately, ISP is used to develop intelligent embedded systems that can pro-cess signals in real-time, such as in intelligent robots, autonomous vehicles, medical de-vices, and other embedded applications. The following paragraphs describe my own expertise followed by four cutting-edge projects in applied AI.
Contribution in Expertise and Research Projects in RC-AAI
My areas of expertise are in electrical engineering and align with the strategic lines of research in the information and communication technologies (ICT) and industrial sys-tems control especially in signal processing, image processing, computer vision, intelli-gent diagnosis of electrical networks, development of intelligent systems in electrical engineering, rapid prototyping of embedded electronic systems, wireless communica-tion, cognitive networks, wireless sensor networks, industrial Internet of things, Cyber-physical systems, industrial informatics, systems control & integration, multi-sensor and sensor arrays information fusion, human-machine interfaces, electronics and data se-curity, also in artificial intelligence such as swarm intelligence/ evolutionary algorithms.
My research activity is predominantly in the field of signal processing and intelligent control where novel kinds of algorithms are designed to process and extract key-features information and where artificial intelligence is used for optimization and classification in preprocessing and/or post-processing. I also have an advanced expertise in electronic systems hardware design and implementation including smart sensors, inter-net of things (IoT) and wireless sensors networks (WSN) devices. Applications are numerous such as in wireless communications, intelligent control, power systems diagnosis and healthcare monitoring. My work aims at comprehensive exploration of new technologies to enable the design and analysis of new wireless communications, adaptive and optimal control DSP systems. While introducing novel DSP algorithms and systems design, I focus on addressing the gap between the design prototype models and implementation of real-world systems which opens the door to industrial collaboration, funding and implementation.
To date, my research has included Artificial Intelligence in DSP applications for array signal processing, smart antennas, Wireless Sensor Networks, IoT devices and cyberphysical power systems control and diagnosis. My plan is to start and develop an initial research program, with four tracks, aimed to implement new applications of security in Cyberphysical systems with Artificial intelligence for the diagnosis and control of indus-trial power systems, wireless channels estimation in cognitive radio networks, and novel signal processing techniques for model-free power systems diagnosis. The objective is to set a solid platform in both instrumentation and algorithms that targets applications and results at the leading edge of research. The four tracks are listed below and sum-marized later:
1. Diagnosis and Security for Cognitive Cyberphysical Systems with Embedded Arti-ficial Intelligence
2. Time-Frequency Mixed Domain Signals Analysis for Electric Unbalance Estimation of Model-Free Power Cyber-Physical Systems: The experimental validation will be based on the setup of a new power network analyzer (hardware device) and signal processing based on novel findings.
3. Joint Timing Synchronization and Wireless Channel Estimation in OFDM for Cogni-tive Radio Networks using Intelligent Learning Control.
4. Intelligent mobile robotics with Internet of thing for Pavement cracks classifica-tion.
All four tracks merge technologies in intelligent signal processing, wireless communi-cations, IoT, WSN and artificial intelligence powered instrumentation targeting the in-dustrial, scientific and medical sectors.
Main Expertise Summary
1. Intelligent Signal Processing and Intelligent Control applied to sensor arrays, radar systems, healthcare and electrical machines control and diagnosis.
2. Mobile robotics and artificial vision: auto-navigation, wireless communication control and pattern recognition.
3. Design, development and Implementation of Wireless Sensor Networks and Indus-trial Internet of Things electronic devices for Industrial and Healthcare monitoring and diagnosis systems (sensors electronic circuit design & implementa¬tion).
4. Design and development of smart sensors electronic systems, interfacing of ana-log and digital sensors, signal processing and transmission, TCP/IP communica-tion.
5. Optimization with multi-objective functions: Swarm Intelligence algorithms (PSO, QPSO, BPSO), Genetic Algorithms, Iterative learning Control Optimization.
6. Electronics and Power Electronics systems design and implementation, Power and energy systems control, Wireless Communications, Signal Processing, Em-bedded Systems, Smart-grid, Cyberphysical Systems.
7. Biomedical signals processing, spectral analysis and features extraction
8. Cybersecurity for Cyber-Physical Systems including lightweight authentication protocols, block and streaming ciphers. Computer/ Network Security with cryp-tographic and steganographic algorithms.
9. Software development with hardware interfacing using Delphi and C++ Builder.
10. Signal and Image Machine Learning/ Deep Learning artificial intelligence algo-rithms with SciLab, Matlab and Python.
First Topic (Security for Cyberphysical Systems with AI Diagnosis)
Diagnosis and Security for Cognitive Cyberphysical Systems/ Smart-Grid with Embed-ded Artificial Intelligence. Abstract— Due to the progress in 5G networks bringing the next generation of mo-bile internet connectivity, Industrial Internet of Things (IIoT) transfor¬mation is accelerat-ing in an un¬prec¬edented way, that will help create a huge rise in Internet of things technology. This work is about the design and implemen¬tation of a cognitive cyber-physical system interface/ IIoT authentica¬tion protocol permitting to secure the access to a web-server controller inte¬grated inside an IIoT device. The authentication is based on a Challenge-Response topology from within a webpage generated by the IIoT de-vice, where a hash algorithm applied to a pre-shared key, is implemented on the ad-ministrator-side, while the IIoT server-side imple¬ments the verification part to grant ac-cess to the administra¬tor. An extra security layer permits a single instance, which locks the device disabling any access attempt, until the device is released by the cli-ent/administrator. The IoT device will be set to transmit multisensory data and receive control commands in a per¬ceived real-time to act on industrial systems. A neural net-work architecture for cyber-physical industrial three-phase motors conditions and faults diagnosis will be used as an artificial intelligence of things (AIoT) imple¬mented on MCUs. The first phase of the neural network architecture is for diagnostic purpose and faults detection; however, this concept is to be extended in a second phase for cyber-security analytics for proactive approach allowing for earlier threat detection and au-tomation of security measures. This work is about the implementation of a microcontrol-ler-based embedded system interface, as an edge computing, to convert a physical system to a Cyberphysical system with embedded network security. The architecture implementing an artificial intelligence diagnosis and faults detection system will be achieved on more powerful computing device as fog computing. This will demonstrate how an IIoT/CPS interface design and algorithms may be implemented and tested in realistic conditions which allows its use in wide range of industrial applications.
