Welcome to ESSPRITS 2011
#Program|#Poster|#Registration|#Technical|#Committee|#Hotels|#CarRent|#Parking|#Contact
October 20 — 21, 2011.
University of Waterloo, Ontario, Canada.
The tremendous success of smartphones and the multitude of software applications available trigger a number of interesting research questions about the use of embedded sensors, signal processing methodologies, and the security of personal and technical data. Clearly, these research questions transgress the boundaries of the established technical disciplines and call for an interdisciplinary approach.
This German-Canadian workshop will include researchers from the fields of IT security, signal processing, control and embedded systems to discuss new directions and solutions in these key areas.
The workshop will feature lecture and poster sessions and a panel discussion. Keynote lectures will be on "Challenges in a World of Ubiquitous Computing" and "On the Security of MAC Schemes in the Multi-User Setting".
List of Confirmed Speakers and Panelists
- S. Fischmeister, University of Waterloo
- T. Florio, Research In Motion
- C. Gebotys, University of Waterloo
- R. Legge, University of Waterloo
- F. Karray, University of Waterloo
- M. Kuntz, University of Waterloo
- R. Martin, Ruhr-Universität Bochum
- A. Menezes, University of Waterloo (Keynote Speaker)
- P. Moss, NSERC
- M. Mönnigmann, Ruhr-Universität Bochum
- C. Paar, Ruhr-Universität Bochum
- A. Schmidt, University Stuttgart (Keynote Speaker)
- S. Sundaram, University of Waterloo
- Z. Wang, University of Waterloo
Program
All lectures and panel discussions are in DC 1302
Day One — October 20, 2011 |
||
| 08:30 — 09:15 | Registration and Coffee | |
| 09:15 — 09:30 | Opening and Welcome Addresses | |
| 09:30 — 10:00 | Keynote 1 Challenges in a World of Ubiquitous Computing |
A. Schmidt |
| Lecture Session (Chair — D. Kolossa) | ||
| 10:00 — 10:20 | Fast Model and Optimization Based Control on Lean Embedded Systems | M. Mönnigmann |
| 10:20 — 10:40 | Real-time Embedded Software Systems | S. Fischmeister |
| 10:40 — 11:00 | Nutrition Break Refreshments available in DC 1301 (Fishbowl) |
|
| Lecture Session | ||
| 11:00 — 11:20 | Teleconferences and Cocktail Parties — Challenges in Embedded Audio Signal Processing | R. Martin |
| 11:20 — 11:40 | Perceptual Image Processing and Visual Communications | Z. Wang |
| 11:40 — 12:00 | Physically Unclonable Functions in Embedded Systems | Tim Günesyu |
| 12:00 — 12:20 | Side Channel Challenges in Embedded Security | C. Gebotys |
| 12:20 — 13:30 | Lunch Buffet lunch in |
|
| 13:30 — 14:00 | Keynote 2 On the Security of MAC Schemes in the Multi-User Setting |
A. Menezes |
| Lecture Session (Chair — T. Guneysu) | ||
| 14:00 — 14:20 | Using Structured System Theory to Identify Malicious Behavior in Distributed Systems | S. Sundaram |
| 14:20 — 14:40 | UW Center for Pattern Analysis and Machine Intelligence: Overview of Research Activities | F. Karray |
| 14:40 — 15:00 | Read My Lips: Audiovisual Speech Recognition for Human-Machine Interaction in Noisy Environments | D. Kolossa |
| 15:00 — 15:15 | Nutrition Break Refreshments available in DC 1301 (Fishbowl) |
|
| 15:15 — 16:45 | Panel Discussion (Moderator: R. Kinne) Discussion topics:
|
T. Florio, D. Kolossa, M. Kuntz, R. Legge, P. Moss |
| Poster Session (Chairs — N. Pohl and X. Fan) | ||
| 17:00 — 18:30 | Researchers from the University of Waterloo and the Ruhr-Universität Bochum present their work (in DC 1301 (Fishbowl)). Refreshment available in Fishbowl. |
|
Day Two — October 21, 2011(For members of the German team) |
||
| 09:00 — 14:00 |
|
|
Poster Presentations
From Germany
Efficient Manycore CHMM Speech Recognition for Audiovisual and Multistream Data
— Dorothea Kolossa, Jike Chong, Steffen Zeiler and Kurt Keutzer
Robustness of speech recognition can be significantly improved by using video recordings rather than audio input only. This is of interest e.g. for human-machine interaction in noisy reverberant environments, or for search in multimedia data. In contrast to conventional speech recognition, the search space increases significantly, preventing straightforward real-time implementation.
