AI-edited Genesis G90 active-seat concept artwork

Active Seat Adjustment
for Personalized Driving Comfort

Genesis G90 Smart Seat Evaluation

Institution
HUVA — Humans and Vehicle Automation Lab
Kookmin University
Supported by
Hyundai Motor Company

Seat concept visualization

Acknowledgments

This industry–university project brought together Hyundai Motor Company, Hyundai Transys, and Kookmin University to study personalized occupant comfort through active seat adjustment.

Research on an Active Seat Adjustment System
for Personalized Occupant Comfort

AI-edited illustration of a black Genesis G90 driving along a city street
AI-edited vehicle illustration

Project objective

Evaluate an active seat adjustment system, develop comfort evaluation methods, and analyze seat pressure alongside perceived comfort to guide personalized seat support.

Study objective

Evaluate driver comfort and its relationship with seat pressure through stationary assessment and prolonged real-world driving.

Completion report · pp. 4, 42 · Driving study · pp. 2–4

Overview

AI-generated illustration of a driver adjusting his sitting posture in a parked Genesis G90
Prolonged driving can lead to discomfort that varies across individuals.

Background

Personalized seat support should reflect individual pressure patterns and comfort preferences.

  • Drivers differ in body shape, habitual posture, and preferred seating positions, which can lead to different comfort needs.
  • During prolonged driving, seat pressure and perceived comfort need to be evaluated together to understand changes in the seated experience.
  • Considering both pressure distribution and subjective comfort can help guide seat adjustments that better reflect individual preferences.
Objectives

Evaluate active seat adjustment for personalized occupant comfort through stationary and real-world driving experiments.

Seat pressure: objective measurement · Perceived comfort: participant-reported rating.
Completion report · pp. 4, 18–19, 23 · AI-generated concept illustration.

Research

How can we evaluate seat comfort from the driver’s perspective?

Our research question

Does a balanced seat always feel comfortable?

AI-generated illustration of seated posture and support at the back, pelvis and thighs
Pressure balance is only part of the seated experience.

Evaluate what the seat measures together with what the driver feels.

Literature Review

Seat pressure should be interpreted alongside subjective comfort.

  • In a review, pressure distribution showed the clearest links with subjective comfort ratings.
  • Applied here: pair seat-pressure data with reported comfort.

Evidence: de Looze et al. (2003) · Figure: Jin et al. (2009)

Pressure-distribution maps reproduced from Jin et al. (2009), Figure 3

Seat evaluations should capture both comfort and discomfort.

  • Earlier research compared separate comfort and discomfort ratings.
  • Applied here: body-region ratings on an adapted −5 to +5 scale.

Study & figure: Kyung et al. (2008), Part I

Body map and separate comfort and discomfort scales reproduced from Kyung et al. (2008), Part I

Seat comfort needs to be observed over time.

  • Discomfort increased across a 100-minute simulated drive.
  • Applied here: repeated comfort ratings during two-hour drives.

Study & figure: Jin et al. (2009)

Time-course ratings graph reproduced from Jin et al. (2009), Figure 4

Figures reproduced in the project literature review · Left: AI-generated concept

How the Active Seat Works

Pressure sensing and air-cell adjustment in the study

14 sensing regions

LeftRegionRight
Upper back*
Mid-back
Lower back
Flanks
Buttocks
Thighs
Backs of knees
  • Road adjustment
  • Conditional
  • Sensing only

10 regions adjusted;
up to 12 with upper-back control.

*Upper-back control depends on the backrest angle. Pressure is measured behind the knees, but this region is not adjusted to protect pedal operation.

Road adjustment order

Buttocks → Thighs → Mid-back → Lower back → Flanks → Upper back*

Repeating cycle

  1. 110 cycles
  2. 2Hold 1 min
  3. 3Upper-body deflation
  4. 420 cycles
  5. 5Hold 1 min
  6. 6Full deflation

Repeat from step 1.

Prescribed
Reference driving posture.
Comfort-adjusted
Prescribed posture adjusted for individual comfort.
Air-cell balanced
Comfort-adjusted posture with left/right pressure balancing.