Index Terms— IIoT, AIoT, Network Security, Cybersecurity, Cyber-Physical, Authentica-tion, Client/ server, Web-Server.
Second Topic (Signal Processing, AI Diagnosis of Power Systems)
Time-Frequency Mixed Domain Signals Analysis for Electric Unbalance Diagnosis of Model-Free Power Cyberphysical Systems. Abstract: This research track will be based on a novel signal processing technique for the estimation of three-phase power systems electric unbalance factor without the need for the power system model (Model-free). The novel method is named radial Cur-rent Unbalance Factor (rCUF). It is based on reducing the three-phase currents into one dimensional signal for time-frequency analysis. The radial signal is defined in the sta-tionary reference frame currents. The rCUF combines high accuracy estimation along with model-free feature which makes it a universal metric for the diagnosis of any three-phase power system. This computational approach will be implemented as a smart-grid monitoring and diagnosis tool for performance analysis of power systems since it will have the ability to estimate unbalance, harmonics distortion and over/under voltage. In terms of spectral sensitivity, the rCUF will be able to detect very low unbalances in frequency domain using high Signal to Quantization Noise Ratio (SQNR). As a case study, the proposed rCUF approach will be applied to a three-phase induction machine (IM) for unbalance factor diagnosis and healthy state control. The importance of this research lies in the fact that it allows remote diagnosis and hence it is appropriate for Cyberphysical systems in smart grid. This work targets the power sys-tems industry and smart-grid and Cyberphysical systems technology providing tools for systems network-based diagnosis and control.
Index Terms— Time-Frequency, Signal Processing, Unbalance, Model-free.
Third Topic (Signal Processing based AI in Cognitive Wireless Net-works)
Joint Timing Synchronization and Wireless Channel Estimation for OFDM Cognitive Radio Networks using Intelligent Learning Control. Abstract: In emerging wireless communication technologies, the cognitive radio is the capability of wireless devices to reconfigure their internal parameters in order to exploit and adapt to the available channels and data spectrum in an intelligent and automatic manner. For this end, frequency selective fading channel estimation consti-tutes an appropriate wireless channels frequency response model used to implement such new technology. In addition to the channel estimation, the proposed approach will also provide synchroniza¬tion and equalization frame. In this work we will use a newly introduced Intelligent Learning Control algorithm called RIO that works under the cate-gory of artificial intelligence and signal processing algorithms, where stochastics is de-fined as a random spread-space search and discovery in a stochastic space that have a large number of optimal solutions. Defined as such it provides an efficient ap-proach to highly non-linear problems solving. The recurrence lies in the fact that once an initial solution is set in a stochastic space, this initial solution is taken and its parame-ters are recursively optimized or filtered in order to minimize the objective function de-fining the goodness of the solution. It is a robust intelligent deterministic/ stochastic evolutionary computation technique that uses its experience to learn and improve it-self.
Index Terms— Channel Estimation, OFDM, Cognitive Radio, Optimization.
Fourth Topic (Intelligent Mobile Robotics with Internet of Things)
Intelligent mobile robotics with Internet of thing for Pavement cracks detection, classifi-cation and reporting. Abstract: This advanced research project endeavors to realize the implementation of an intelligent mobile robot, designed to excel in a multitude of field missions encom-passing exploration, monitoring, data collection from sensor networks, and more. Among its array of missions, a specific focus lies in pavement analysis, crack detection, and the crucial decision-making process for pavement repair, all driven by cutting-edge image processing techniques. The mobile robot, remotely controlled via a highly secure wireless communication link based on a 256bits symmetric algorithm (AES, Ge-netic Cryptography are envisaged), boasts an integrated Internet of Things (IoT) mod-ule for precise navigation. The wireless communication control encompasses both long-range IoT at 2.4GHz through 4G mobile communication and a relative short-range 100MHZ Frequency Hoping Spread Spectrum multi-channel communication system. Furthermore, it harnesses an autonomous navigation system grounded in field explora-tion, driven by an ensemble of sensors including cameras, ultrasound, infrared, lasers, GPS and Gyroscope. The core intelligence of this remarkable mobile robot relies on quantized swarm intel-ligence and the innovative Recursive Intelligent Optimization algorithms for route plan-ning, marking a significant advancement in the realm of robotics. This research en-deavor merges diverse engineering disciplines, including intelligent control, sensor ar-ray signal processing, electronics & communications, mobile robotics, mechanical en-gineering, and civil engineering. Beyond its primary mission, the mobile robot harbors the potential for an array of applications, such as environmental monitoring, vegeta-tion monitoring, field surveillance and beyond, signifying its broader impact and signif-icance in multifaceted domains. This project represents a pioneering leap towards achieving enhanced efficiency and precision in autonomous robotic systems, promis-ing transformative implications across various sectors.
Index Terms— Intelligent Robotics, Image Processing, Autonomous Navigation, Signal processing, Optimization.