Thus, for real-time constrained applications, using current multiprocessor computing capability can be vital. This poster describes how general purpose graphics processors can be used to speed up decoding by exploiting their massive parallelism, allowing for real-time performance of audiovisual speech recognition.
Robust Automatic Speech Recognition
— Steffen Zeiler, Dorothea Kolossa and Rainer Martin
Speech recognition using close-talking microphones is already reliable and commonly used. In contrast, for distant talking applications such as car navigation, home entertainment systems, or human-machine interaction in public spaces, speech recognition is still insufficiently robust.
In order to overcome the main problems of such setups - noise, reverberation, and cross-talk - statistical signal processing methods in the front-end are valuable tools. This poster presents approaches based on de-noising, de-reverberation and source separation, and describes ways of achieving a tight integration of pre-processing with speech recognition by means of so-called uncertainty-of-observation approaches.
COPACOBANA - A Cost Optimized Parallel Code Breaker
— Tim Güneysu
COPACOBANA (Cost-Optimized Parallel COde Breaker) is an FPGA-based computing cluster built for less than 10,000 € and optimized to run cryptanalytical algorithms. COPACOBANA is suitable for computational problems which are parallelizable, and have low communication and memory requirements.
Post Quantum Cryptography for Embedded Devices
— Stefan Heyse
In the last years, the need for embedded systems has arisen continuously. Spanning all aspects of modern life, they are included in almost every electronic device: mobile phones, personal digital assistants (PDAs), domestic appliances, and even in cars. This ubiquity goes hand in hand with increased need for embedded security. For instance, it is crucial to protect a car's electronic door lock from unauthorized use. These security demands can be solved by cryptography. In this context, many symmetric and asymmetric algorithms, such as AES, DES, RSA, ElGamal, and ECC, are implemented on embedded devices.
Most public-key cryptosystems frequently implemented have been proven secure on the basis of the presumed hardness of two mathematical problems: factoring the product of two large primes (FP) and computing discrete logarithms (DLP). Both problems are well known to be closely related. Hence, solving these problems would have significantly ramifications for classical public-key cryptography, and thus, for embedded devices the algorithms are implemented on. Nowadays, both problems are believed to be computationally infeasible with an ordinary computer. However, a quantum-computer having the ability to perform computations on a few thousand qbits could solve both problems by using Shor's algorithm. Although a quantum computer of this dimension has not been reported, development and cryptanalysis of alternative public-key cryptosystems seem suitable. Cryptosystems not breakable using quantum computers are called post-quantum cryptosystems.
Published post-quantum public-key schemes are focused on the following approaches: Hash-based cryptography (e.g. Merkle's hash-tree public-key signature system), Multivariate-quadratic-equations cryptography (e.g. HFE signature scheme), Lattice-based cryptography (e.g. NTRU encryption scheme), and Code-based cryptography (e.g. McEliece encryption scheme, Niedereiter encryption scheme).
To achieve acceptance and attention in practice, post-quantum public-key schemes have to be implemented efficiently. Furthermore, the implementations have to perform fast while keeping memory requirements small for security levels comparable to conventional schemes. Additionally, the newest attacks e.g. SPA, DPA, Fault injection etc. have to be taken into account by the implementer.
