Connected Robotics!

↓Jobs↓

This application-focused research project is going to explore how the introduction of 5G will revolutionise robotics.

Motivation

Robotics often makes leaps when new technologies from other domains are introduced. In this project, we want to investigate what 5G can do for robotics. Our aim is to implement a diverse range of relevant robotics applications and evaluate the impact of connectivity properties on them. The following three applications can best describe our approach:

Investigate how lower latency improves user experience for manipulator teleoperation with force feedback
Leverage edge computing and high-throughput connections to process RGB-D data on off-board GPU clusters
Scale real-time low latency collaborative mapping and collision avoidance to large robot fleets

Outline

The project is a collaboration between three labs, their share is depicted in the columns above. The project aspect Evaluate is concerned with implementing typical robotics building blocks, optimising these together with the network, and documenting the performance. Verify is about novel, compact teleoperation interfaces with haptic feedback and user perception of the network performance through those devices. Compare focuses on scaling the previously conducted research to a real-world intralogistics scenario and understanding connectivity requirements of tightly-coupled distributed algorithms.

Available positions

Six positions in total are available within this project. Further details are given below.

MASc: Haptics Demonstrator: Build a sensible and compact haptics input device to illustrate the effects of network latency and reliability
MASc: Connected Robotics at Scale: Implement & analyse a complex real-world demonstration showing the effect of wireless communication on robotics algorithms
PhD: Connected Robotics at Scale: Explore & analyse the effects of wireless communication on tightly-coupled robotics fleets
PhD: Multi-access Edge Computing for Mobile Robotics: Investigate the benefits of edge-computing on robotics with a focus on real-time applications
PostDoc: Mobile Teleoperation User Experience: Investigate the impact of typical wireless communication on teleoperation for manipulators with force feedback
PostDoc: Novel teleoperation interfaces for mobile teleoperation: Investigate novel interfaces for teleoperation and how wireless networks impact the user experience

The positions have aspects of independent work, leading to independent publications and thesis-relevant results. They also have aspects of intense collaboration to enable swift progress in this cross-domain project.

Overall project goals

Compare and document the performance of robotics applications using 5G and Wi-Fi
Understand which network properties are most relevant for which application
Analyse the scalability of applications running on fleets
Provide traffic patterns typical for robotics applications
Verify application performance when using 5G mmWave and roaming between multiple cells

Network

Connected Robotics is a project within the collaboration of the University of Waterloo and Rogers Communications, Inc., a leading Canadian telecommunications company. It leverages Mitacs provincial government support and utilises the existing infrastructure of RoboHub provided by Innovation Canada and the Advanced Manufacturing Consortium. Rogers provides a 5G campus network including mmWave support and network know-how. On the academic side, three labs form the project: RoboHub, the Wireless Sensors and Devices Lab and the Haptic Experience Lab. Those labs are respectively led by William Melek (Mechanical and Mechatronics Engineering), George Shaker (Electrical and Computer Engineering), and Oliver Schneider (Management Sciences).

People

William Melek

William Melek is the director of the Robohub and a Professor in the Department of Mechanical and Mechatronics Engineering at the University of Waterloo. His research focus is in mechatronics and machine learning.

George Shaker

Prof George Shaker is a wireless and IoT expert, with a huge industrial experience. He is a professor of Department of Mechanical and Mechatronics Engineering as well as the Department of Electrical and Computer Engineering and leads the wireless sensors and devices lab.

Oliver Schneider

Prof Oliver Schneider is a professor at the University of Waterloo in the Faculty of Engineering (Department of Management Sciences). His research focus is in human-computer interaction, specifically haptic technology.

Alexander Werner

Alexander Werner is a researcher at the RoboHub. His focus is on control of complex robotic systems, centered around torque/force control systems.

Brandon DeHart

Prof DeHart is the Manager of the RoboHub. His research focus is span from bioinspired robot design over interactive architecture and includes bipedal locomotion.

Open Positions

These positions are fully-funded. The positions are funded by Rogers Communications and MiTacs, a provincial government research agency. This funding implies a focus on education and industrial expertise. For more information about other aspects of studying at the University of Waterloo, please see UW-Engineering: Graduate Students. In case you have questions about the research direction, requirements, or any other aspects contact us through robohub@uwaterloo.ca. Other methods of reaching out: Twitter: @uwrobohub.
The positions are contingent on the acceptance of our proposal by Mitacs, which is influenced by provincial politics. This statement will be removed once the proposal is accepted.

Schedule:

Submit your documents by 2023-01-31 to robohub@uwaterloo.ca. A response is guaranteed within 21 days after deadline. Please include: which position you are interested in, a current curriculum vitae, and an up to date grades summary. A detailed description of projects you worked on is also of value.
Please be aware of the start dates given in the description of each position. If positions are not filled in the first round, we will update the proposal accordingly.

MASc: Haptics Demonstrator

Build a sensible and compact haptics input device to illustrate the effects of network latency and reliability.

