Robotics Research Projects

Introduction

Robotics Research Projects will provide an overview of the latest research in the field of robotics, as well as a hands-on approach to bring critical skills together. This is done in a project-oriented course where students will design mechanical, electrical, and software subsystems of an overall functioning robot.

In this module, students will come up with a user case for a robotic solution, develop a hypothesis about the effect of a design criterion on the user experience (e.g. across gender, culture, or well-being), develop the robotic solution (experimental set up), test on human participants, obtain measurements, analyse data, and write a report in a conference paper format (6-8 pages).

This requires students to fuse elements of embedded programming, control, and mechanical fabrication in the design and build of highly contextualized smart technologies. They must consider the interaction between the human users and their robot systems, and demonstrate a systematic methodical approach to the conception, development, and validation of their ideas. This has the aim of proposing new methods for the development and application of robotics in new contexts across the world.

Background

A 2013 report by McKinsey [1] estimated that advanced robotics could generate a potential economic impact of USD1.7-4.5 trillion by 2025. It mentions that “These advanced robots have greater mobility, dexterity, flexibility, and adaptability, as well as the ability to learn from and interact with humans. In advanced economies, some workers might find new job opportunities in developing, maintaining, or working with robots”. The users of these future robotic co-workers will be diverse in terms of gender and culture. Therefore, it is pertinent to develop a robotics module that accounts for how gender matters in the design process. Recent work shows that user experiments must cover gender to obtain a robust picture of how different features of products and services affect user safety and satisfaction [2]. Recent work at the MIT media lab has found that gender plays an important role in human-robot interaction in terms of the persuasiveness or the ability to influence a user to change behavior [3]. Therefore, we will choose several domains of human-robot interaction that allow us to raise questions about how gender and culture of robot users and designers should be integrated into the design process of robots.

[1] McKinsey global report 2013: Disruptive technologies

[2] Schiebinger, Londa. “Gendered innovations: harnessing the creative power of sex and gender analysis to discover new ideas and develop new technologies.” Triple Helix 1.1 (2014): 9.

[3] Siegel, Mikey, Cynthia Breazeal, and Michael I. Norton. “Persuasive robotics: The influence of robot gender on human behavior.” Intelligent Robots and Systems, 2009. IROS 2009. IEEE/RSJ International Conference on. IEEE, 2009.

Please also look at this video:

Project scopes

Please review the following project outlines and choose an area you would like to investigate further. You can discuss with the module leaders Thrishantha Nanayakkara, Petar KormushevNicolas Rojas, and Weston Baxter to identify a refined research question to investigate deeper. You can do a project in a team of up to 3 members. You will work in the robotics research labs in the basement of the Dyson Building to conduct experiments. You will also get the support of the research teams in the robotics labs.