Microphone-Arrays and Beamforming in Embedded Systems
— Sebastian Gergen and Rainer Martin
Mobile phones and hearing aids are the most prominent applications for embedded systems which capture sound for signal processing and transmission. In a real world scenario, sound of a desired source is always contaminated by room related reverberation and noise, which may originate from sound sources differing from the desired one. The spatial diversity of sound sources in a room can be exploited to reduce the effects of reverberation and noise if multiple microphones arranged in a microphone-array are used for the sound pickup. By adequately filtering each of the microphone signals and summing them up an improved version of the originally received signal can be achieved. This beamforming technique allows a flexible generation of a frequency dependent spatial directivity. Once beamforming is applied the output signal of the beamforming system can be used for further processing, e.g. single channel noise reduction and/or source localization. In this poster presentation
We will discuss some recent advances in one- and two dimensional microphone-array and beamformer design.
Single-channel Speech Enhancement: A Real-time Demonstration
— Rainer Martin, Timo Gerkmann, Colin Breithaupt, Dirk Mauler, Martin Krawczyk
In this demonstration we present a low-latency, real-time framework for single channel speech enhancement. The demonstrator allows to select and switch between many state-of-the-art algorithms. It includes the first real-time implementation of the newly developed temporal cepstrum smoothing concept for output quality enhancement. This concept has been introduced at ICASSP 2008 by the authors and has attracted considerable interest since. The demonstrator can be used to compare different spectral windows, a priori SNR estimators and speech presence probability estimators with and without temporal cepstrum smoothing.
Jackstraws: Picking Command and Control Connections from Bot Traffic
— Ralf Hund and Thorsten Holz
A distinguishing characteristic of bots is their ability to establish a command and control (C&C) channel. The typical approach to build detection models for C&C traffic and to identify C&C endpoints (IPs and domains of CC servers) is to execute a bot in a controlled environment and monitor its outgoing network connections. Using the bot traffic, one can then craft signatures that match C&C connections or blacklist the IPs or domains that the packets are sent to. Unfortunately, this process is not as easy as it seems. For example, bots often open a large number of additional connections to legitimate sites (to perform click fraud, query for the current time), and bots can deliberately produce "noise" - bogus connections that make the analysis more difficult. Thus, before one can build a model for C&C traffic or blacklist IPs/domains, one first has to pick the C&C connections among all the network traffic that a bot produces.
We present JACKSTRAWS, a system that accurately identifies C&C connections. To this end, we leverage host-based information that provides insights into which data is sent over each network connection as well as the ways in which a bot processes the information that it receives. More precisely, we associate with each network connection a behavior graph that captures the system calls that lead to this connection, as well as the system calls that operate on data that is returned. By using machine learning techniques and a training set of graphs that are associated with known C&C connections, we automatically extract and generalize graph templates that capture the core of different types of C&C activity. Later, we use these C&C templates to match against behavior graphs produced by other bots. Our results show that JACKSTRAWS can accurately detect C&C connections, even for novel bot families that were not used for template generation.
Integrated SiGe Transceiver Circuits for Ultra-Wideband 80 GHz FMCW Radar Systems
— Nils Pohl and Timo Jaeschke
The increasing use of radar systems in automotive and industrial fields demands for inexpensive high precision implementations. Modern SiGe bipolar technologies are well-suited for the use at mmWave frequencies around 80/94 GHz. The use of FMCW frequency ramps with very high bandwidth (>10 GHz) enables very robust and precise measurements and open new markets for radar systems, which are dominated by costly and susceptible laser measurement systems.
In this poster, the bandwidth limitations of common radar systems will be discussed on system and circuit level. Afterwards, an ultra-wideband realization will be presented, which clearly exceeds the state of the art.
