Speculative Case Study Automotive / Autonomous HMI Coming Soon

Tesla FSD — drive with certainty.

Reimagining Tesla's autonomous drive UI with LiDAR-infused vision. Not as a physical crutch, but as a tool to rebuild driver trust and system transparency — turning a tentative, flickering estimate of the road into a rock-solid rendering of the world.

Subject
Tesla FSD
(speculative)
Role
Lead UX
Product Design
Focus
In-cabin HMI
& Trust
Status
In progress
2026
01 / Executive Summary

A vision-only bet, tested by the real world.

The Premise

Tesla's approach to autonomous driving has been famously dogmatic — relying on Tesla Vision, a pure camera system, while discarding radar and ultrasonic sensors. The decision was driven largely by manufacturing simplicity and cost, leaning on neural networks to perceive depth from 2D images.

But real-world edge cases — severe weather, blinding glare, sudden sensor occlusion — repeatedly expose the physical limits of a vision-only setup. This study explores a speculative but critical challenge: how would we design the UI and UX for a premium, LiDAR-integrated Tesla FSD option?

Rather than treating LiDAR as a crutch, we treat active laser range-finding as a way to rebuild driver trust — transitioning the in-cabin experience from a flickering estimation to a reassuringly accurate rendering of the vehicle's surroundings.

02 / The Challenge

Visual anxiety in pure vision systems.

The Friction

The primary friction point in today's FSD (Supervised) experience isn't a lack of capability — it's auditory and visual anxiety.

In a vision-only system, the in-cabin screen tries to show the driver what the car "sees." In poor weather, cars, pedestrians, and lane lines flicker, jump, or briefly vanish. Even when the pathing engine is performing safely, an erratic representation forces the driver to hover a foot over the brake, anticipating failure.

When the system is unsure of an obstacle's depth — a vehicle kicking up heavy spray — it escalates with high-frequency chimes and takeover strikes. That creates a reactive UX where the user must continuously judge whether the system is about to err or simply overreacting.

01
Ghost visuals.
Obstacles flicker and disappear on screen in rain, spray, or snow. The visual instability triggers immediate anxiety, regardless of whether the car is actually safe.
02
Cognitive overload.
Uncertain depth becomes high-pitched chimes and "blue screen" prompts, forcing constant evaluation of whether the system is failing or overreacting.
03
Abrupt handoff.
In heavy weather the system can disengage with an "Autopilot Degraded" alert, giving the driver mere split-seconds to reorient — the opposite of calm.
03 / Strategic Objective

Rebuilding trust with visual certainty.

The Goal

Design an interface that leverages the physical certainty of LiDAR — which measures distance directly via photons rather than estimating it through pixels — to redefine the autonomous driving UX.

The design must serve three human-centric pillars: visual stability (eradicating the jumping of estimated obstacles), confidence leveling (distinct visual styles for the system's certainty), and proactive environmental feedback (telling the driver why the car is confident, even in zero-visibility weather).

[ Traditional Vision Only ]  ->  Unstable depth estimation  ->  Visual flicker   ->  User anxiety
[ LiDAR-Infused Vision   ]  ->  Direct distance measure    ->  Solid rendering  ->  High trust
Model — from estimation to measurement
"

Trust isn't built by hiding uncertainty. It's built by showing the driver exactly how certain the car really is.

04 / Proposed Redesign

Three moves toward a calmer cabin.

The Concepts

Each feature translates a physical capability of LiDAR into a visible, reassuring signal — so the driver understands the system's confidence without decoding raw data.

01
LiDAR Scan-Lines
A subtle laser "halo" emanates from the vehicle vector. When LiDAR resolves depth in low visibility, detected obstacles gain a stable, neon-hued mesh instead of shaky vector blocks — signalling that the rangefinders have locked on even when cameras are partly blinded.
02
Active Confidence Path
The FSD line becomes a gradient. Solid deep blue means 100% path certainty; a softer, wider turquoise appears when navigating construction, merges, or zero-map areas where LiDAR is carving boundaries in real time. Peripheral vision reads it without a chime.
03
Environmental Attenuation
A calm status toast replaces the beep-and-disengage. Normal: "LiDAR + Vision Active." Degraded: "Cameras Occluded — Navigating via LiDAR Precision." The system reassures the driver of its redundancy layers instead of panicking them.
            [ FSD ENGAGED ]

     .-.     .-.     .-.        <- blinding spray (camera)
    ( o )   ( o )   ( o )       [ LiDAR locks solid 3D mesh ]
     '-'     '-'     '-'
              |
              v
        +-----------+
        |    CAR    |            <- zero flicker on screen
        +-----------+
              ^
              |   (active laser range lines)
         [ My Tesla ]
Concept — LiDAR stabilises the obstacle even as cameras degrade
05 / Architecture Comparison

The same car. A steadier interface.

Existing vs. Proposed

The redesign isn't about prettier pixels — it's about replacing binary, reactive signals with progressive, legible ones that keep the driver informed.

Component
Existing Vision-Only UX
Proposed LiDAR-Infused UX
Object Representation
Shaky, flickering 3D boxes that jump or disappear in bad weather.
Stable, high-precision models with clean edge-mesh highlights.
Path Visualization
Binary blue line (engaged vs. disengaged).
Dynamic gradient path showing real-time confidence and its source.
System Status
High-pitched warnings and sudden takeover strikes.
Calm, persistent notifications confirming sensor redundancy.
Error Handling
Abrupt FSD disengagement when cameras lose visibility.
Progressive fallback states — driver stays informed, autonomy stays safe.
06 / Recommendations

Rolling it out without confusion.

Next Steps

Integrating physical distance sensors into a historically vision-exclusive suite needs a strategic rollout so the change reassures rather than confuses.

01
Dynamic confidence states.
Process LiDAR into clean, simplified meshes rather than raw point clouds. Reveal scan lines only in low-visibility or congested areas to emphasise active sensing without overwhelming the driver.
02
Standardised sensor status.
A dedicated "laser pulse" icon beside the Autopilot wheel lets the driver instantly confirm that both optical and physical sensors are fully operational.
03
Refined alert thresholds.
With depth-sensing redundancy, reduce takeover chimes to true, high-probability collision risks — a calmer cabin, not a nervous one.
The Takeaway

Certainty is a design material. LiDAR's real product value here isn't more sensors — it's the ability to show the driver a stable, honest picture of how confident the car actually is, and to fail forward gracefully instead of handing back control in a panic.

This case study is still in progress. Full interactive mockups of the scan-line visualizer and the gradient confidence path are on the way.

Back To / Selected Work

More case studies.

See all work