Present - FMP
U&S
FocusOn focuses on neurodivergent employees (ADHD and autism) in shared offices, typically designed around the neurotypical average (PI/V). I worked towards an inclusive, needs-based design, beginning with a literature review, user interviews and an expert interview, letting task initiation and distractions emerge as the central themes. I translated these into scenarios and claims (scenario-based design) to understand the users' problems and identify design opportunities (PI/V). I then ran a study using an adapted Co-Constructing Stories method, where participants experienced a task starter and three LED conditions, followed by interviews. Interview data were processed with thematic analysis, and video data visualised in TU/e's Data Foundry to triangulate self-reported experiences with observed behaviour (MD&C). Combining thematic analysis with the scenario claims sharpened my analysis, helping me select which themes to take forward.
A second focus was learning to actively involve stakeholders. I set myself the intention to do this and made a stakeholder map to guide me, since I knew little about neurodiversity myself. I asked each expert to point me towards others (snowball sampling), which gradually expanded my network.
Critically, I learned that involving the right expertise is about when as much as who. Bringing in a light expert on the artistic and realisation side was valuable, but I should have added research-oriented expertise on visual ergonomics and luminance ratios (see T&R) earlier, alongside it.
T&R
T&R connected closely to U&S and C&A: I first realised a working test version to evaluate the concept with users (U&S), and the controller–user interaction, such as feedback when a button is pressed, shaped the experience (C&A). This connects to my PI/V, in which bridging T&R and U&S is central (PI/V).
I built a fully functional, real-time system around two ESP32 microcontrollers communicating over ESP-NOW. My main learning was the systematic approach: I mapped the communication in a flowchart, tested each component in isolation, then integrated step by step. I used AI to understand how the components and ESP-NOW worked and to support debugging, but I directed the process and kept the overview, so I could build on it independently.
The realisation went beyond software. Working from each component's power requirements, I calculated that the battery needed a boost converter for a stable 5V. I modelled the assembly in SolidWorks to keep the controller compact, printed the housing in PLA (tested to transfer capacitive touch), and soldered the wiring onto a compact PCB insulated against short circuits. The battery is removable and rechargeable for energy efficiency. For the light, a bare LED strip proved too subtle in a pilot, so after testing white PLA I chose an acrylic diffuser that reflects light off the housing before it passes through (expert advice, U&S): unlike PLA, acrylic let me enlarge the housing for better desk integration while needing less brightness, saving energy and giving a wider brightness range and more control (PI/V).
Critically, I explored how light communicates but did not verify its physical side: I never measured the luminance and contrast ratios in a lab, as the expert's standards came too late (see U&S). In the user test no one reacted negatively to the brightness, only the green colour, so it seems to support rather than irritate, but fatigue from luminance often works unconsciously, which only a lab measurement could confirm (PI/V).
C&A
C&A is where my main goal for this semester came together: exploring how light can serve as a means of communication. At the start I was unsure whether to adjust the light in the space or use it as a modality to communicate with, and through literature, an expert conversation and user input I moved towards the latter.
During ideation I explored how light could communicate, through brightness, movement, colour, and form, using techniques such as brainstorms per theme, a morphological chart, and scenario-based design. I moved between intuition and theory: generating light behaviours and checking them against principles, and at other times translating a principle directly into a behaviour. One example is the continuous stimulation in focus mode, which came from combining theory, experts and users (optimal stimulation theory (Zentall & Zentall, 1983), the designer of Breathing Light, and participants who described how interest in a tool fades): a slowly shifting, water-like light that stays engaging without demanding attention. The controller form was deliberately based on familiar products (a kitchen timer, a tap dial) so the interaction would read intuitively.
My key learning is that light spans a wide communicative range: a small peripheral change can steer someone unconsciously, as several participants responded to the brightness in C1 without noticing it, whereas a stronger signal explicitly redirects them. This makes it essential to choose where on that range to sit, which is why personalisation matters (PI/V). A point for development is the controller's intuitiveness: the buttons could communicate their function more clearly, and it was not obvious that you slide your finger along the curve to adjust brightness. These are affordance and legibility, central to my PI/V, that I would refine next (PI/V).
B&E
B&E was a smaller part of this project, but it shaped how I positioned the design for Ahrend. To make the value explicit I used a Value Proposition Canvas, mapping the jobs, pains and gains of a neurodivergent employee (see U&S) against what the system offers. I positioned the concept with an ERRC grid and a benchmark of competing products, which showed most current tools address only one theme and stay static, defining the gap FocusOn fills.
A key consideration was that impact runs through Ahrend: however strong the design, it only reaches users if Ahrend can carry it. I therefore designed it to fit their existing furniture rather than as a standalone device, lowering the barrier to adoption. This connects to my vision of creating designs that genuinely reach the people the standard product passes by, which means designing not only for the user but also for the party that can bring it to them (PI/V).
I also handled risks as part of the process, social visibility in a shared space, habituation over long-term use, and real-world adoption, which I mapped through the user test and translated into design choices.
MD&C
MDC mostly ran underneath the realisation in T&R, focusing on the underlying computing and mathematics. Before building the system I mapped its behaviour in flowcharts, structuring it as a finite state machine with three states (default, session, break) that the system moves between on events such as starting a session or the timer expiring (T&R). For head detection I used MediaPipe, which recognises facial landmarks; from its output I used the roll angle and set a threshold, calibrated by measuring realistic values from a first-person perspective (C&A), to detect when the user looked away or back.
The animations rely on mathematical curves: combined sine functions for the water animation, a cosine for the brightness pulse, and a quadratic falloff for the light ball. I worked from existing examples, adjusting the parameters until the movement felt natural and coming to understand how each curve shaped the light (C&A).
For data, I quantified in Excel how often themes recurred across the scenarios and visualised this to decide which to take forward. I used TU/e's Data Foundry to log and visualise the video observations and triangulate them with the interviews (U&S). I kept the code structured so it stays readable for further development.