An Integrated Broadband FMCW Radar Sensor for Ultra High Resolution SAR Imaging at 80 GHz
— Timo Jaeschke, Michael Vogt, Nils Pohl
Synthetic aperture radar (SAR) is used in many different applications (e.g. environmental monitoring, surveillance or airport security body scanner systems) for imaging of a radar target or scene. The image resolution after SAR processing in azimuth direction is determined by half of the antenna aperture width. Using high frequencies of 80 GHz in combination with short distances and a small horn antenna a very good resolution of 1.5 cm can be achieved, whereas the image resolution in range direction depends on the bandwidth of the radar sensor. The record bandwidth achieved by existing radar systems is 10 GHz.
Here, a complete frequency modulated continuous wave (FMCW) experimental radar system with a record bandwith of 24.5 GHz is presented. This allows almost isotropic resolution cells for SAR imaging of 1.5 cm x 1.3 cm (-6 dB width) with applied hanning window function for sidelobe reduction in signal processing.
First ultra high resolution SAR images are shown to prove this good resolution achieved by the broadband 80 GHz radar sensor.
In addition to this a body scanner as a possible application for the high resolution radar imaging system is presented.
Towards Suboptimal Explicit Nonlinear Model Predictive Control with Guaranteed Stability
— M. Schulze Darup, M. Jost, M. Mönnigmann
Model predictive control (MPC) has successfully been applied in a variety of industrial fields. While used for systems with faster and faster dynamics, the high computational demand of MPC is still an obstacle to applications at high sampling times of, say, less than a millisecond. Explicit MPC methods are an alternative in these cases, because they provide a solution to the MPC optimization problem in the form of an explicit feedback law u(x) and thus do not require online optimization. For example, in the case of a linear dynamical system with linear input and output constraints and a quadratic performance criterion (linear-quadratic MPC), the control law is known to be a continuous piecewise linear function that can be calculated explicitly. While this solution is structurally simple, it can be complex in that the piecewise control law may have tens of thousands of pieces even for low dimensional linear dynamical systems and short prediction horizons. We discuss new techniques to calculate and represent control laws of this form efficiently. Moreover, we present ideas on how to extend linear-quadratic explicit MPC to a suboptimal but tractable case for nonlinear systems.
From University of Waterloo
High Performance GHASH
— Jithra Adikari, M. Anwar Hasan, Nicolas Mèloni, and Christophe Nègre
This work presents a new hardware implementation of the GHASH function involved in the Galois/Counter Mode of operation for block ciphers. This function is usually computed by using successive multiply-and-add operations over the binary field of dimension 128. We have considered a recent method for GHASH computation which first performs a reduction modulo the characteristic polynomial of the hash key and then performs several multiply-and-add operations to compute the final hash value. In this work we report our implementations of the characteristic polynomial based GHASH computation using FPGA and ASIC.
Integrity Verification of Multiple Data Copies over Un-trusted Cloud Servers
— Ayad F. Barsoum and M. Anwar Hasan
Increasingly more and more individuals and organizations are opting for outsourcing data to remote cloud service providers (CSPs). This is primarily to reduce the maintenance cost and the burden of large local data storage. Customers can rent the CSP's storage infrastructure to store and retrieve an unlimited amount of data by paying fees metered in GB/month. For an increased level of scalability, availability and durability, some customers may want their data to be replicated on multiple servers across multiple data centers. The more copies the CSP is asked to store, the more fees the customers are charged. Therefore, customers need to have a strong guarantee that the CSP is storing all data copies that are agreed upon in the service contract, and all these copies are not being tampered with or partially deleted over time.
Consequently, the problem of provable data possession (PDP) has been considered in many research papers. Several previous PDP schemes focus on the provable possession of a single copy of the file, and provide no guarantee that the CSP stores multiple copies of customers’ data. In this presentation, we address the issue of creating multiple copies of a data file and verifying those copies stored on cloud servers. We propose a multi-copy provable data possession (MC-PDP) scheme that is provably secure against colluding servers. The proposed scheme allows the authorized users of the data file to seamlessly access the copies received from the cloud servers. Through theoretical analysis and experimental results, we demonstrate the performance of our scheme.
