HoloNav
Alex Pina (ID), Julian Cunningham (ID), Kareen Yeung (ID, ME), Levi Tucker (ECE), Trevor Goo (ECE), George Lee (ECE)
My Role: Research Phase Lead · UX Lead · UI Design · User Testing
The Problem
As city dwellers turn to micro-mobility solutions, they navigate in streets not often designed to
include them.
How might we identify, measure, and prevent risk?
The Brief
Design an augmented-reality navigation solution for personal electric vehicles (PEV) that is safer than navigating with a phone.
How might we minimize distraction?
50%
of traffic injury cases, phone use was involved within one minute of the incident
28%
of ride time is spent looking at a phone-mounted screen while cycling
87%
of PEV owners surveyed would spend $100+ on AR navigation technology
Discovery
Scope: Primary Research, Stakeholder Identification, Usage Scenario, Value Proposition
Methods: Literature Reviews, Interviews, Surveys, White Space Analysis, Personas, Journey Mapping
We started broad — exploring the characteristics and needs of micro-mobility users through interviews, field observations, secondary research, and a survey of PEV owners via Prolific. Our research surfaced a clear pattern: Phone-based navigation methods are dangerous because they are distracting, and there is an opportunity to create a less distracting alternative.
1. Distraction is the core risk
Studies showed riders using phone-mounted navigation spent significant portions of their ride looking away from the road, which is directly correlated with increased accident rates.
Prolific Survey · n=30
2. PEV Information Systems Vary
Handlebar vehicles provide real-estate for screens and add-ons, making them reliable for commuters. Electric skateboards and unicycles are more popular amongst hobbyists. Their lack of handlebar support requires companion apps/watches for niche performance tracking.
3. No AR solution exists for this niche
Competitors like REKKIE, FORM, and ENGO 2 are locked into other sports.
No one targets PEV riders with navigation-first AR.
Engo 2
Running — burn, heart rate, distance
REKKIE
Skiing — slope, altitude, speed
FORM
Swimming — pace, stroke rate, splits
Prototypes & Testing
Approach: Rapid iteration — build, test, measure distraction, redesign
Metrics: Eyes-off-road time, Dwell time per glance, Qualitative exit surveys
We built 3 generations of prototypes in 14 weeks, each one testing a specific hypothesis about how to reduce rider distraction.
Generation 1 → 1.5
Helmet-Mounted Opaque Display
Our first prototype used an opaque TFT screen on an adjustable helmet arm, connected to a phone via WiFi.
Initial test results were inconclusive — so we refined our methodology, filming riders' faces to capture glance frequency, glance duration, and total eyes-off-road time.
Gen 1.5 results showed the head-mounted display was less distracting than a phone mount:
5.8% eyes-off-road vs. 7.3% for the control, with 0.33s dwell time per glance vs. 0.65s.
Gen 1 prototype — opaque TFT display mounted on a helmet, tested in Piedmont Park.
The control and experimental rides rotated across 2 routes, minimizing environmental bias.
Generation 2+3
AR Smart Glasses — Brilliant Frames
Our final working prototype used the Brilliant Frame, an open-source AR glasses platform, running custom navigation software. Our team of engineers built a companion Android app (Flutter +Mapbox API) that sends
turn-by-turn data to the glasses via Bluetooth.
Gen 3 prototype — The final interface was informed through information density testing performed in-person and through prolific.
Results
Using the same testing methodology across prototypes, we measured a consistent reduction in distraction compared to phone-mounted navigation.
Phone Mount (Control)
11.5%
Eyes-off-road time
0.55s
Dwell time per glance
HoloNav AR HUD
6.3%
Eyes-off-road time
0.71s
Dwell time per glance
The AR HUD reduced eyes-off-road time by 45% compared to the phone mount. Though Dwell time was higher, we attributed this to unfamiliarity with AR interfaces, occasional software errors, and a fluctuating battery indicator that was removed in the final UI.
Qualitative feedback revealed riders felt a constant need to glance up to avoid missing instructions, indicating that auditory or haptic cues should be prioritized in future iterations.
Interface Design
Tools: Figma, FigJam, Surveys
Focus: Minimal cognitive load, information hierarchy, display customization
The UI evolved across three display generations. Early interfaces used simple directional arrows with distance countdowns. The final UI concept introduced tiered information density— letting users toggle between turn-by-turn only, turn + street name, and turn + street + arrival info.
Survey data showed simplicity and clarity scored highest, with turn-by-turn ranked as the most important feature (83%), followed by street name (75%). When forced to sacrifice one of four features, only 13% of users kept the map view.
HUD information density tiers (L→R: minimal, detailed).