Completion report · pp. 23, 56, 60

Experimental Design

From stationary assessment to on-road evaluation

Experimental Objective

Evaluate perceived comfort and pressure distribution with and without active air-cell adjustment.

Study sequence

  1. Stationary assessment
  2. Participant selection
  3. Two-day road evaluation
  4. Outcome measures
32

Stationary participants

4 sequential postures

  1. Habitual
  2. Prescribed
  3. Comfort-adjusted
  4. Air-cell balanced
12

Selected drivers

Selected for pronounced stationary comfort responses

Same 12 drivers · Both days · Fixed order

Day 1

No air-cell adjustment

Adjustment OFF throughout the 120-minute drive.

OFF
Day 2

Adjustment after 60 min

OFF for 0–60 min; active adjustment for 60–120 min.*

OFFON*

Pressure
+
Comfort

Regional pressure
−5 to +5 ratings

14 pressure-sensing regions

Genesis G90 · Approx. 122-km route · Scheduled ratings: 0, 30, 60, 90 and 120 min

*Day 2: adjustment paused for 3 minutes at minute 100. Fixed day order; not randomized.

Completion report · pp. 23–26, 42, 52, 55–59

Experiment

Stationary Vehicle Experiment

IRB protocol: KMU-202503-HR-461

The actual Genesis G90 stationary study setup

Experiment

Equipment
Seat-pressure sensors
+ air-cell control system
Vehicle
Genesis G90 limousine
Pilot sessions
3
Main experiment
32 participants
Session duration
Approx. 75 min per participant

Experimental Procedure

  1. 1. Briefing &
    body measurements
  2. 2. Seat familiarization
  3. 3. Posture sequence

    Habitual → Prescribed → Comfort-adjusted → Air-cell balanced

  4. 4. Final comparison & interview
Independent Variable
Seating posture (4 sequential postures)
Dependent Variables
  • Regional pressure
  • Comfort ratings and comments

Summary and Limitations

  1. Pilot sessions standardized the posture, measurement and questionnaire procedures.
  2. Across the full stationary sample, neither pressure level nor left/right pressure ratio showed a significant correlation with comfort.
  3. Static seating findings require evaluation during prolonged real-world driving.

Completion report · pp. 23–26, 47–48

Experiment

Real-World Driving Experiment

IRB protocol: KMU-202503-HR-461

Actual pressure monitoring during the Genesis G90 driving study

Experiment

Equipment
Seat-pressure sensors
+ air-cell control system
Vehicle
Genesis G90 limousine
Route validation
4 pilot drives
Main experiment
12 drivers · 24 valid sessions
Driving sessions
Two 120-min drives
Approx. 122-km route

Experimental Procedure

Same 12 drivers · Both days · Fixed order

  1. 1. Briefing & seat setup
  2. 2. Two-day drive

    Day 1 OFF 0–120 min

    Day 2 OFF 0–60 / ON 60–120 min*

    During each drive: scheduled ratings at 0, 30, 60, 90 and 120 min.

  3. 3. Final interview
Independent Variable
Air-cell adjustment condition
Dependent Variables
  • Regional pressure
  • −5 to +5 comfort ratings and comments

*Day 2: 3-min adjustment pause at minute 100.

Summary and Limitations

  1. The protocol captured pressure and repeated comfort ratings across 24 valid driving sessions.
  2. The small cohort and fixed day order limit causal interpretation and generalization.
  3. Further trials should add a comparison seat-operation condition and a larger sample.

Completion report · pp. 52, 55–59, 88–90

Comfort Questionnaire

Participants rated whole-body and regional comfort on an 11-point scale (−5 to +5). Researchers recorded their ratings and comments.

English adaptation of the study body map and 11-point comfort rating guide, from −5 to +5
Comfort rating guide · English adaptation · Click to enlarge
English translation of the blank study form for recording whole-body and regional comfort ratings
Draft recording form · English translation · Click to enlarge

Source versions: the guide uses “Neither comfortable nor uncomfortable” at 0; the draft recording form uses “No discomfort”.

English translations of the original diagrams · Completion report, slide 21

Evaluation Framework

Physical measurements and the driver’s experience

Actual in-vehicle study photograph showing the driver, road context and pressure-monitoring laptop
Actual Genesis G90 study setup

Participant (driver)

Drives the vehicle
and rates seated comfort

Survey researcher

Administers the comfort survey
and records responses

Pressure recording

The laptop logs pressure
from 14 sensing regions

Pair measured seat data with the driver’s reported experience.

