Final year project ideas by ARQ (robotics)

In the following you can find a list of potential final year projects for final year (BEng/MEng/MSc) students.

MSc Students (Extended Research Project - DENM100)

Professor Kaspar Althoefer, k.althoefer@qmul.ac.uk

• Human integratable force and tactile sensor for medical application 
  There is growing interest to measure the forces that are involved when humans interact with their environment. This is of particular interest in the field of medicine where information on these interaction forces could be used for training purposes or to log relevant medical information while a clinician is conducting an examination, for example of soft tissue. The project will focus on developing force/tactile sensors that can be “integrated” with the hands of clinicians or the instruments they hold, to acquire distributed force/tactile information.  The project will make use of the group’s expertise in the field and the student is expected to closely collaborate with the supervisor's team (SAPHIR).

• Embedded tactile sensors based on optical tomography
  Tactile sensing technology based on optics has a number of advantages compared to traditional systems that use electrical signals for the transmission of tactile and force information, including safety when used in areas where humans can get into contact with the sensors, MR compatibility, reduced interference of and from other electrical systems in the vicinity. The student will design new tactile sensors made of transparent silicone and optoelectronic circuitry for the measurement of distributed tactile information. Embedding the sensor into surgical instruments will be attempted.  Interfacing the sensor for signal acquisition and classification is also an important part of this project. The student is expected to closely collaborate with the supervisor's team (SAPHIR).

• Snake-like continuum manipulator design and control
  Continuum-style robots, including those inspired by biology, increasingly arouse the attention of researchers due to the compliance capability and the wide range of motion. There are different continuum-style robot constructions, such as spring based design, concentric tube design, super-elastic rod/tube design, etc. The task here would be for the student to do explore different continuum robot designs, build a prototype and implement appropriate control strategies. The project is particularly suited for students with background in control and system interfacing. Having good programming skills could be beneficial too.

• Inflatable, growing robots
  The supervisor's team is involved in the creation of new robotics concepts: inflatable, growing robots, also called eversion or vine robots. These new types of robots have applications in construction, the nuclear industry as well as minimally invasive surgery. The project will focus on designing and realising prototypes. The student will be required to develop control strategies for these robots;  machine learning techniques can also be considered in this context. Support will be provided by the supervisor's team (SAPHIR).

• Robots for firefighting services
  This project is in collaboration with the West Midland Fire Service (WMFS). WMFS are very keen to employ robotics for their services. One area of interest is to develop robots that can reach into remote areas, such as a collapsed buildings or caves. These robots could then provide provisions and telecommunications means to people being trapped. In this project, growing or eversion robots will be investigated. The student will be expected to develop eversion robots and test those in realistic environments. Work is to be conducted in collaboration with WMFS. 

Dr Lorenzo Jamone, l.jamone@qmul.ac.uk

• Gentle robotic grasping of delicate objects ("Grasp me, tender, ...")
  Robotic arms and grippers are employed in industry to grasp (i.e. pick) objects and move them around. However, those are typically hard objects (e.g. metal parts) and the gripper can apply large forces to pick them successfully. Instead, humans use the sense of touch to apply just the right amount of force, and therefore they can easily pick delicate objects as well, gently (tender!). You will develop simple tactile sensors and integrate them in a small robotic arm+gripper, and program a control strategy for the autonomous grasping of delicate objects, e.g. a raspberry. Examples of our innovative tactile sensors here, here and here. Electronics integration and good programming skills are required for this task. Note: this is a VERY CHALLENGING PROJECT, especially if you do not have any robotics/electronics/control background.
• Embodied Artificial Intelligence: a baby robot learning how to move
  How do we control our body movements? At birth, babies are not able to precisely control their bodies, e.g. touching their head with the right hand; however, during early childhood (and already prenatally, in the womb), they start moving the limbs and touching their own body and the external environment, gradually learning how to move with purpose. In this project you will develop exploration strategies for a “baby robot” to gradually learn a body model from multimodal sensory data (touch, proprioception, vision). The robot learning might be based on ML/AI techniques such as Reinforcement Learning, Deep Learning, Autoencoders. Good programming skills (Python) and some expertise in ML/AI are required for this project; expertise in robotics and computer vision is a plus.
• Data Glove with fingers tracking and tactile feedback for fully immersive VR and telerobotics applications
  With the booming interest in VR and robotics, wearable input devices for human-computer interaction are rapidly evolving; however, solutions to provide finger tracking and haptic/tactile feedback to the human user are costly and not always very effective, ending up in expensive and/or cumbersome devices. This project will explore novel technologies and design solutions to create a comfortable and nice-looking data glove that can: 1) measure human finger movements and 2) provide haptic/tactile feedback to the user. Recommended solutions are: flex sensors (for task 1) and vibrating motors (for task 2). The potential application could be something like this. Electronics integration and good programming skills are required for this task; expertise in design is a plus.
• Tactile intelligence for dexterous robots
  Humans can recognize objects just by touching them. We can estimate the shape of objects, their size, their texture/material, and other properties. Novel sensors developed at QMUL can be used to provide robots with the sense of touch. The data coming from these tactile sensors must be analyzed with machine learning techniques to make sense of it. The objective of the project is to train classifiers based on tactile data, so that specific object properties can be estimated and object identity can be recognized. The project requires good programming and machine learning skills; previous knowledge of tactile sensing and robotics is optional. For reference, you can see the output of a previous MSc project on this topic.

