1
Division of IT Convergence Engineering Related Work Knee Rehabilitation Using Range of Motion Exercise Feedback Yeongrak Choi 1 , Sangwook Bak 1 , Sungbae Cho 1 , Changsuk Yoon 2 , John Strassner 1 , M. Jamal Deen 1 and James Won-Ki Hong 1 1 Division of IT Convergence Engineering, POSTECH, Pohang, Korea 2 Department of Computer Science and Engineering, POSTECH, Pohang, Korea Motivatio n Our Design Conclusio n Overview Results Importance of Knee Rehabilitation – Difficult to return to its original state after injury or operation • Stable, enduring and customized rehabilitations required – Feedback on the health of knee required • Accuracy of monitored data is essential for customized knee exercise plan and to ensure the overall safety of the knee rehabilitation process • Beneficial to both patients and doctors Knee Joint ROM (Range of Motion) Exercise – Helpful for knee rehabilitation – Criteria for checking the health of knee Knee Rehabilitation Monitoring and Inference System – Monitor the knee ROM exercise Maximum/minimum angle, period per ROM activity, moving count, # of sets, … – Analyze exercise data How much exercise per day? – Infer the health of the knee and recommend changes if necessary Determine if the health of the knee is improving based on measurement data Is more exercise needed, or is current exercise sufficient? Our work – Better accuracy - Uses 3-axis accelerometer and gyroscope – Popular technology - Uses Bluetooth to communicate – Light-weight, less than 400g Activity sensor using WBAN – Two-axis accelerometer - less accurate – ZigBee used for communication - less popular AKROD (Active Knee Rehabilitation Orthotic Devices) • Large size and Heavy (3.18kg); no network functionality Sensors - use two Wiimotes – 3-axis MEMS accelerometer (ADXL330) • Measuring magnitude and direction – 2-axis MEMS gyroscope (IDG-600) in MotionPlus • Gyroscope for tracking movement Inference using Ontologies – Inferring rules • Ability: Evaluating maximum and minimum angles • Intensity: Checking the number of sets Design Objectives: Portable, User-friendly and Smart! Sensors installed into knee support Implemented server-based user- interface Sensor Data Monitoring Results / Inference Patient Doctor Sensor data Symptoms Ontology for Knee Rehab. x y z g x y z Sensor Experiment – Scenario: Regularly bends and unbends leg for 10 times (30-40˚ to 120˚) – Evaluation: Use of Kalman filter minimizes errors from rapid movement Conclusion – Our system provides better knee rehabilitation – accurate, light weight and cheap – Filtering technique to calibrate the data from different sensors Future Work – Enhance the accuracy of measuring knee angles – Develop ontologies with rules to augment knowledge – Improve user interface - smart phone application and better server interface – Apply to other joints and new situations Daily Result Exercise Guidance (from Dr.) Installation & Implementation Examples of Inferring Rules Ability if (Max-Min > Guided Angle) Good Intensity if (Daily Set > Guided Set) Enough Bluetooth (PAN) LAN / WLAN (Socket Programming) Knee sensors Measuring data Local Server Infer results Analyzes data Receives data Receiver (PDA / Smartphone) Communication with user Display data & info. g x y z x y z

Division of IT Convergence Engineering Related Work Knee Rehabilitation Using Range of Motion Exercise Feedback Yeongrak Choi 1, Sangwook Bak 1, Sungbae

Embed Size (px)

Citation preview

Page 1: Division of IT Convergence Engineering Related Work Knee Rehabilitation Using Range of Motion Exercise Feedback Yeongrak Choi 1, Sangwook Bak 1, Sungbae

Division of IT Convergence Engineering

Related Work

Knee Rehabilitation Using Range of Motion Exercise Feedback

Yeongrak Choi1, Sangwook Bak1, Sungbae Cho1, Changsuk Yoon2, John Strassner1, M. Jamal Deen1 and James Won-Ki Hong1

1 Division of IT Convergence Engineering, POSTECH, Pohang, Korea2 Department of Computer Science and Engineering, POSTECH, Pohang, Korea

Motivation Our Design

Conclusion

Overview

Results

• Importance of Knee Rehabilitation– Difficult to return to its original state after injury or operation

• Stable, enduring and customized rehabilitations required

– Feedback on the health of knee required• Accuracy of monitored data is essential for customized knee exercise plan

and to ensure the overall safety of the knee rehabilitation process• Beneficial to both patients and doctors

• Knee Joint ROM (Range of Motion) Exercise– Helpful for knee rehabilitation– Criteria for checking the health of knee

• Knee Rehabilitation Monitoring and Inference System – Monitor the knee ROM exercise

• Maximum/minimum angle, period per ROM activity, moving count, # of sets, …

– Analyze exercise data• How much exercise per day?

– Infer the health of the knee and recommend changes if necessary• Determine if the health of the knee is improving based on measurement data• Is more exercise needed, or is current exercise sufficient?

• Our work– Better accuracy - Uses 3-axis accelerometer and gyroscope– Popular technology - Uses Bluetooth to communicate– Light-weight, less than 400g

• Activity sensor using WBAN– Two-axis accelerometer - less accurate– ZigBee used for communication - less popular

• AKROD (Active Knee Rehabilitation Orthotic Devices)• Large size and Heavy (3.18kg); no network functionality

• Sensors - use two Wiimotes– 3-axis MEMS accelerometer (ADXL330)

• Measuring magnitude and direction

– 2-axis MEMS gyroscope (IDG-600) in MotionPlus• Gyroscope for tracking movement

• Inference using Ontologies– Inferring rules

• Ability: Evaluating maximum and minimum angles• Intensity: Checking the number of sets

• Design Objectives: Portable, User-friendly and Smart!

Sensors installed into knee support

Implemented server-based user-interface

Sensor DataMonitoring Results / Inference

Patient Doctor

Sensor data Symptoms

Ontology forKnee Rehab.

x

yz

gxy

z

• Sensor Experiment– Scenario: Regularly bends and unbends leg for 10 times (30-40˚ to 120˚)– Evaluation: Use of Kalman filter minimizes errors from rapid movement

• Conclusion– Our system provides better knee rehabilitation – accurate, light weight and cheap– Filtering technique to calibrate the data from different sensors

• Future Work– Enhance the accuracy of measuring knee angles– Develop ontologies with rules to augment knowledge– Improve user interface - smart phone application and better server interface– Apply to other joints and new situations

Daily Result

Exercise Guid-ance(from Dr.)

• Installation & Implementation

Examples of Inferring Rules

Abilityif (Max-Min > Guided An-

gle) Good

Intensityif (Daily Set > Guided

Set) Enough

Bluetooth(PAN)

LAN / WLAN(Socket Programming)

Knee sensors

Measuring data

Local Server

Infer results

Analyzes data

Receives data

Receiver(PDA / Smartphone)

Communicationwith user

Display data & info.

g

xyz

xyz