  1. Robo-patient: Medical students learn to perform an abdominal physical examination of patients following scientific briefing sessions and demonstrations from experts. Complex factors and their interactions sometimes unknown to the expert, condition these demonstrations. For instance, our recent findings show that haptic information gain during physical examination of a soft tissue phantom depends on how we control stiffness, force, and speed of fingers. Though speed can be observed, a student finds it hard to estimate how the expert controls internal body variables like stiffness and force of fingers. These complexities can be further compounded by gender and culture biases of interpreting patient’s facial expressions as a feedback to test a medical hypothesis during examination. A systematic graduation from simple to complex scenarios of learning can potentially improve teaching and learning in medical education. Since the availability of the right patient cases within a learning period is beyond the control of students and trainers, a robotic patient with controllable and sensorised phantom organs to present key abnormalities like liver enlargement, appendicitis, gastrointestinal infections, and a controllable face with different ethnic and gender backgrounds to present facial expressions of pain during physical examination will allow us to develop a new teaching and learning paradigm in medical education. The key objective of this project is to understand what human participants do to maximize haptic information gain during a physical examination in different patient contexts. The primary patient contexts considered in this project are gender, age, and ethnicity rendered by the robotic face. The internal organ conditions will be presented using a robotic phantom abdomen already available. The student(s) can choose to project an animated face onto a soft robotic face and use a clever combination of animation and facial tissue actuation to present pain expressions to render a maximum sense of agency. An animated video projection could potentially allow presenting different gender and culture backgrounds of the patient without having to fabricate many soft robotic faces. Students will also measure palpation forces, movements, and muscle activity of the physical examiner using sensors already available in the laboratory.
  2. Robotic companions: Students will use the Pepper robot for this project. Robotic companions will likely be a useful part of society to address loneliness, medical therapy, and as domestic assistants. A robot can engage a human counterpart better if there is a way to estimate the level of engagement using some measurable variables. In this project, students will develop a tactile sensor kit that can be integrated on the hand of the Pepper robot, accelerometers, and a proximity sensor to monitor if there are any behavioral differences in the way male and female users interaction with the robot when the robot takes a male or a female gender. Students can dress the robot to give it a gender identity, and they can choose an interaction task like approaching and shaking the hand as a comparable context to see differences in user behaviors. Several projects can be made for different engagement tasks.
  3. Haptic feedback displays: Students will be briefed to design a wearable haptic feedback interface to allow a user to feel from a remote probe that examines a soft tissue to locate a buried hard nodule representing a clinical scenario of locating a tumor in a soft tissue. The user will be able to control the stiffness of the remote probe to improve haptic perception. This scenario allows the students to understand gender and culture-specific behaviors of users to improve haptic perception, and to understand best ways to design haptic feedback interfaces to best suit different user groups.
  4. Robotics guiders: Students will be briefed to use a scenario of guiding blindfolded participants using haptic feedback given via reins or the Pepper robot to understand how physical measurements can be used to estimate the psychological conditions like trust in the guider in different contexts of guiding. Our past results show that the forces felt by the guider due to changes in the degree of the voluntary following can be used to estimate the trust level of the follower in the guider. This provides a good experimental scenario to allow students to understand how gender and culture of followers exhibit different variations in the trust for the same guiding context. This information will be very important to inform design approaches to guiding robots for different user categories.

Papers to read

  1. Fong, T., Nourbakhsh, I., & Dautenhahn, K. (2003). A survey of socially interactive robots. Robotics and autonomous systems42(3-4), 143-166. PDF
  2. Lee, H. R., Sung, J., Šabanović, S., & Han, J. (2012, September). Cultural design of domestic robots: A study of user expectations in Korea and the United States. In 2012 IEEE RO-MAN: The 21st IEEE International Symposium on Robot and Human Interactive Communication (pp. 803-808). IEEE. PDF
  3. Lee, H. R., & Sabanović, S. (2014, March). Culturally variable preferences for robot design and use in South Korea, Turkey, and the United States. In Proceedings of the 2014 ACM/IEEE international conference on Human-robot interaction (pp. 17-24). ACM. PDF
  4. Carpenter, J., Davis, J. M., Erwin-Stewart, N., Lee, T. R., Bransford, J. D., & Vye, N. (2009). Gender representation and humanoid robots designed for domestic use. International Journal of Social Robotics1(3), 261. PDF
  5. Mumm, J., & Mutlu, B. (2011, March). Human-robot proxemics: physical and psychological distancing in human-robot interaction. In Proceedings of the 6th international conference on Human-robot interaction (pp. 331-338). ACM. PDF
  6. Libin, A. V., & Libin, E. V. (2004). Person-robot interactions from the robopsychologists’ point of view: The robotic psychology and robotherapy approach. Proceedings of the IEEE92(11), 1789-1803. PDF
  7. Blake, M. K., & Hanson, S. (2005). Rethinking innovation: context and gender. Environment and planning A37(4), 681-701. PDF

Weekly schedule: Each project will be discussed in a weekly 2-hour meeting. This will be the opportunity to discuss how to solve problems, related theory and techniques, related work done by others, report or paper writing, practicing presentations, troubleshooting, and any other project related topics. The first 3-weeks will be used to discuss some related research publications listed above, to refine a research question to be investigated, and to plan the resources and methods. The last 2-weeks will be spent on polishing the demonstration and the report.