Time-aware Instrumentation on the Worst-case Path
— Hany Kashif and Sebastian Fischmeister
Program tracing through instrumentation is a well established method for debugging. Real-time embedded programs are time sensitive and thus the instrumentation mechanism must honor the programs' timing constraints. Our approach introduces slack-based conditional instrumentation. The underlying idea is to keep track of slack and execute instrumentation code whenever possible even on the worst-case path.
Cross-band Interference Reduction Trade-offs in SISO and MISO OFDM-based Cognitive Radios
— Ehsan Haj Mirza Alian, Hamidreza Ebrahimzadeh Saffar, and Patrick Mitran
Cognitive radio is a promising approach for efficient utilization of radio spectrum. Among all the proposed signaling schemes for cognitive radio physical layer, such as multi-carrier modulation, filter bank multi-tone modulation, and single-carrier frequency division multiple access, orthogonal frequency division multiplexing (OFDM) seems to be a promising one due to its robustness against multipath fading, high spectral efficiency, and capacity for dynamic spectrum use.
In this research, we investigate the problem of cross-band interference minimization in OFDM-based cognitive systems. Cross-band interference is mainly caused by high OFDM sidelobes. In the first part of our work, we propose a framework to study the trade-off between two recently proposed techniques, adaptive symbol transition (AST) which is performed in the time domain, and active interference cancellation (AIC) which is performed in the frequency domain. We use the trade-off study results to maximize the useful data rate for a desired level of interference. Simulation results show that the best trade-off depends on the configuration of spectral opportunities and specifically, whether there is one large primary band or multiple smaller primary bands.
In the second part, a new method for interference reduction in multiple-antenna cognitive systems is developed. We show that with knowledge of the channel, the secondary transmitted sequences can be jointly optimized over multiple antennas such that the interference at the primary receiver location is better minimized. Computer simulations also demonstrate an improvement of almost 10 dB compared to separate-antenna optimization.
Performance Analysis of Null-steering Beamformers in Cognitive Radio Systems
— Nadia Jamal and Patrick Mitran
We evaluate the performance of a secondary system with a multi-antenna transmitter and a single-antenna receiver. The secondary system coexists with a primary system in an underlay cognitive setting where both systems share the same frequency bands simultaneously. A secondary beamforming vector is characterized such that the interference power at each primary receiver is nullified (null-steering beamforming). While constraining the secondary transmitted power, we aim to achieve the maximum received power at the secondary receiver. With perfect channel state information (CSI), we show that the secondary system can achieve a mean received power that grows linearly in the number of secondary transmitting antennas and is directly proportional to the the power of the line of sight (LOS) component between the secondary transmitter and the secondary receiver (secondary LOS power). Furthermore, in the case of imperfect CSI, it is shown that a moderate secondary LOS power can significantly reduce the effect of estimation error on the performance. Thus, almost the same mean secondary received power is yielded as obtained without channel estimation error.
Securing Passive RFID Communication Through Physical Layer Enhancements
— Qi Chai and Guang Gong
Although RFID systems offer many noteworthy characteristics, security and privacy issues associated with them are not easy to address. In this paper, we investigate how to solve the eavesdropping, modification and one particular type of relay attacks toward the tag-to-reader communication in passive RFID systems without requiring lightweight ciphers or secret credentials shared by legitimate parties using a physical layer approach. To this end, we propose a novel physical layer scheme, called Backscatter modulation- and Uncoordinated frequency hopping-assisted Physical Layer Enhancement (BUPLE). The idea behind it is to use the amplitude of the carrier to transmit messages as normal, while to utilize its periodically varied frequency to hide the transmission from the eavesdropper/relayer and to exploit a random sequence modulated to the carrier's phase to defeat malicious modifications. We further improve its eavesdropping resistance through the coding in the physical layer as BUPLE ensures that the tag-to-eavesdropper channel is strictly noisier than the tag-to-reader channel. Three practical Wiretap Channel Codes (WCCs) for passive tags are then proposed: two of them are constructed from linear error correcting codes, and the other one is constructed, for the first time to the best of our knowledge, from resilient vector Boolean functions. The security and usability of BUPLE in conjunction with WCCs are further confirmed by our proof of concept implementation and testing on the software defined radio platform with a programmable WISP tag.
