How to use this list
This list is for master's and doctoral researchers in electrical and electronic engineering (EEE) who need a project topic.
Treat every row as a starting point. Before you commit, check what equipment, software licences and data your department can give you, and search for studies published in the last three years. Related engineering pages: engineering dissertation titles and mechanical engineering topics.
Theme 1: Renewable generation and power systems
Power systems research has shifted from large synchronous machines towards networks dominated by converters, solar and wind. That shift makes classic questions (tracking, inverter choice, transmission, transients) worth asking again with current devices and data. Wide bandgap converters are now being reviewed specifically for microgrids, which shows how quickly the device layer is changing (Burmester et al., 2026).
| Topic | Possible research question | Likely design and data | Why it matters now |
|---|---|---|---|
| Dual-axis solar tracking | How much more energy does a dual-axis tracker deliver than single-axis and fixed mounting at one site, and does the gain cover its extra cost and energy use? | Side-by-side field measurement over several months, or a validated model with local irradiance data | Tracking gains depend strongly on climate and latitude, so local evidence is needed |
| Micro-inverters versus string inverters for rooftop PV | Under partial shading, which architecture yields more energy, and what are the efficiency and reliability trade-offs? | Bench or rooftop comparison with controlled shading; efficiency curves; failure data where available | Inverter choice drives yield and lifetime cost in small PV systems |
| Voltage-source-converter HVDC link | How does a VSC-HVDC link respond to an onshore AC fault, and which control settings limit the disturbance? | Simulation in an established power system tool with a published benchmark network | HVDC is the usual route for connecting distant offshore wind |
| Switching transients with inverter-based generation | How do capacitor-bank and breaker switching transients change as inverter-based generation rises in a distribution network? | Electromagnetic transient (EMT) simulation of a benchmark feeder, varying the share of inverter-based generation | Protection settings designed for synchronous machines may not suit converter-rich networks |
Theme 2: Power electronics and devices
Silicon carbide (SiC) and gallium nitride (GaN) devices switch faster and with lower losses than silicon, which allows higher frequencies and smaller passive components; SiC is generally preferred above about 650 V and GaN below it, although the devices still cost more than silicon (Kumar et al., 2022). Comparing a silicon IGBT with a SiC MOSFET in the same circuit is a well-defined experiment (Kinyua & Aoki, 2025). Reliability is a live question too: in one modelling study of a photovoltaic string inverter, the GaN device showed about five times more plastic energy dissipation in its die attach under cloudy conditions than a SiC MOSFET, despite lower losses, which the authors attributed to its smaller package (Van De Sande et al., 2020).
| Topic | Possible research question | Likely design and data | Why it matters now |
|---|---|---|---|
| Silicon versus wide bandgap devices in a converter stage | How do switching loss, efficiency and junction temperature compare when a Si IGBT is replaced by a SiC MOSFET (or a GaN HEMT) in the same converter? | Double-pulse tests and efficiency measurement on one test board; thermal imaging or thermal model | Device choice is now a design decision, not a default (Kumar et al., 2022; Kinyua & Aoki, 2025) |
| Output ripple in a DC-DC converter | Which combination of switching frequency, filter design and control method minimises output ripple for a given efficiency target? | Simulation followed by a prototype buck or boost converter with oscilloscope measurement | Ripple limits sensitive loads such as sensors and processors |
| Resonant inductive wireless power transfer | How do efficiency and delivered power change with coil distance and misalignment, and which compensation network is most tolerant? | Bench rig with measured coil parameters; comparison of two compensation topologies | Wireless charging is spreading from phones to vehicles and implants |
| Thermomagnetic energy conversion | Can a bench-scale thermomagnetic generator using a magnetic phase-change material produce measurable power from low-grade heat, and what limits it? | Model of the magnetic and thermal cycle, then a small proof-of-concept rig if materials are available | Low-grade waste heat is plentiful but hard to convert |
Theme 3: Control and embedded systems
Many good EEE projects combine sensing, computation and actuation in one small system. Framing such a system as a cyber-physical system, with layers from data acquisition to decision-making, helps you decide what to build and what to measure (Lee et al., 2015). For connected devices, a reference architecture gives a shared vocabulary for sensing, networking and application domains (ISO/IEC, 2024).
