• Nenhum resultado encontrado

Mihai V. Micea

N/A
N/A
Protected

Academic year: 2023

Share "Mihai V. Micea "

Copied!
113
0
0

Texto

Platforma CORE-TX (Collaborative Robotic. Environments – The Timisoara Experiment) a fost concepută și implementată la nivel de prototip pentru studiul și dezvoltarea de sisteme în timp real, aplicații distribuite pentru percepție artificială, rețele de senzori inteligenți și medii robotice colaborative. Collaborative Robotic Environments - The Timisoara Experiment) a fost realizat în faza de prototip pentru studiul și dezvoltarea de sisteme în timp real, aplicații de percepție artificială, rețele de senzori inteligenți și medii robotice colaborative.

Introduction and Fields of Expertise

Adapted versions of the classical real-time scheduling algorithms Earliest Deadline First (EDF) and Minimum Ease First (MLF), with a focus on non-preemptive operation: EDFNP and MLFNP, respectively. Development and demonstration of a fully predictable hard real-time scheduling technique, FENP (Fixed Execution Non-Preemptive), able to solve the classic execution jitter problem in real-time systems.

Post Doctoral Research and Results

Based on the CORE-TX platform and other robotic systems, such as the LEGO Mindstorms NXT, the complex topic of robot positioning and location management has been addressed by our research team in the DSPLabs. Based on the previous research and results in the field, a location management and robot positioning methodology [A38] was developed in the DSPLabs by our team.

Head of the DSPLabs Research Group

Currently, our paper, including the most relevant results of this research, has been cited by more than 12 scientific articles, published by other authors in major journals or conference proceedings in the field. Dan Chiciudean) and an average of 10  16 students (undergraduate, master's and PhD), who usually collaborate for more than 2 years on R&D projects in the laboratory.

Initiator and Manager of Partnership Programs

Partnership with Continental Automotive Romania in real-time operating systems and tools for the automotive industry started in 2013: coordination of R&D contract [G7], creation of "UPT  Continental RTS Group". I also maintain a good collaboration that includes co-authorship of articles, joint research and development projects and mobility campaigns with State University of New York, Stony Brook, USA (Prof. Alexandru Doboli), Salzburg Research, Austria (Prof. Ulrich Hofmann), Stefan University cel Mare from Suceava (Prof. Constantin Filote [G5], [G6]).

Manager of R&D Grants and Projects

The FALX-DACIAE project addresses the fields of multimedia, mobile applications, multi-core architectures, digital image and video processing, real-time systems, parallel processing, compilers and software tools. The main goal was to research different types of advanced software applications for modern multi-core systems, with a focus on the field of multimedia in mobile phones, and to develop suitable semi-automatic tools and processes to increase the productivity of such applications.

Course Development and Teaching Activity

ADSP: Advanced Digital Signal Processing", 1st/3rd semester, Master of Computer Engineering and Master of Information Technology, Faculty of Automation and Computing. 34;ADSP: Advanced Digital Signal Processing", 1st/3rd semester, Master of Computer Engineering and Master of Information Technology, Faculty of Automation and Computing [M11].

Scientific Referee

34;RT: Research Topics in Computer Systems/Information Technology", semester 1/3, Master of Computer Engineering and Master of Information Technology, Faculty of Automation and Informatics. UPT Scientific Bulletin: Transactions on Automatic Control and Computer Science, Polytechnic University of Timisoara, Romania, ISSN 1224-600X.

Management and Administrative Experience

Department Public Image Manager, Computer and Software Engineering Department, Politehnica University of Timisoara, Romania (2000 - present). Member of the Central Board for Bachelor Entrance Examination, Politehnica University of Timisoara, Romania (1996  present).

Affiliations, Awards and Recognition

The most difficult problem of the HARETICK inter-task communication and synchronization mechanism is to ensure the correct communication between the HRT and the SRT contexts. Measurement of the Layer 3 Receive SRT task (top), the SPI Receive ModX and the HSCD hard real-time scheduler (bottom) (source: [A21]).

Figure 2-1. Specifications of the data communication and processing application  (source: [A21])
Figure 2-1. Specifications of the data communication and processing application (source: [A21])

Hybrid Scheduling Techniques to Improve the Efficiency of Highly

In this case, hard real-time task scheduling is performed only by the FENP component of the H2RTS algorithm. The ROM (Support and Operation / Robotics Module) manages the mobility and activation functions of mobile WITs.

