FYP / MSc projects

Deployable structures are mechanical systems that transform from a compact stowed configuration to a larger operational one, and they are critical in applications ranging from spacecraft antennas and solar arrays to biomedical implants. Recent work has shown that wafer-scale microfabrication offers a powerful route to producing planar precursors that mechanically deploy into complex 3D geometries, combining the precision of semiconductor processing with the programmability of structural mechanics. However, exploiting microfabrication for deployable structures at Imperial requires first establishing a baseline process flow using the College cleanroom facilities.

This project will familiarise the student with Imperial's microfabrication capabilities and use them to produce a set of bistable deployable micro-structures. The student will design planar precursors with engineered stress or geometric features that snap into a stable 3D configuration upon release, fabricate them using photolithography and etching, and characterise their deployment behaviour. The project will produce both a documented process recipe for use by future students and a working demonstrator of micro-scale bistable deployment.

A willingness to work in a cleanroom environment is essential, along with basic proficiency in CAD and Python. No prior microfabrication experience is required. For inspiration, see previous results from our group.

Automated repair of damaged structural surfaces - aerospace components, industrial tooling, or infrastructure - is a long-standing goal of advanced manufacturing. The core challenge is the tight coupling between sensing, motion planning, and material deposition: the robot must identify the damaged region, generate an appropriate toolpath, and deposit material in a way that conforms to the existing geometry.

This project will develop the hardware and control pipeline for a robotic extrusion-deposition system aimed at local surface repair. The student will interface a robotic arm with an extrusion head, develop motion-control software that follows a specified toolpath, and integrate the system end-to-end with input from a 3D scan of a damaged region. A simplified curvature-based metric (provided) will be used to identify the repair region, and the student will demonstrate closed-loop deposition on a set of test specimens with engineered defects. The project suits a student with strong interests in robotics, mechatronics, and hands-on system integration, and requires good Python skills. Prior exposure to ROS is helpful but not required.

Robotic manipulation of soft, yield-stress materials such as clay sits at an unusual intersection of mechanics, control, and craft. It is also a domain in which contemporary artists - most famously in works such as Sun Yuan and Peng Yu's Can't Help Myself - have used robots not to perform a utilitarian task, but to stage an open-ended, absurd, or meditative action. This project embraces that framing: a robot that engages repeatedly and imperfectly with a ceramic medium, producing an artefact whose value lies in the process rather than the end product.

Project 3a: hardware and motion control

This project will design, build, and operate a robotic system that performs a repeated shaping or throwing action on clay. The student will interface a robotic arm with a simple end effector, program a control routine that executes an open-ended shaping behaviour, and document the resulting artefacts and process. The engineering contributions are the hardware integration, the motion-control routine, and an analysis of how the clay's rheology constrains the achievable motions; the broader output is a physical art piece that can be exhibited. The project suits a student with strong mechatronics and coding skills who is interested in the intersection of engineering and art. A willingness to iterate experimentally and tolerate messy outcomes is essential.

Project 3b: perception and behavioural control

A robot that repeatedly engages with a deformable medium - clay, for example - requires a perception system that tracks the state of the material and adapts its actions accordingly. When the goal is an open-ended, aesthetically driven interaction rather than a specified end shape, the perception and control problem is particularly interesting: the robot is not pursuing a target but reacting to what is there, producing behaviour that is responsive without being goal-directed in a conventional sense.

This project will develop the perception and behavioural-control layer for a robotic ceramic art installation. The student will integrate a camera system with a robotic arm performing shaping operations on clay, train a computer-vision pipeline to track the geometry and deformation of the clay in real time, and design a behavioural control loop in which the robot's actions are modulated by what it perceives. The work builds on the parallel hardware-focused project (3a) and offers the opportunity to contribute to an exhibited artefact as well as a technical thesis. The project requires strong programming skills in Python, familiarity with computer vision and/or deep learning, and an interest in robotics. Willingness to engage with the artistic framing of the project is essential.

Modular reconfigurable materials - imagine a collection of small units that can rearrange themselves into different global configurations - have long been a goal of robotics and programmable-matter research, with the microbots of Big Hero 6 serving as a popular cultural reference. Achieving this in practice requires modules that can couple to their neighbours, change their coupling state on command, and slide or rotate relative to each other under some form of local interaction. This project focuses on the fundamental building block: a single module and its interaction with one or two neighbours.

This project will design, fabricate, and characterise a reconfigurable module with a controllable local interaction. The student will propose and prototype a coupling mechanism (magnetic, mechanical, electro-adhesive, or other), characterise its engagement and release behaviour, and demonstrate at least one reconfiguration action - typically sliding, rotation, or snap-attach - between a pair of modules. The project will produce a working prototype, a characterised coupling mechanism, and a proposed scaling strategy for larger assemblies.

The project suits a student with strong design and fabrication instincts and an interest in mechatronics. Prior exposure to CAD and basic electronics is expected.

Textiles - both traditional (knits, weaves) and non-traditional (chainmails, crochets) - derive much of their mechanical behaviour from frictional interactions between the filaments that make them up. Additive manufacturing opens the possibility of designing filaments with programmed surface textures, which could in turn programme the frictional and therefore mechanical response of an entire textile. Understanding this connection requires accurate characterisation of filament-scale friction, including how it depends on filament surface topography at the microscopic scale.

This project will characterise frictional behaviour of 3D-printed filaments with a range of deliberately varied surface textures. The student will design and print filament specimens with controlled surface features, develop a benchtop friction-testing setup suitable for filament-on-filament contact, and quantify the dependence of friction on texture, load, and sliding direction. The resulting dataset will form the foundation for subsequent textile-design work.

The project suits a student with interests in experimental mechanics, additive manufacturing, and tribology. Good programming skills in Python are expected for data processing.

Transforming a flat sheet into a target three-dimensional shape by prescribing local length changes is a classical problem in differential geometry - and one with direct implications for deployable structures. A triangular grid whose edges can individually extend or contract provides an experimental realisation of this idea: by commanding each edge to a prescribed length, an initially planar grid deploys into a target 3D surface. When the prescribed lengths are derived from a conformal map, the deployed shape approximates a target surface in a geometrically controlled way.

This project will build and demonstrate a robotic deployable meta-surface based on a triangular grid with individually actuated edges. The student will design a modular triangular unit with an extensible edge actuator (linear actuator, shape-memory alloy, or equivalent), assemble a small grid of these units, and implement open-loop control that commands each edge to a specified length. A conformal-map-derived length field (provided or computed using existing tools) will be used to deploy the grid into a target surface. The project will produce a working demonstrator and a characterised mapping from commanded lengths to deployed geometry.

The project suits a student with strong design, fabrication, and coding skills. Prior exposure to robotics or Grasshopper is advantageous but not required.

Interlocking plate assemblies - in which flat elements are arranged so that frictional and geometric contact produces a coherent structural response - offer a route to reconfigurable, demountable, and topologically interlocked structures. When the interlocking is engineered together with a bistable mechanical response, the assembly acquires discrete stable configurations, switchable by applied load, that could be exploited for deployment, locking, or energy absorption.

This project will design, fabricate, and characterise a set of frictional interlocking plates with an engineered bistable response. The student will iterate on plate geometry and contact features, fabricate specimens using 3D printing or laser cutting, and characterise the snap-through behaviour and the conditions under which interlocking is preserved across snap events. The project will produce a documented design with a characterised force-displacement response and a demonstration of controlled switching between stable configurations.

The project suits a student with strong design and hands-on instincts. Good CAD skills are expected; prior experience with mechanical testing is beneficial.