Moshin Brace | My Site 2

MOSHIN KNEE BRACE

HUMAN MOVEMENT AND ARTIFICIAL INTELLIGENCE

This knee brace integrates an accelerometer-based sensing system to measure real-time knee kinematics during rehabilitation. Motion data is processed through a machine learning model to identify potentially harmful movement patterns that could hinder ACL recovery. By combining wearable technology with intuitive mechanical design, the brace supports safer, more informed rehabilitation.

TYPE

California College of the Arts

Undergraduate Senior Thesis

DURATION

24 Weeks (2023 - 2024)

PROCESS

01 BACKGROUND

In the United States, about 250,000 anterior cruciate ligament (ACL) injuries occur annually, a high-impact injury that leads to long rehab periods.

The knee is not a simple hinge. It’s a coupled system where small abnormal motions, especially in rotation and valgus, are strongly linked to ACL injury and failed rehab.

Problem Context

Approximately 250,000 ACL injuries occur per year in the U.S. alone, requiring extensive rehab. Yet only ~65% return to pre-injury activity levels, and nearly 30% experience further injury—suggesting that current rehab tracking is insufficient.

Need for Objective Tracking

Clinical research shows wearable sensors, including accelerometers, have potential to capture meaningful biomechanics data, but larger studies and better tools are needed to integrate this into real-world rehab routines.

Market Opportunity

The rehabilitative knee brace market is expanding and smart assist braces are growing—indicating users and clinicians are receptive to innovations that combine support with data tracking.

EXPLORATION

02 UNDERSTANDING THE MOVEMENT

BIOMECHANICS

DESIGN INTEGRATION

03 BIOMECHANICS DRIVEN FORM

Sketch exploration and fabric prototyping were guided by biomechanical research showing that ACL injury risk is driven by excessive valgus and rotational motion rather than flexion alone. Form, strap layout, and material behavior were iterated to support sensor placement capable of detecting lateral acceleration and asymmetric loading during dynamic movement.

EARLY GRAFT

EXTERNAL / INTERNAL ROTATION

Graft is still stronger than the native ACL

PROLIFERATION

LIGMENTISATION

Graft is in the most vulnerable state where it enters a remodeling stage

Material honesty drives the design of this Brutalist-inspired .

ABDUCTION / ADDUCTION

EXTENSION / FLEXION

Abduction is when the knee collapses inward and adduction is when the knee moves outward. This is the motion most associated with ACL injury.

Range: 5 - 7° degrees of knee movement.

Abduction is when the knee collapses inward and adduction is when the knee moves outward

Range: 0° at full extension | 130–150° at full flexion

During rehab: Walking: ~0–60° | Stairs: ~0–90°

Running: ~40–120°

ACL risk increases when:

Internal rotation occurs near extension
Combined with valgus collapse

04 COLOR, MATERIAL, FINISH

SOFT FABRICATION

An existing knee brace cutting pattern was modified to preserve proven anatomical fit while accelerating material testing. Neoprene’s elasticity introduced edge-binding challenges, particularly under dynamic tension. Iterative testing of binding methods and fabric adhesives was conducted to identify assembly solutions that maintained stretch, seam durability, and wearer comfort.

Neoprene

Velcro

BOA Lacing / Sensor housing

Internal rotation is when the tibia rotates inward relative to femur and external rotation is when the tibia rotates outward. This motion is small, fast, and dangerous.

At 90° flexion: 20° - 40° degrees of knee movement.

Near extension: Rotation drops to 5°-10°

MACHINE LEARNING

06 BACKGROUND

The micro controller unit initially gave haptic feedback through sound when the acceleration of the knee was too aggressive. The parameters were

Accelerometers can detect:

Lateral acceleration spikes
Asymmetric loading events
Sudden valgus collapse patterns

Sensor relevance:

Accelerometers + ML can detect:
Rotational acceleration
Coupled valgus + rotation events
“Near-extension rotation spikes”

Hesitation
Stiff-knee gait
Asymmetrical cadence

In context scenario

FINAL

07 DIGITAL PRODUCT

DIGITAL PRODUCT

A mobile app concept translates knee motion data into progress tracking, movement alerts, and weekly rehabilitation summaries. The interface explores how real-time feedback and long-term trend visualization could support safer movement patterns and patient engagement during recovery.

Daily Movement Summary – steps, sessions, safe vs risky motions
Real-Time Alerts – valgus or rotational warning indicators
Weekly Progress – trend lines, symmetry improvements
Clinician View – aggregated recovery metrics

UI/UX concept exploring how motion data could be transformed into clear feedback, progress tracking, and injury-risk awareness.