Description

Design and develop a teleoperation interface that allows haptic (force) feedback. This compact interface will focus on haptic rendering quality at limited complexity and cost. It will allow the user to experience differences in various connectivity properties, especially abstract metrics like latency and reliability. You will integrate this interface in a teleoperation setup and will compare its capabilities against existing solutions. You will document your work and contribute to the publication in peer-reviewed journals.
Supervisor: Oliver Schneider. Before applying, feel free to familiarise yourself with the ongoing research in my lab.
Tentative start date: 2023-09-01

Requirements:

An undergraduate degree in Electrical or Mechatronics engineering
Knowledge in controls
Knowledge in mechanical design
Proficiency in English, both spoken and written, to communicate daily

Preferred qualifications:

Experience in actuator design
Experience in haptics
Experience writing real-time control software

Your duties:

Design a teleoperation interface with force feedback
Implement & verify the haptic rendering performance
Assist with user study to analyse user effectiveness in teleoperation

Your experience:

Understanding haptics
Mechatronic design
Industry collaboration
Scientific approach and publications
Teamwork

MASc: Connected Robotics at Scale

Implement & analyse a complex real-world demonstration showing the effect of wireless communication on robotics algorithms

Description

You will be tasked with applying the research prototypes implemented in other work packages to an industrial-scale scenario. The scale will allow you to test the performance of typical robotics applications like collaborative mapping and teleoperation of manipulators under realistic conditions. Industrial partners like Attabotics and Purolator will be eager to provide test cases, enabling them to understand the implications a 5G or Wi-Fi deployment would have. One planned demonstrator setup is to drive an autonomous ground vehicle through the coverage of multiple network cells and monitor how applications are affected at handovers. This demo will also include the teleoperation of an arm mounted on the vehicle, integrating it with other parts of this project. The primary result will be the understanding of the intrinsic limitations of those mobile networking options, including details of which applications require which level of reliability and latency. An auxiliary output of your research will be traffic data which allows Rogers to further improve their network to support these applications.
Supervisor: George Shaker. Before applying, feel free to familiarise yourself with the ongoing research in my lab.
Tentative start date: 2024-09-01

Requirements:

An undergraduate degree in engineering or computer scinence
Knowledge of robotics
Basic knowledge of ROS
Knowledge of Linux
Proficiency in English, both spoken and written, to communicate daily

Preferred qualifications

Experience in integrating complex (robotic) systems
Experience in robotics: mapping
Experience in computer networking

Your duties:

Integrate communication interfaces
Realise trajectory execution on the vehicle
Realise manipulator teleoperation on this platform
Record & document your results
Contribute to publications

Your experience:

Deployment of robotics application to complex systems
Behavior of wireless networks, especially networks consisting of multiple cells
Industry collaboration
Scientific approach and publications
Teamwork

PhD: Connected robotics at scale

Explore & analyse the effects of wireless communication on tightly-coupled robotics fleets

Description

This PhD will focus on understanding which mobile robotics applications profit from close-by computing infrastructure. One example is deep-learning-based object tracking which makes it necessary to process rich, high-bandwidth sensor data in real time. For this class of applications, the question is: Should a continuous high-bandwidth data stream flow from a mobile robot through the network to a cloud data center, or should it be processed closer - in terms of physical distance as well as latency? The obvious solution is to process all data on board, however, many reasons stand against this: cost, power consumption, and size. Another solution is to process data on shared computing located as close as possible to the robot, in the extreme case next to the wireless base station the robot is connected to. For 5G, this is called multi-access edge computing (MEC).
In your research you will evaluate the different placement options for computing infrastructure using multiple robotics applications as test cases. Building on existing code, you will realise a complex real-world test scenario. This will yield valuable data which enables others to design their robots and networks accordingly. Going along with the scheme of the whole project, the approaches will will be evaluated using 5G and Wi-Fi.
Supervisor: George Shaker. Before applying, feel free to familiarise yourself with the ongoing research in my lab.
Tentative start date: 2024-05-01

Requirements:

A MASc degree in engineering or computer science
Experience in robotics
Knowledge of deep learning
Basic knowledge of ROS
Knowledge of Linux
Proficiency in English, both spoken and written, to communicate daily

Preferred qualifications:

Deployment of robotics application to complex systems
Experience in deep-learning based pose estimation
Experience in visual servoing

Your duties:

Compare performance of edge-computing and traditional cloud-based solutions
Analyze off-board sensor data processing code for different compute options
Evaluate behavior at handover
Supervise MASc students helping with preparation of experiments
Publish results

Your experience:

Integration of complex robotic systems
Understand real-time algorithms for robotics
Master reliable distributed applications for robotics
Industry connection
Scientific approach and publications
Teamwork

PhD: Multi-access Edge Computing for Mobile Robotics

Investigate the benefits of edge-computing on robotics with a focus on real-time applications