Dr Ketao Zhang, ketao.zhang@qmul.ac.uk

• Multi-fingered robotic gripper for agriculture applications
  Underactuation is a widely used concept in various designs for robotic grippers where the mechanical systems are capable of adapting to the shape of objects with a less complex control system. This project will investigate a novel approach for developing underactuated robotic grippers with variable stiffness joints for grasping and manipulation towards agriculture applications. The main objectives are: 1)Design and modelling of flexure hinges and reconfigurable modules for a multi-fingered robotics gripper;  2) Develop actuation schemes and control systems; 3) Build a light-weight and low-cost robotic gripper with integrated force and vision sensors; 4) Evaluate the grasping performance of the gripper in different grasp modes.
• Perching small aerial robots with electroadhensive (EA) anchors
  Development of the anchoring mechanisms and EA pad patterns with limitation of the size (40 mm x 40 mm) to make it force maximize (can be found in reference papers). Implement simulations/experimental tests of the different patterns designed and optimize the pattern using software. The main objectives are: 1) Optimisation of pattern design of the pad; 2) Exploring a novel design of quick releasing for application; 3) Experimental tests in various materials of substrates; 4) Integration of the pad with a free-form structure and small aerial robots (multirotor UAVs) for perching on complex structures.
• Design and control of a four-legged robot prototype
  Among mobile robotic research field, legged locomotion is largely applied for advanced robotic systems due to the higher degree of versatility compared to wheeled robots, which allows them to successfully move and interact in unstructured environments; nevertheless, legged robots present several designing problems and require a much more complex control system. This project will investigate a small-size 2-DOF mechanism for a legged robot that is aimed to achieve high performances in both velocity and versatility. The main objectives are: 1) Exploring the motion characteristics of the closed-loop mechanism; 2) Reveal the various configurations of the mechanism; 3) Modelling and simulation of kinematics of the closed loop mechanism; 4) Develop a prototype of the leg mechanism with smart servo actuators; 5) Validating the kinematic performance of the leg mechanism with experimental tests.
• A 5-DoF Parallel Robot for Additive Manufacturing
  Additive manufacturing is becoming a bigger element in new technology development. The purpose of this research is to consider the feasibility of a 3-DOF parallel platform to be used for additive manufacturing processes. This project aims to design a new 3-DOF parallel robot and develop a proof of concept prototype aimed at advanced additive manufacturing. The main objectives are: 1) Kinematic design and analysis of the 3-DOF parallel robot; 2) Develop 3D model of the parallel robot using SolidWorks; 3) Prototype development of the parallel by integrating the actuation system and creating kinematics-based control algorithms.

Dr Hasan Shaheed, m.h.shaheed@qmul.ac.uk

• TBC

Dr Thilina Dulantha Lalitharatne, t.lalitharatne@qmul.ac.uk

• Soft haptic glove for physical examination
  The physical examination methods such as palpation used by physicians/GPs involves examining a patient to make initial diagnoses of issues such as an enlarged liver, tumours, and abnormalities in the abdomen, breast, etc. This project aims to develop a soft haptic glove capable of tactile sensing to measure the forces physicians use when palpating, investigate palpation behaviour, and apply it in remote palpation applications. Required skills: good hardware/mechanical fabrication skills, instrumentation, prior experience in the use of soft fabrication will be an advantage.
• Brain signal controlled assistive robots
  This project will explore the use of brain-computer interfaces (BCI) to control assistive robots such as meal-assistance robots and prosthetic hands designed for disabled individuals. We will use the g.tec unicorn hybrid black (https://www.unicorn-bi.com) EEG signal acquisition system to acquire the brain signals and explore the use of AI or machine learning (ML) techniques to control the assistive robots, which are already developed and available in the lab, in real-time. Required skills: good programming skills (MATLAB/Python/Java/C/C++), mechatronics, prior experience in signal processing and ML will be an advantage.
• Development of a pneumatically controlled soft hand for a humanoid social robot to generate of humanlike hand gestures
  Hand gestures are an essential part of human social communication. These gestures are typically culturally diverse and convey various meanings. Likewise, it is crucial to equip humanoid social robots with humanlike hand gestures. Therefore, in this project, we will develop a pneumatically controlled soft hand capable of generating humanlike hand gestures. Required skills: good hardware/mechanical fabrication skills, mechatronics, prior experience in the use of soft fabrication, pneumatic systems will be an advantage.
• Development of a soft social robot
  The primary aim of this project is to develop an abstract modular soft social robot. Each modular part will add a degree of freedom (DoF) to the social robot. Through a series of user studies, we will systematically assess the impact of adding more DoFs and the level of abstraction on human-robot social interaction. Required skills: good hardware/mechanical fabrication skills, mechatronics, prior experience in the use of soft fabrication, conducting user studies will be an advantage.

UG Students (Final Year Project)

Dr Ildar Farkhatdinov, i.farkhatdinov@qmul.ac.uk

• rehabilitation robotics (exoskeletons, wearable robots)
• control for physical human-robot interaction
• virtual  and augmented reality interfaces for remote robot control
• haptics and sense of touch for virtual reality and telerobotics
• machine learning for human electromyography data analysis
• wearable robotic tail for balancing control
• haptic motion simulator for virtual reality
• robotic tilting platform for vestibular research with animals
• graphical user interfaces and gamification for rehabilitation robots

Dr Thilina Dulantha Lalitharatne, t.lalitharatne@qmul.ac.uk

• humanlike social robots for human-robot social interactions
• robotic simulators for medical training
• robotic doctors for remote medical examinations
• control of assistive robots (eg. prosthetics, meal assistance robots) using brain signals (BCI/BMI)
• bioinspired underwater robots