  1. Week 1: Research paper discussion to set out the structure of a research project. Students go out and reflect on what it involves to do a research project – a. Formulate a research question, b. Plan a measurable method to test it, c. Obtain measurements, in this case using human participants with informed consent, d. Analyze the data, e. Interpret the results, and f. Arrive at conclusions.
  2. Week 2Discuss the project scopes outlined above (0.5 hr of brief, 1 hr of discussion about possible research questions, 0.5 hr of instructions to prepare a brief project proposal for the 3rd week). Students go out and make their groups, formulate research questions and experimental methods, sketch a Gantt chart. This will be submitted as a project proposal.
  3. Week 3: Students submit project proposals. Meet and discuss to help sharpen the research questions and methods. Students go out and come up with an experimentation plan and schedule. If there are collaborating PhD students involved, discuss their availability to do experiments.
  4. Weeks 4 – 92-hour Research meeting (5min update from each group followed by 10 mins discussion. Remainder of the time will be spent to discuss relevant analysis techniques, troubleshooting, sharing ideas, and feedback). Students are expected to spend on average about 8-hours per week doing experiments and analysis.
  5. Week 10Review reports, finalise analysis, interpretations, and conclusions.
  6. Week 11: Student groups do technical presentations (15 min conference style talks). A panel representing the Imperial Robotics Forum will give marks (25% of final grade).
  7. Week 12: Students submit an 8-page report following an IEEE conference paper format (75% of the final grade).

Learning outcomes: On successful completion of the module, students should be able to:

  • Apply robotics system design methods and biological principles to the development and construction of a novel intelligent robotic system.
  • Select and utilize appropriate software in the programming and demonstration of a robotic system.
  • Select and utilize appropriate fabrications techniques to develop and build a small robot that includes integrated electrical and mechanical components.
  • Discuss the opportunities and limitations of current robotics technologies.
  • Demonstrate an understanding of the potential scope of application for robotic systems.

The assessment will be in two parts. 1) A 15-minute viva: We will invite staff members of the Imperial Robotics Forum and members from the industry to join us to award marks for your 10 min presentation and 10 min viva. This will carry 25% of your final grade. An excellent presentation (A grade) will show a clear research question, a sound research methodology, concise results, deep interpretations, and clear conclusions. 2) An 8-page report: This should be written in a typical conference paper format. The report should have a clearly written abstract, an introduction citing related work with a clear description of the research objective, a clearly written methods section, a results section with plots and figures, a discussion, and a conclusion section. We will help you to write a good report. This report carries 75% of your final grade. If the paper is suitable for a peer-reviewed conference, we will further help you to go that extra mile, which will make a significant difference in your CV.

Resources available to you

Pepper robot

3D printers, standard sensors like force sensors, servo motors, 3D motion capturing systems, EMG sensors, our own soft sensors and actuators, force sensors, workshop area to fabricate things.

Related information

If you are interested in reading some related work done by my PhD students and postdocs, here are some links:

  1. J. Konstantinova, M. Li, M. Gautam, P. Dasgupta, K. Althoefer and T. Nanayakkara. “Behavioral Characteristics of Manual Palpation to Localize Hard Nodules in Soft Tissues”,  IEEE Transactions on Biomedical Engineering, pp. 1651 – 1659, vol. 16, no. 6, 2014. PDF
  2. Ranasinghe A, Dasgupta P, Althoefer K, Nanayakkara T (2015) Identification of Haptic Based Guiding Using Hard Reins. PLoS ONE 10(7): e0132020.  PDF
  3. Sornkarn N, Dasgupta P, Nanayakkara T (2016) Morphological Computation of Haptic Perception of a Controllable Stiffness Probe. PLoS ONE 11(6): e0156982.  PDF
  4. Nantachai Sornkarn and Thrishantha Nanayakkara, “Can a soft robotic probe use stiffness control like a human finger to improve efficacy of haptic perception?”, in press, IEEE Transactions on Haptics, 2016. PDF
  5. Jelizaveta Konstantinova, Giuseppe Cotugno, Prokar Dasgupta, Kaspar Althoefer, Thrishantha Nanayakkara, “Palpation Force Modulation Strategies to Identify Hard Regions in Soft Tissue Organs”, PLoS ONEPONE-D-16-31711R2 PDF
  6. Nicolas Herzig, Perla Maiolino, Fumiya Iida, Thrishantha Nanayakkara, “A Variable Stiffness Robotic Probe for Soft Tissue Palpation”, IEEE Robotics and Automation Letters (RA-L), pp. 1168 – 1175, vol. 3, issue 2, 2018. PDF

A talk I gave to a public audience about the challenges for robots to survive in natural environments:

A talk by Petar Kormushev on robot intelligence

A talk by Nicolas Rojas