A Lightweight Pseudorandom Number Generator for EPC Class 1 Gen2 RFID Tags
— Kalikinkar Mandal, Xinxin Fan, and Guang Gong
Radio Frequency Identification (RFID) is a promising technology for automatic identification of remote objects. For most RFID applications, security is an important or even crucial requirement. Since most protocols for securing RFID systems proposed so far are based on the use of an on-board true random and/or pseudorandom number generator (TRNG/PRNG), a number of solutions have been proposed in literature for implementing TRNGs/PRNGs on RFID tags. In particular, the EPCglobal Class-1 Generation-2 (EPC C1 Gen2 in brief) standard uses random numbers in the tag identification protocol. All of the proposals for TRNGs are based on analog circuits that sample a random physical phenomenon like thermal noise. To the best of our knowledge, only three PRNGs have been proposed for EPC C1 Gen2 tags, among which two proposals use TRNGs as a component and the security properties of these two PRNGs rely on the security of TRNGs. Considering the high power consumption, large area and low throughput of TRNGs, we propose a lightweight PRNG for low-cost EPC C1 Gen2 tags in this contribution. The basic idea of our design is to replace the TRNG by a lightweight pseudorandom sequence generator with good statistical properties. To this end, nonlinear feedback shift registers (NFLSRs) have been fully exploited in our design. The proposed PRNG can be implemented in hardware using around 1,242 logic gates.
An Ultra-Efficient Key Recovery Attack on the Lightweight Stream Cipher A2U2
— Qi Chai, Xinxin Fan, and Guang Gong
Although Radio Frequency Identification (RFID) technology provides many attractive and exclusive characteristics, the constrained computational and storage capabilities as well as the extremely low manufacture cost of RFID tags have posed a new challenge that goes beyond the traditional cipher design paradigm and stimulates the brand-new design of lightweight stream/block ciphers. In this poster, we report on an ultra-efficient key recovery attack under the chosen-plaintext-attack model against the stream cipher A2U2, which is the most lightweight cryptographic primitive (i.e., it costs only 284 GE in hardware implementation) proposed so far for low-cost Radio Frequency Identification (RFID) tags. Our attack can fully recover the secret key of the A2U2 cipher by only querying the A2U2 encryption twice on the victim tag and solving 32 sparse systems of linear equations (where each system has 56 unknowns and around 28 unknowns can be directly obtained without computation) in the worst case, which takes around 0.16 second on a Thinkpad T410 laptop.
OAuth and ABE based Access Token in Semi-Untrusted Cloud Computing
— Anuchart Tassanaviboon and Guang Gong
To protect cloud users from locking their data and applications into a single cloud provider is one objective of open cloud. The first stage of this open environment is inter-operation between users' data in cloud storage or database providers with web applications. However, outsourcing data to cloud storage or database can provide data and application interoperability. For users, this situation bring forth losing data control-ability. Therefore, we propose an new authorization system that users still take control of data when the data resting with cloud storage and until arriving at the legitimate receiver. To this end, we leverage CP-ABE scheme, ElGamal encryption and OAuth framework in our authorization our model to provide user-centric and end-to-end characteristics. We compare this model's security and efficiency with standard protocols, and present the benefits of and future work for this design.