Completion report · pp. 56–60

Stationary Findings

Comfort across three evaluated seating postures

Mean whole-body comfort in 32 participants: prescribed posture −1.47, comfort-adjusted posture +2.09, air-cell balanced posture +1.97.
Whole-body comfort · n = 32 · Rating scale: −5 to +5Open full-size chart ↗
Personal comfort preferences matter when adjusting the seat.

Prescribed → comfort-adjusted

Mean comfort increased from −1.47 to +2.09 after the prescribed posture was adjusted to the individual.

Comfort-adjusted → air-cell balanced

The means were similar: +2.09 and +1.97. Their whole-body comfort difference was not significant.

Evaluation focus

Read the pressure measurements alongside the posture the driver finds comfortable.

Completion report · pp. 19, 36, 86

Individual Responses

Different comfort responses to the same adjustment

Stationary comfort-response groups: increasing 14 participants, mixed 12, decreasing 6; groups were defined by regional comfort changes from comfort-adjusted to air-cell balanced posture.
Stationary response groups · Total n = 32Open full-size chart ↗
The overall mean includes different individual responses.

Increasing · n = 14

Regional comfort scores increased with air-cell balancing.

Mixed · n = 12

Some regions improved while others declined. “Mixed” describes variation across regions.

Decreasing · n = 6

Regional comfort scores declined with air-cell balancing.

Groups were defined from stationary comfort responses, rather than assigned before the experiment.

Completion report · pp. 42–43

Pressure and Comfort

Does a balanced seat also feel comfortable?

Stationary analysis of pressure level and left/right pressure ratio against comfort; no significant correlation was found across the full stationary sample.
Full stationary sample · Pressure level and left/right pressure ratioOpen full-size chart ↗
No significant pressure–comfort correlation was detected.

Objective measurements

Pressure level and left/right pressure ratio describe physical loading on the seat.

Full-sample result

Neither measure showed a significant correlation with comfort across the full stationary sample.

Implication for evaluation

Keep subjective comfort as a separate outcome when assessing seat adjustment.

This result is specific to the measured conditions and sample.

Completion report · pp. 47–50

On-Road Findings

Comfort changes during prolonged driving

  1. Stationary
  2. Driving
  3. Design proposals
Whole-body comfort means over two-hour drives. Day 1 without adjustment declined from 3.17 to 1.25. Day 2 rose from 1.67 at minute 60 to 2.17 at minutes 90 and 120.
Raw cohort means · n = 12 · Full rating scale: −5 to +5Open full-size chart ↗

Day 2 adjustment schedule

0–60 min · OFF60–120 min · ON*

*Three-minute pause at minute 100.

Day 1 — adjustment OFF

Mean whole-body comfort declined from +3.17 to +1.25 during the two-hour drive.

Mean ratings remained positive but declined over time.

Day 2 — adjustment after 60 min

Mean comfort rose from +1.67 at 60 minutes to +2.17 at 90 and 120 minutes.

The observed trajectory supports further testing. The fixed day order and small cohort limit causal interpretation.

Completion report · pp. 56, 65, 75, 88, 90

On-Road Pressure Response

Measured pressure during the two-day drive

  1. Stationary
  2. Driving
  3. Design proposals
Mean whole-body and upper-body sensor pressure during the two-day drive. On Day 2, whole-body pressure rose from 15,930 at 60 minutes to 17,201 at 120 minutes; upper-body pressure rose from 5,509 to 6,736 ADC units.
Mean sensor aggregates · n = 12 · ADC unitsOpen full-size chart ↗

Whole body

On Day 2, mean pressure rose from 15,930 at 60 min to 17,201 at 120 min.

Upper body

The corresponding upper-body mean rose from 5,509 to 6,736.

These descriptive sensor values do not establish a change in comfort.

