smart pen technology

Feasibil ity Study of a Smart Pen for Autonomous Detecti on of

Concentrati on Lapses During ReadingGuided by:Gopika D KAsst.Professor

Presented by:Kenneth Thomas.N1MJ10EC108

Contents:AbstractIntroductionMethods & Materials

• Motion detection• Alarm system• Testing• Evaluation method

Results• Power Management• Form Factor• Accuracy

AdvantagesConclusionReferences

Abstract: Accelerometer and a microcontroller are embedded in a pen to

record data as user reads the text using the pen as a pointer. An alarm is generated when there is a substantial pause in

reading. An software program communicates with the pen through usb

interface to customize system parameters,record data & graph collected data.

The prototype was tested on 11 normal volunteers and 3 persons with ADD.The system had a false alarm rate of 19% and sensitivity of 82%.

With refinement, our system can be used to help students with ADD/ADHD maintain concentration while reading in different scenarios.

Introducti on In the US alone, there are 4.5 million school aged children suffering from ADD/ADHD. A technological solution is required enabling readers to maintain concentration and

makes the acquisition of behavioral data and delivery of interventions more efficient for educational professionals.

Currently available systems are:• Point system• Timed alarm system

It is desirable to automate the role of supervisor in a portable and discreet way such that the user feels comfortable using it in a variety of environments.

The system is based on the assumption that when user’s concentration lapses,the pattern of pen movement is different from normal reading.

Methods & Materials3-axis accelerometer(MMA7260,Freescale)Microcontroller(Atmega 328,Atmel)Alarm system(vibrating element and LED)Supplemental Java program

3axis accelerometer Microcontroller(Atmega 328,Atmel)

Contd..

A. Moti on Detecti on A signal processing algorithm detects features in accelerometer signals to distinguish

between reading motion and lack of motion. There are 4 categories of signals.They are:

• Reading motion• Vigorous motion• Negligible motion• Incorrect orientation

B. Alarm System Alarm system was designed to be discreet with the goal of increasing user independance. In accordance with this goal,the system does not contain any sound emitting device,rather

a vibration is generated when necessary. The warning consists of a surface mount LED that an be disabled by prompt action from

the user.

C. Tests Using Human Parti cipants

Testing was performed in 2 phases 8 volunteers (aged 20 to 25) were recruited in the 1st phase,3 of whom were diagnosed

with ADD/ADHD. Volunteers were made to read a sample text for 10 minutes with hand motion being

video taped and system outputs being recorded. 6 volunteers were recruited for the 2nd phase (none of whom had ADD/ADHD). They were made to read continuously for a period of 1hr, with hand motion being

video taped and system outputs being recorded. To ensure that the pen was exposed to non-reading scenarios,the volunteers were

periodically engaged in light conversation,forcing them to temporarily ignore the material.

D. Evaluati on Method The practicality and performance of our system was evaluated in terms of power

management, size and accuracy. Accuracy of system outputs was evaluated by comparing user behavior with video tape

recordings. An event is the occurrence of a behavior that our system detected. The different events are:

• Not reading: 32 seconds• Incorrect orientation: 5 seconds• Vigorous motion: 3 seconds• Negligible motion: 6 seconds Initial settings can be customized for different users using the supplemental java

program. Recording parameters:

• True positive(TP)• False negative(FN)• False positive(FP)• True negative(TN) Sensitivity (TP/(TP+FN))

Specificity(1-FP/(FP+TN))

ResultsA. Power Management

Power consumption• 800 A in reading detection mode• 4 mA in warning mode• 31 mA in alarm mode• < 10 A in sleep mode

Therefore the estimated current consumption/reading session is about 2 mA. The smart pen runs on a AAA battery. The system can thus run continuously for about 550 hours.

B. Form Factor The system is portable,freedom to be used in public. The system was small enough to fit into a highlighter casing with the filly populated

printed board measuring 15mm x 97mm. Warning is generated using blinking LED. Alarm is generated using vibration,which notifies the user discreetly.

C. Accuracy

1st phase of testing was designed to measure specificity.

Reading task was only 10 mins long & system was not tested for different reading scenarios.

Subjects 2 & 8 showed high count of false positives due to a typical reading styles.

Very few true positives as reading time was limited to 10 mins.

Due to short period of reading,there are no false negatives.

Contd..

2nd phase of testing was designed to evaluate both sensitivity and specificity during a reading time of 1hr,with some deliberate distractions.

The algorithm performed well when predictable movements were executed by the user with pen orientation of 45 degree during reading.(eg.subject 1).

During distractions,our system correctly detected true positive events with a sensitivity of 81.9%.

Advantages Small footprint and can fit into a highlighter casing. Low power consumption. Alarm system is discreet. The algorithm can be customized for different users and different reading styles

using the supplemental pc program. The role of the supervisor has been made automated.

Conclusion Implemented a prototype instrumented “smart pen”. It has a small footprint,operates at low power and has low probability of false

alarm Future work will be on customizing the pen to adapt to individual reading habits &

detection of concentration lapses for persons with ADD.

References[1] W. R. Jenson, H. K. Reavis, and G. Rhode. The Tough Kid Book:Practical Classroom Management Strategies. Longmont, CO: SoprisWest, 1993.[2] A. Morrison, “On-task learning system and method,” U.S. Patent0188860, 2006.[3] L. Herron, “Portable electronic device for assisting persons withlearning disabilities and attention deficit disorders,” U.S. Patent6042383, 2000.[4] “The Invisible Clock,” addwarehouse.com. [Online]. Available:http://addwarehouse.com/shopsite_sc/store/html/invisible-clock.html.[5] J. Green, “Learning device for attention deficit children,” U.S. Patent5288233, 1994.[6] S. Kalson, “Adaptive Array CFAR Detection,” IEEE Trans. Aero. Elec.Sys., vol. 31, No. 2, Apr. 1995, pp. 535-542.

Thank you