Zero Correlation Zones of Binary/QAM Golay Sequences and Their Applications
— Guang Gong, Fei Huo, and Yang Yang
Sequences with desired correlation properties have wide implications in modern communications, radar and sonar applications. In this work, we present our new findings on some constructions of single H-ary Golay sequence and 4q-QAM Golay sequence with a large zero autocorrelation zone, where H>=2 is an arbitrary even integer and q>=2 is an arbitrary integer. The potential applications on binary Golay sequences and QAM Golay sequences with this property include system synchronization and detection at the receiver end.
Registration
There is NO FEE for attending the workshop, BUT space is limited. Researchers of the sponsoring organizations and other individuals wishing to attend the workshop, should email their registration requests including names and affiliations to: Chris Schroeder cschroeder@uwaterloo.ca before October 12, 2011.
The Workshop speakers, panelists, poster presenters and invited guests are automatically registered and DO NOT need to send any additional registration information.
Technical Program Committee
- Guang Gong
University of Waterloo, Canada - Anwar Hasan
University of Waterloo, Canada - Rainer Martin
Ruhr-Universität Bochum, Germany - Martin Mönnigmann
Ruhr-Universität Bochum, Germany - Christof Paar
Ruhr-Universität Bochum, Germany - Nils Pohl
Ruhr- Universität Bochum, Germany
Organizing Committee
- Dave Dietz
Research In Motion, Waterloo, Ontario, Canada - Guang Gong
Department of Electrical & Computer Engineering, University of Waterloo, Ontario, Canada - Joann Halpern
German Center for Research and Innovation, New York, USA - Anwar Hasan
Department of Electrical & Computer Engineering, University of Waterloo, Ontario, Canada - Rolf Kinne
ConRuhr North America, New York, USA - Rainer Martin
Ruhr-Universität, Germany - J. R. Sack
German Embassy, Ottawa, Canada
Hotels in Waterloo/Kitchener
Waterloo Inn
Address: 475 King Street North, Waterloo, Ontario, N2J 2Z5
Phone: 1-800-361-4708 1-519-884-0220
Best Western St Jacobs Country Inn
Address: 50 Benjamin Rd. East, Waterloo, Ontario, Canada, N2V 2J9
Phone: 1-519-884-9295 1-800-972-5371
The Waterloo Inn
Address: 475 King Street North, Waterloo, Ontario, Canada, N2J 2Z5
Phone: 1-519-884-0220 1-800-361-4708
The Waterloo Hotel
Address: 2 King Street North, Waterloo, Ontario, Canada, N2J 2W7
Phone: 1-519-885-2626 1-877-885-1890
Delta Hotels and Resorts
Address: 105 King Street East, Kitchener, Ontario, Canada, N2G 2K8
Phone: 1-519-744-41416 1-888-890-3222
Radisson Hotel
Address: 2960 King Street East, Kitchener, Ontario, Canada, N2A 1A9
Phone: 1-519-894-9500 1-800-333-3333
Car Rental
Enterprise Rent-A-Car
Phone: 1-519-884-6001 1-519-772-0888
Discount Car and Truck Rentals
Phone: 1-519-888-0576
Campus Parking
UW Parking Services
The suggested parking is N lot, which is close to DC. There is also limited visitor parking behind Engineering 5 (please enter off Phillip Street by East Campus Hall and you will see a sign for B lot. From this lot �C one can walk around the building �C across Ring Road into the Davis Centre). Both lots are Pay and Display - $5.00 CANADIAN COINS ONLY ACCEPTED.
Campus Map and Direction
Campus map
Phone: 1-519-888-4567
Toronto International Airport to University of Waterloo
Airways Transit
Phone: 1-519-886-2121 (ADVANCE RESERVATIONS ARE REQUIRED)
Contact Information
Chris Schroeder
Electrical & Computer Engineering
University of Waterloo
200 University Avenue West, E5, #4043
Waterloo, Ontario
Phone: (519) 888-4567 x 31466
Email: cschroeder@uwaterloo.ca