| Topic | Possible research question | Likely design and data | Why it matters now |
|---|---|---|---|
| PID versus model predictive control of a DC motor drive | Does model predictive control improve speed tracking and disturbance rejection enough to justify its computational load on a low-cost microcontroller? | Hardware-in-the-loop or bench experiment with identical load steps; tracking error and execution time | Low-cost processors now make advanced control practical in small drives |
| Finite-state-machine controller on an FPGA | Can a safety-relevant controller (for example, traffic lights or a lift) be specified as a finite-state machine and formally verified before deployment? | HDL design, simulation, formal property checking, then FPGA implementation | Formal methods catch design errors that testing misses |
| Automatic control of vessel navigation lights | Can a low-power controller switch anchor and navigation lights automatically and report lamp faults reliably in marine conditions? | Prototype with light and current sensing; environmental testing; fault-injection trials | Small vessels often rely on manual switching and checks |
| Sensor fusion for a small mobile robot | How much does fusing wheel odometry, an inertial sensor and a low-cost range sensor reduce localisation error indoors? | Repeated runs on a marked course with ground-truth positions; Kalman filter variants compared | Affordable autonomy depends on cheap sensors used well |
| Biometric authentication in an e-voting prototype | What false-acceptance and false-rejection rates, and what usability problems, does a fingerprint-based voter check show in a supervised trial? | Prototype plus usability study with consenting volunteers; no real elections or real voter data | Election technology needs evidence on security and accessibility together |
Theme 4: Signal processing and sensing
Signal processing topics work well when the signal, the hardware and the performance measure are all named. The four rows below each pair an algorithm or circuit with a concrete measurement, from audio noise to the voltage and current traces of an electrical discharge machine.
| Topic | Possible research question | Likely design and data | Why it matters now |
|---|---|---|---|
| Real-time audio noise reduction on low-cost hardware | Which noise-reduction method gives the best quality for its processing delay on a microcontroller or small DSP? | Implementation of two or three methods; objective quality scores and listening tests on recorded speech | Hearing aids and voice interfaces need low delay and low power |
| Low-noise JFET input stage for an audio filter | Does a JFET input stage give lower noise and distortion than an op-amp-only design in an active audio filter? | Circuit design, simulation, then prototype measurement of noise floor and total harmonic distortion | High-input-impedance stages matter for sensors and audio sources |
| Embedded vision for early warning | Can an edge device detect smoke or intrusion in camera images with acceptable accuracy and delay, while keeping images on the device? | Public or consented datasets; compressed model on edge hardware; accuracy, latency and power measured | On-device processing reduces bandwidth use and privacy risk |
| Monitoring electrical discharge machining from process signals | Can features of discharge voltage and current signals classify stable and unstable machining in real time? | Signals logged on an EDM machine under planned settings; feature extraction and classification | Process monitoring can cut scrap and electrode wear |
Theme 5: Flexible, printed and optoelectronic devices
Device topics need access to fabrication or characterisation facilities, so confirm these first. The rows below are deliberately narrow: each tests one property of one device type under stated conditions.
| Topic | Possible research question | Likely design and data | Why it matters now |
|---|---|---|---|
| Printed interconnects in flexible hybrid electronics | How does the resistance of printed conductive tracks joining silicon chips change under repeated bending? | Printed test structures; cyclic bending rig; resistance logged against cycle count | Flexible hybrid devices combine printed tracks with conventional chips |
| 3D-printed conductive hydrogel strain sensors | How stable are the conductivity and sensitivity of a printed conductive hydrogel sensor over repeated strain and drying? | Fabrication of sensor samples; cyclic strain tests; humidity-controlled ageing | Soft sensors are needed for wearables and soft robotics |
| OLED test devices under drive conditions | How do drive current and temperature affect luminance decay and efficiency of OLED test devices? | Accelerated ageing of fabricated or purchased test devices; luminance and current-voltage measurement | Lifetime remains a limit for some OLED applications |
How to test a topic before you commit
Use the FINER criteria: feasible, interesting, novel, ethical and relevant (Hulley et al., 2013). In EEE, feasibility usually decides the matter. List the equipment, licences and components you need, check lead times, and agree laboratory safety arrangements for high voltage or high current work before you start. Plan how you will store and share data and code; research funders such as the EPSRC expect research data to be preserved and made accessible where appropriate (EPSRC, 2022).
For the wider process, see how to choose a PhD topic, how to select a PhD topic and PhD topic selection support.
Common mistakes
- Choosing a subject heading ("OLEDs", "control systems") instead of a question with a measurable outcome.
- Relying on simulation alone, without validation against measurement or published data.
- Planning a prototype that needs equipment or components you cannot obtain in time.
- Comparing two designs under different conditions, so the comparison is not fair.
- Leaving high-voltage safety approval and data management until the experiments are due.
Key takeaways
- A topic becomes researchable when it names a system, a measured outcome and a fair comparison.
- Wide bandgap devices have turned device choice into a research question in power electronics (Kumar et al., 2022; Kinyua & Aoki, 2025).
- Reliability, not only efficiency, deserves measurement in converter studies (Van De Sande et al., 2020).
- Architectures for cyber-physical and connected systems help scope embedded projects (Lee et al., 2015; ISO/IEC, 2024).
- Check feasibility, safety and data plans before committing to an EEE topic (Hulley et al., 2013; EPSRC, 2022).