Figure 2-5 illustrates an example of H 2 RTS task scheduling and execution within  the three contexts defined by the proposed method
Figure 2-5 illustrates an example of H 2 RTS task scheduling and execution within the three contexts defined by the proposed method

Predictable Data Communication System for Smart Sensors in Hard

This research started during the development process of the CORE-TX platform [A9], with additional support from R&D grants [G2] and [G3]. Finally, geometric methods use these parameters to calculate the position of the robot in the room. The proposed robot algorithm, [A40], is the first stage of a more complex procedure, which determines the relative distance between two mobile nodes (WITs) of the CORE-TX platform [A9], with the ultimate goal of calculating the location of the nodes.

To support the location control of the WITs, their Perception/Sensor Module (SEM) is specially designed with an extension tower containing two ultrasonic transducers (see Figure 4-1). On the one hand, rotation of the entire robot is avoided during the adjustment process. In the final phase of the alignment (Pre_align == 0 at line 22 of Listing 4-1), the algorithm traces back the position with the highest measured values.

Figure 3-5. PARSECS communication protocol stack (source [A27]).
Figure 3-5. PARSECS communication protocol stack (source [A27]).

MTDOA Inter-Robot Distance Measurement Technique

Our research team performed a large number of experiments and tests in the DSPLabs to evaluate the performance of the proposed MTDOA distance measurement technique. The difference between the packet reception times for the XBee modules of the two robots involved in the distance measurement procedures was measured with a logic analyzer. This difference varies randomly each time, mainly due to the operation of the XBee modules and the associated protocol stack.

This is another indication of the dominant contribution of XBee operating delays affecting the results.

Table 4-2. Distance measurement results for the MTDOA method (source [A26]).
Table 4-2. Distance measurement results for the MTDOA method (source [A26]).

CTOF Inter-Robot Distance Measurement Method

To evaluate the performance of the proposed CTOF method, a series of experiments were performed in the DSPLabs. This is a direct result of the independence of this technique from the random delays introduced by the wireless modules. Overall, the experimental results indicate that the CTOF method, with its accuracy of 4.8 cm for distances of 3 m and its linear behavior, outperforms the MTDOA and other similar techniques used in state-of-the-art site monitoring systems.

To further improve the accuracy of the CTOF distance measurement method, a Kalman filter was applied to various number of repeated measurements for the same distance to be determined.

Figure 4-14. Operating principle of the CTOF method (source [A26]).
Figure 4-14. Operating principle of the CTOF method (source [A26]).

Collaborative Localization Methodology for Autonomous Robots

The localization reliability,  N, measures the importance and reliability of the position coordinates of a network node. Therefore, in conjunction with the hardware subsystem, the software application layer has a significant impact on the power consumption [B56] of the system. In another set of experiments, the energy consumption signatures were measured and evaluated considering the execution of the same workload of the algorithm on a different number of threads.

Based on the proposed framework, we have shown how multithreading applications affect the power consumption of the single and multi-core battery-powered mobile devices.

Figure 4-18. Case study robotic environment for the collaborative localization  methodology (source [A38])
Figure 4-18. Case study robotic environment for the collaborative localization methodology (source [A38])

Battery Management System and Methodology for Online State-of-

This SoH estimation principle adopted for BMS integration is based on previously stored measurements of the total available discharge capacity of the battery pack. The algorithm uses the entire charge and discharge history of the battery, up to the current cycle. The nmin value is equal to the size of the moving time window defined by the SoH estimation algorithms.

Several approaches have been used to evaluate the performance of the proposed SoH estimation algorithms.

Figure 5-5. Proposed electrical model for the Ni-MH battery (source [A31]).
Figure 5-5. Proposed electrical model for the Ni-MH battery (source [A31]).

Relevant Works Published by Mihai Micea

Cretu, Voicu Groza, "Voorspelbare datacommunicatie-interface voor harde real-time systemen", in Proceedings of the International Workshop on Robotic and Sensors Environments, ROSE 2008, IEEE, Ottawa, Canada, 17-18 oktober 2008, pp. Petriu , "Fish Shoal Inspired Movement in Robotic Collectives", in Proceedings of the International Workshop on Robotic and Sensors Environments, ROSE 2008, IEEE, Ottawa, Canada, 17-18 oktober 2008, pp. Cretu, Voicu Groza, "Emergent Exploration and Resource Gathering in Collaborative Robotic Environments", in Proceedings of the International Workshop on Robotic and Sensors Environments, ROSE 2008, IEEE, Ottawa, Canada, 17-18 oktober 2008, pp.