Completion report · pp. 60, 68–69 · Driving paper · p. 6

Driving Outcomes by Stationary Response

Stationary responses did not determine a single driving trajectory

  1. Stationary
  2. Driving
  3. Design proposals
Day 2 mean comfort at minutes 60, 90 and 120: stationary increasing group, n=4, 2.75, 2.75, 2.75; mixed group, n=5, 1.20, 1.80, 1.60; decreasing group, n=3, 1.00, 2.00, 2.33.
Driving cohort · Stationary increasing n = 4 / mixed n = 5 / decreasing n = 3Open full-size chart ↗

Increasing group · n = 4

2.75 → 2.75 → 2.75

Day 2 mean comfort remained stable from 60 to 120 minutes.

Mixed group · n = 5

1.20 → 1.80 → 1.60

The mean rose at 90 minutes, then declined slightly at 120 minutes.

Decreasing group · n = 3

1.00 → 2.00 → 2.33

The driving mean rose despite the group’s decreasing stationary response.

Values are group means at 60 → 90 → 120 min on Day 2. They do not establish that every individual improved.

Completion report · pp. 55, 78 · Driving paper, p. 9 · Aggregated study data

Exploratory Associations

Selected relationships for further investigation

  1. Stationary
  2. Driving
  3. Design proposals

Pressure sampled every second; indicators summarized in 30-minute windows.

Two selected exploratory correlations: lower-body comfort and balance-maintenance duration at 120 minutes, r=.605, p=.037; whole-body comfort and a movement-related indicator at 90 minutes, r=.651, p=.022.
Selected exploratory associations · Unadjusted p-valuesOpen full-size chart ↗

Lower body · 120 minutes

Comfort was positively associated with time within a 4% left–right pressure-ratio difference.

r = .605p = .037

Whole body · 90 minutes

Comfort was positively associated with inflation count ÷ (recognizable + repositioning pressure-change event counts).

r = .651p = .022

Air-cell intake volume

After activation, air-cell intake volume showed no significant linear correlation with comfort in any assessed region (all p > .05).

Most tests were not significant. These selected, unadjusted associations are exploratory and do not establish causation.

Completion report · pp. 60, 80, 82 · Driving paper, pp. 9–10

Seat Strategies by Cluster

Proposals documented in the completion report

  1. Stationary
  2. Driving
  3. Design proposals

Cluster 1 · Low pressure

Smaller body size and lower BMI. Lower pressure and comfort scores.

  • Softer cushioning
  • Fine adjustment of seat height, lumbar support and backrest angle
  • Additional support behind the knees and under the thighs

Cluster 2 · Intermediate

Average body size. Intermediate pressure and comfort scores.

  • Maintain a stable, general-purpose seat design
  • Support whole-body, upper-body and lower-body balance
  • Provide lumbar, seat-angle and cushion-tension adjustment

Cluster 3 · High pressure

Larger body size and higher BMI. Higher pressure and comfort scores.

  • More durable and resilient cushioning
  • Stronger back and thigh support
  • Improved ventilation and heat management

Proposed strategies from the report; separate from the increasing / mixed / decreasing comfort-response groups.

Completion report · p. 87

Research Contributions and Future Work

Conclusions and follow-up studies stated in the source materials

  1. Stationary
  2. Driving
  3. Design proposals

Research contributions

Comfort evaluation process

Established and implemented comfort evaluation procedures for stationary and on-road conditions.

Completion report · p. 90

Experimental database

Collected air-cell balancing data and presented initial analyses for follow-up research.

Completion report · p. 90

Future work

Additional analyses

Analyze movement counts, fixed-position duration and posture; examine the effects of body support and movement induced by air-cell operation.

Completion report · p. 90 · Driving paper · p. 11

Comparison conditions

Add seat-operation conditions to address placebo effects and compare body balancing with movements such as massage.

Completion report · p. 90 · Driving paper · p. 11

Larger sample

Increase the sample size to strengthen statistical power.

Completion report · p. 90

Completion report · p. 90 · Driving paper · p. 11, §5.4

AI-edited rear-view Genesis-branded G90 concept on a winding road at dusk

Thank you.

Active Seat Adjustment for Personalized Driving Comfort

Yeongje Kim
Institution
HUVA — Humans and Vehicle Automation Lab
Kookmin University
Supported by
Hyundai Motor Company

In collaboration with Hyundai Transys

AI-edited concept visualization