Description

This PhD focuses on understanding all possible uses of edge computing within wireless networks, primarily 5G, for mobile robotics applications. You will be conducting research in the robotics fields of visual-servoing and collaborative mapping, leveraging the multi-access edge computing capabilities of a 5G network. First, you will thoroughly investigate the requirements for different robotics use cases such as the required bandwidth or how the application reacts to varying latency and reliability. You will then integrate the different test setups with the 5G research network provided by our industry partner Rogers and evaluate the real-world performance. These results will then be compared to alternative methods of connecting robots such as Wi-Fi.
Concretely speaking two benchmark applications are to be implemented. (1) Visual-servoing using RGB-D data streamed over the wireless network to a GPU-accelerated computed cluster on the network edge. (2) Collaborative mapping & collision avoidance in crowded and dynamic environments of a robot fleet.
Your results can then be used in the collaborative WP3 where they will be applied in an industrial-scale use case.
Supervisor: William Melek
Tentative start date: 2023-10-01

Requirements:

A MASc degree in engineering or computer science
Experience in Kinematics & Controls for manipulators
Experience in (time-discrete) control
Proficiency in English, both spoken and written, to communicate daily

Preferred qualifications:

Experience in Deep-learning-based object pose estimation
Experience in Visual-servoing
Experience in Mapping, especially collaborative mapping
Computer networking

Your duties:

Integrate 5G interfaces & understand provisioning algorithm on edge-computing devices
Implement state-of-the-art deep-learning based 6D object real-time tracking
Deploy visual-servoing algorithm and analyse object tracking performance
Understand performance as a function of connectivity metrics latency, throughput and reliability
Generate experimental results for 5G & Wi-Fi connections
Publish results

Your experience:

Reactive robotics: e.g. visual-servoing
Evaluating distributed real-time applications
Behavior of wireless networks
Industry collaboration
Scientific approach and publications
Teamwork

For the following PostDoc positions, consider the additional information available at University of Waterloo: Become a postdoc and Engineering Faculty: Postdoctoral Fellows.

PostDoc: Mobile Teleoperation User Experience

Investigate the impact of typical wireless communication on teleoperation with force feedback

Description

Your task in this project is to produce leading research about limitations in the area of teleoperation with force feedback over optimised 5G networks and compare these with Wi-Fi connectivity. Based on an existing teleoperation implementation, you will improve statics gathering, and optimise feedback control strategies for teleoperation with force feedback. These algorithms are to be tested on a range of manipulator robots including Kuka IIWA and Franka-Emika Panda. After the integration of 5G and Wi-Fi network interfaces, you will holistically analyse and optimise any existing bottlenecks in the setups. A user study will be used to measure the impact of network properties of latency and reliability on the users' ability to conduct reference tasks. As an extension of this work, you will collaborate on a mobile device teleoperation application utilising virtual reality technologies to maximise the usefulness of such an approach (Demo).
Supervisor: William Melek
Tentative start date: 2023-09-01

Requirements:

A Ph.D. degree in engineering or computer science
Experience in robotics
Experience in feedback control
Experience in C++
Proficiency in English, both spoken and written, to communicate daily

Preferred qualifications:

Experience in teleoperation
Experience in networking
Experience in robotics: force control
Experience in digital control

Your duties:

Adopt existing teleoperation code
Integrate connectivity to robot systems
Understand and quantify performance, investigating bottlenecks
Design and execute a user study
Publish results

Your experience:

Real-time analysis of distributed/hybrid systems
Optimizing real-time control code
Practical stability analysis of teleoperation with force feedback
Industry connection
Scientific approach and publications
Teamwork

PostDoc: Novel Teleoperation Interfaces for Mobile Teleoperation

Investigate novel interfaces for teleoperation and how wireless networks impact the user experience

Description

We have identified a new research area that brings existing high-transparency teleoperation approaches to mobile applications. We consider two cases where reliable and low-latency wireless networks are key to these developments: mobile manipulators and mobile teleoperation interfaces. Teleoperation has been studied intensively for extreme cases such as perfect communication and long latencies. Optimising approaches for these new conditions has tremendous potential through the sheer number of applications and users. Your work will be split into an analysis of novel interface concepts and the design of such. Together with a MASc student under your supervision, you will realise a simplified reference hardware interface that allows conducting a user study to understand the impacts of network reliability and latency on the user experience. Together with collaborators, you will contribute your expertise to a mobile-phone-based interface. These directions will enable you to publish relevant research and further your scientific career.
Supervisor: Oliver Schneider. Before applying, feel free to familiarise yourself with the ongoing research in my lab.
Tentative start date: 2024-02-01

Requirements:

A Ph.D. degree in Electrical or Mechatronics engineering
Experience in robotics
Experience in Mechatronics design
Proficiency in English, both spoken and written, to communicate daily

Preferred qualifications:

Experience in teleoperation
Experience in actuator control
Experience in analysis of real-time systems

Your duties:

Review existing teleoperation interfaces
Supervise design of compact teleoperation interface
Collaborate on mobile-phone based interface
Publish results

Your experience:

Teleoperation under realistic network condition
User studies in haptics
Industry collaboration
Scientific approach/publications
Teamwork