I explored how a passenger could contribute navigation, hazard and journey information without becoming another interruption — then translated that problem into a connected Passenger Display + Driver HUD interaction model.
Drivers and passengers already regulate communication in context. The design opportunity was to make that informal behaviour visible, structured and less dependent on guesswork.
Automotive interfaces often concentrate interaction around the driver. Research suggested a different role split: let the passenger handle information-heavy secondary tasks, while the driver receives only the information that needs their attention (Ma et al., 2024; Berger et al., 2023).
How might I let passengers contribute useful information while protecting the driver's attention and decision-making authority?
At higher workload, the passenger had to guess when to speak while the driver had no visible timing signal or feedback loop.
Low load. Navigation is already set.
Merge, roundabout or unfamiliar junction.
Roadblock / route change is spotted.
Input is dismissed, absorbed or overridden.
Useful information arrives too late.
No structured debrief; the pattern repeats.
This was an individual MSc major project. I owned the research, synthesis, interaction architecture, UI, coded prototype and validation.
Because this was a concept project rather than a shipped product, I did not have a commercial KPI. I defined success around behaviour: could the passenger take on richer secondary tasks while the driver received concise, understandable information with limited visual demand?
These are the moments where evidence forced the design to change, including an option I considered and rejected.
The passenger has more freedom to explore information; the driver has less visual bandwidth. The architecture needed to reflect that asymmetry.
Passenger displays can shift information-heavy tasks such as navigation and media away from the driver (Ma et al., 2024).
Passenger Display for explore / prepare / submit; Driver HUD for concise, driver-relevant delivery.
One shared centre-stack for both occupants. It keeps both roles competing for the same interaction space.
The first HUD hierarchy made the hazard cue too close to the speedometer, creating an unintended association.
In low-fi testing, 60% of drivers interpreted the earlier hazard placement as related to the speed information.
Separate the hazard spatially and keep the category explicit with icon + label pairing.
Keep the compact hazard cue attached to the speedometer to save space.
A simple Send action hid too much state. Both people needed to understand whether a request was delivered, waiting or timed out.
Participants wanted visible communication status rather than silent suppression after sending.
Visible states plus an 8-second timeout when the driver is occupied, followed by a clear re-send path.
A one-way “Send” button with no feedback loop.
The final system keeps a shared visual language, but gives each occupant a distinct interaction job.
The passenger can research a destination, inspect route context and decide when to propose it to the driver.
A route change becomes a structured handoff rather than a verbal instruction alone.
The passenger sees when a request has reached the driver layer and how long it will remain active.
Four explicit hazard categories keep safety-critical information distinct from routine journey requests.
The two interfaces share typography, iconography, spacing and semantic colour, while the interaction model remains role-specific. Visual hierarchy and glanceability were designed with automotive human-factors guidance in mind (NHTSA, 2013; BSI, 2017; ISO 2575:2021; ISO/TS 21957:2023).
The final evaluation combined usability, workload, eye tracking, reaction time and safety-compliance measures in a controlled two-screen prototype.
The project became more useful as the constraints became explicit. Those constraints now define what I would test next.
The evaluation used a controlled browser-based simulation. Real traffic, vehicle motion, glare, vibration and genuine driving consequences were not represented.
I would test richer workload-aware gating, notification persistence and physical in-vehicle conditions before considering production-level implementation.
The low-fi hazard misunderstanding changed the final hierarchy. In automotive UX, spatial placement can change the meaning people infer.