direkt zum Inhalt springen

direkt zum Hauptnavigationsmenü

Sie sind hier

TU Berlin

Inhalt des Dokuments

Dr.-Ing. Vlado Handziski

Lupe

I am senior researcher at TKN, coordinating the research and teaching activities of the group in the areas of networked embedded systems, internet of things and cyber-physical systems.

During the winter semesters 2014/2015 and 2015/2017 I was interim professor at the Chair for Embedded Systems at TU Dresden.

I received my doctoral degree in Electrical Engineering from TU Berlin (summa cum laude, 2011) and my M.Sc. degree from Ss. Cyril and Methodius University in Skopje (2002).

I have participated and led research activities in more than ten projects with external funding, at European, national and local level and have contributed to international standardization activities in the area of evaluation of indoor localization systems.

My research focus lies on developing platform solutions and system abstractions for wireless networked embedded systems and cloud-supported Internet of Things (IoT) as well as their application in specific application domains like industrial automation and indoor localization. In my research methodology, prototyping, experimental work and testbeds play a central role. I was chief architect of the popular TWIST testbed and one of the core developers of TinyOS.

 

CVResearch | Publications | ProjectsTeaching | ServiceGoogle Scholar

Research Profile

My research efforts have led to significant contributions in multiple research areas, briefly outlined below (references relate to the PDF version (PDF, 495,3 KB) of the publication list).

 

Platforms and System Abstractions

One of the core challenges for the design of runtime environments for networked embedded devices is to balance between the conflicting needs of extracting maximal capabilities from the underlying resource-constrained hardware while promoting portability, interoperability and convenient application development.

As part of my PhD research, I have proposed a novel software architecture for wireless sensor networks that achieves the above goals by concentrating on the organization principles for the hardware abstraction code and by providing a convenient high-level publish/subscribe interface promoting interoperability and streamlined application development [T1]. As a core developer of TinyOS and in close cooperation with UC Berkeley, I led the integration of these architectural concepts as part of the TinyOS 2.x Hardware Abstraction Architecture (HAA) ([C25], [NR5], [TR6]), enabling fine control of the trade-offs between portability and energy-efficiency and simplifying the integration of new hardware platforms. Our related work on the concept of “power-locks” [C22] that tightly integrates the process of access arbitration for a shared hardware resource and its (implicit) energy-management, has highlighted the potential and limits of fully automatic energy-management through centralized resource control in embedded operating systems. The viability of the concept was confirmed by the subsequent broad application of similar ideas in other domains, for example the use of “wake-locks” in Android.

With the help of these architectural features, TinyOS 2.x was ported to tens of sensor node platforms with MCUs ranging from 8-bit Atmel AVRs to 32-bit ARM Cortexes. It was the dominant runtime environment for sensor networks research, reaching more than 25,000 annual downloads at its peak. It was also used in numerous commercial products. 

 

Experimental Methodology

In the early days of sensor networks research, the evaluation practice was dominated by bench-top experiments with just a handful of nodes and simulations that frequently made very artificial assumptions about the connectivity, traffic, failure patterns, etc. These approaches were not able to capture the challenges associated with real-life deployment scales of several hundred nodes. To address these limitations, I have lead the development of the TKN Wireless Indoor Sensor Network Testbed (TWIST), one of the first and most popular remotely-accessible indoor wireless sensor network testbeds ([C23],[C24]). 

In the years since it was deployed, TWIST has become an indispensable resource for the research community, as witnessed by the large number of papers published at top conferences and journals using experimental data collected on TWIST: SenSys 2009 2x, EWSN 2010 3x, IPSN 2011 1x, ACM TOSN 2011 1x, IPSN 2012 2x, SenSys 2012 2x, NSDI 2013 1x, SenSys 2013 1x, IPSN 2014 1x, etc. TWIST has also been replicated by other academic institutions like WUSTL and companies like Siemens. 

Through my participation in several projects and activities, part of the FIRE initiative and EIT Digital, I have recently led the transformation of TWIST from a pure wireless sensor testbed, to a testing environment supporting more complex and heterogeneous CPS scenarios, enabling concurrent experimentation with different wireless technologies, robotic mobile platforms and building automation sensors and actuators ([PD5],[PD6],[PD7]). 

TWIST and similar testbeds like MoteLab at Harward, and the more recent offerings like Indriya at NUS and IoTLab at INRIA, have contributed to a fundamental change in the evaluation methodology in our community. Thanks to our pioneering work, large-scale experimental evaluation on one of the remotely accessible testbeds is now commonly seen as precondition for serious consideration of any protocol or system proposal in the domain of wireless sensor networks. I am recognized expert in the domain and continue to actively influence the broader agenda of developing large-scale experimental facilities for IoT and CPS ([IT1],[IT7]).

 

Industrial Automation

The use of wireless communication in the industrial automation domain enables new application scenarios not possible with wired links, it leads to reduced costs and increased levels of spatial fidelity with which the processes are monitored and controlled. Due to the stringent reliability and latency requirements and the need for energy-efficiency in some scenarios, traditional wireless solutions from other domains are not directly applicable. The increased interest in the “Industry 4.0” concept puts additional emphasis on simple integration of the local industrial automation systems with external cloud-based infrastructure. This raises the need for wireless solutions that offer high levels of reliability and predictability while being able to natively interface with IP-based networks. In cooperation with academic partners like SICS, INRIA and UoC and with industrial leaders like NXP, I have been on the forefront of promoting the use of IEEE 802.15.4e / TSCH protocol and the IPV6 over TSCH (6TiSCH) adaptation layer, as solution to the above challenges [J1]. As part of a collaborative effort bringing together the four most popular operating systems for networked embedded systems: Contiki, TinyOS, OpenWSN and RIOT, I have coordinated the implementation of a TSCH-based stack in TinyOS and performed a horizontal analysis on the impact of these protocols on the run-time environments and the requirements for the hardware/software interface to the timer and transceiver modules. 

 

Localization

Indoor localization services are crucial building-block for many Future Internet applications that have to be context-aware, leading to plethora of research and commercial solutions. Unfortunately, most of these solutions are only evaluated under custom, non-comparable and non-repeatable conditions. As a result, many of them fail to attain the expected performance levels when migrating from development to real-life conditions. As part of the EVARILOS project, I have contributed to the development of a common methodology and a set of metrics enabling fair comparison of indoor localization solutions [C17]. I have also served as member of the editorial committee for the international standard ISO/IEC 18305 “Test and evaluation of localization and tracking systems”, under leadership of NIST, that codifies more precisely a set of concrete of testing scenarios and metrics. Building on my experience with developing large-scale testing infrastructures, I contributed to the design of a platform for fully automatized benchmarking of RF-based indoor localization solutions under controlled interference ([J2],[J4],[C8],[C9],[C11]).

I have participated in the establishment of a novel indoor localization competition led by Microsoft, as part of the CPS Week and the IPSN conference ([J5],[C10]). Since its first edition during the IPSN 2014 (for which I served as demo chair), the Microsoft Indoor Localization Competition has established itself as the primary venue for testing and showcasing the state-of-the art indoor localization solutions under common conditions. 

Zusatzinformationen / Extras

Quick Access:

Schnellnavigation zur Seite über Nummerneingabe

Auxiliary Functions