34;Foodball: Emergent resource gathering in collective robotic environments", in Proceedings of International Workshop on Robotic and Sensors Environments, ROSE 2011, IEEE, Montreal, Canada, September 17-18, 2011, p.

Papers Published by Other Authors in the Field

Cretu, "Energy Efficiency in Real-Time Systems: A Brief Review," in Proceedings of the 8th IEEE International Symposium on Applied Computational Intelligence and Informatics, SACI 2013, IEEE, Timisoara, Romania, May 23-25, 2013, p. 34; A Simulator-Based Study of Robot Alignment and Localization", in Proceedings of the International Symposium on Robotic and Sensor Environments, ROSE 2013, IEEE, Washington, USA, October 21-23, 2013, pp. Martinoli, "Indoor Navigation Survey with the Khepera III Mobile Robot : An experimental basis with a case study of ultra-wideband positioning", in Proc.

Kuc, "A simple ultrasonic GPS system for indoor mobile robotic system using Kalman filtering", in Proc.

Relevant R&D Grants, Contracts and Projects

Cretu (co-director), Artur Kuczapski, Mihai Novac, Horia Ciocarlie, „PERL: Modeling and Simulation of Automated Manufacturing Equipment and Process Planning Techniques in Semiconductor Manufacturing”), Contract de cercetare-dezvoltare si implementare, 2007, Advanced Clean Production Information Technology SRL. Cretu (co-director), Delia Golcea, Mihai Fagadar-Cosma, Cosmina Chise, et al., „Geodata Platform: Application for Handling Geographical Information - GIS AURORA Platform Geodata: Application for Handling Geographical Information - GIS”), R&D and implementation din Acord, 2007, Alcatel-Lucent Romania SA., Timisoara, Romania. Micea, et al., „MERLIN: Platformă IT pentru modelare, simulare, planificare și optimizare automată a producției în industria semiconductoarelor MERLIN: Platformă IT pentru modelare, simulare, planificare și optimizare automată a producției în industria semiconductoarelor”), Grant R&D, PNCDI II, Programul 5 - Inovaţii, PDP CNMP, Ministerul Educaţiei, Cercetării şi Inovării, Bucureşti, România.

Cioarga, Bogdan Ciubotaru, „DSPLABS: Dezvoltarea unui Laborator Multidisciplinar de Procesare Digitală a Semnalului, Multimedia, Telecomunicații Digitale, Rețele de Senzori și Robotică”), Proiect de Dezvoltare, Program Laboratoare Licențiate, AC Ministerul Educației, Cercetării și Inovării și Universitatea Politehnică din România. Timisoara, Romania.

Patents Pending for Inventions

Relevant Monographs, Book Chapters and Didactic Materials

Micea, Andrei Stancovici, Sinziana Indreica, "Distance Measurement for Indoor Robotic Collectives", in Mobile Robots - Control Architectures, Bio-Interfacing, Navigation, Multi Robot Motion Planning and Operator Training, section 16, Janusz Bedkowski (Eds.), InTech, Rijeka, Kroasië, Des., 2011, pp. Micea, Valentin Stangaciu, Cristina Stangaciu, Constantin Filote, "Sensor-Level Real-Time Support for XBee-Based Wireless Communication", in Advances in Intelligent and Soft Computing (AISC), volume 145, afdeling 20, Ford L. Micea, "Dense and Sparse Optic Flows Aggregation for Accurate Motion Segmentation in Monocular Video Sequences", in Lesingnotas in Rekenaarwetenskap (LNCS), volume 7324, afdeling 25, Aurelio Campilho, Mohamed Kamel (Reds. .), Springer Berlin Heidelberg, Aveiro, Portugal, Jun., 2012, pp.

Prizes, Awards and Recognition

Imagem

Figure 2-1. Specifications of the data communication and processing application  (source: [A21])
Figure 2-2. HARETICK data and program memory map for the application HRT  tasks (source: [A21])
Figure 2-6. Performance of the H 2 RTS hybrid scheduling, compared to the FENP  algorithm (source: [A37])
Figure 2-7. Caption of task execution and timing measurement for the wireless  communication system (source: [A37])
+7

Referências

Documentos relacionados

A maior frequência foi para realização de avaliação odontológica preventiva 75,0% e para avaliação pré-cirurgia cardí deManda do serViço de odontologia clínica do HosPital santa