1 Department of Electrical and Computer Engineering Advisor: Professor Hollot Team RCA March 1, 2013 Cumulative Design Review.

1Department of Electrical and Computer EngineeringDepartment of Electrical and Computer Engineering Advisor: Professor Hollot

Team RCAMarch 1, 2013

Cumulative Design Review

2Department of Electrical and Computer EngineeringDepartment of Electrical and Computer Engineering Advisor: Professor Hollot

Kenneth Van Tassell, EEUser Interface & Communication

Scott Rosa, CSEServer & Data Analysis

Justin Kober, EESensor Network & Power

Timothy Coyle, EEImpact Processing & Communication

RCA (Real-Time Concussion Analyzer)

3Department of Electrical and Computer Engineering

Current concussion detection• Train coaches to recognize symptoms

Players may hide or not experience symptoms right away

RCA will monitor each player and alert the coach with the risk of concussion after each collision

Customer feedback• College trainers and high school athletic director

Concussion Detection in High School Football

4Department of Electrical and Computer Engineering

Our Solution: Block Diagram

Bluetooth

Android

Server

5Department of Electrical and Computer Engineering

Demonstration of Impact Data Collection• Single accelerometer interfaced with processor• Helmet processor transmission

Demonstration of Base Station/UI Interaction• Using test data

− Receive from helmet− Run algorithm

• UI able to receive and display test results

Review of MDR

6Department of Electrical and Computer Engineering

Sensors

ADXL 78 MEMS Accelerometer• Single axis• +/- 70 g range• 27 mV/g sensitivity• 5 V and 1.3 mA

PCB Design• 3 connection wires• 1 noise filtering capacitor• 1 sensor

7Department of Electrical and Computer Engineering

Battery

System Power Requirements • Maximum current ≈ 300 mA• 5 volts• 5 hour duration• Total energy consumption = 1390 mAh

USB Battery Pack• 2000 mAh• 5.5 V, 700 mA • Rechargeable• Packaging in helmet

8Department of Electrical and Computer Engineering

Sensor Packaging in Helmet

Sensors

IC and battery

9Department of Electrical and Computer Engineering

Human Skull

Loaded vs empty helmet

Simulate the neck

Controlled test for impact duration

• Small increase

10Department of Electrical and Computer Engineering

Application / User Interface and UI Communication Design Requirements

• Simple Operation / User Friendly• Displays Meaningful Information• User Adaptable

− Coach vs. Trainer• Player Adaptable

− Adaptive Threshold Monitoring

Challenges• Improving Processing • Maintaining Continuity• Redesigning Bluetooth Parsing for Variable Hits

11Department of Electrical and Computer Engineering

Application / User Interface and UI Communication

12Department of Electrical and Computer Engineering

Application / User Interface and UI Communication

13Department of Electrical and Computer Engineering

Risk Function

C1= -12.531C2= 0.0020

14Department of Electrical and Computer Engineering

Calculating Magnitude Hit Vector

15Department of Electrical and Computer Engineering

Isolated Algorithm Error Characterization

16Department of Electrical and Computer Engineering

Isolated Algorithm Error Characterization

17Department of Electrical and Computer Engineering

Impact Processing & Communication

ATmega32U4 8-bit AVR Microcontroller• MDR sample rate = 2.08ms per sample• 0.152ms per sample (92.7% increase)• Threshold triggering

Wireless Transmission

 Maximum Data Rate (Kbps)

Effective Data Rate (Kbps)

Maximum Range (m)

Effective Range (m)

Bluetooth 240 154.2 100 30XBee 20 15.1 610 100Bluetooth vs Xbee

12xRate 10.2xRate 0.2xRange 0.3xRange

18Department of Electrical and Computer Engineering

Experiments• Threshold Estimation

- Mean, Standard Deviation

Impact Processing & Communication Experiments

19Department of Electrical and Computer Engineering

Demonstration of Complete System Functionality• Show implementation of battery powered system• Impact the helmet with a known force• Transmit impact data with required sample rate from the

sensor array to Android device• Display risk of concussion with confidence interval on

Android device• Display player impact history on Android device upon

user request

Proposed CDR Deliverables

20Department of Electrical and Computer Engineering

Demo

Demo

21Department of Electrical and Computer Engineering

Timeline

22Department of Electrical and Computer Engineering

Fully Integrated and Wearable Helmet• Stable battery• Player specific adaptability

Reliable User Friendly Android Application• User specific settings

Risk Calculation with Confidence Interval• Probability of Error

DEMO• Impact Dummy• Server, Impact Location & Application

FPR and Demo Day

23Department of Electrical and Computer Engineering

Questions

Thank You

24Department of Electrical and Computer Engineering

Questions

Thank You

25Department of Electrical and Computer Engineering

Data Processing and Storage

Requirements From MDR• Calculate a magnitude hit vector from the

accelerometer data• Calculate risk for each hit as well as cumulative risk• Be able to store raw accelerometer data

Accomplished• Calculates the hit vector in an average of 411.6 ms.

The overall program calculates risk in an average of 1.06 s

• Tested and graphed data• Set up server with database to store raw

accelerometer data and hit data

26Department of Electrical and Computer Engineering

Material Costs

This cost could be reduced by $1,200 purchasing the RN-41 and designing the BlueTooth modem ourselves

27Department of Electrical and Computer Engineering

Computational Analysis

Storage and computation can be done on a phone

Server hosts MySQL database

Max phone memory size: 64 GB

Min phone memory size: 2 GBTop Ten Paid apps in Itunes 11/26/12

Max App memory size: 791MBMin App memory size: 3.1MB

Average of max and min memory size: 397.05 MB

Size of Entry in table is

40Bytes

0.00004 MB9926250 hits in total

90

Avg hits per college season per player: 1177seasons played by college player: 4

hits for college team in 1 years 423720Team data for 4 years 17MB

Team data per year 4MBStore data for 99 teams per yearStore data for 23 college teams for 4 years

28Department of Electrical and Computer Engineering

Preliminary Weight Analysis

NFL Helmet Approximately 6lbs. or 2.722kg ATMega328P = 2g MEMS each approximately 1g Gyroscope approximately 2g Power approximately 23g Estimated total system weight (not including

packaging) = 35g-40g

5% of helmet is approximately 136g

29Department of Electrical and Computer Engineering

Requirements Analysis: Specifications

Real-Time continuous impact measurements Player specific adaptability Equipment weight increase less than 5% Effective range 150 m Responds in under two seconds Robust

• Interference• Durable

30Department of Electrical and Computer Engineering

Measured System Energy Consumption

31Department of Electrical and Computer Engineering

Worst Case Power

Worst Case Power AnalysisDevice Max Voltage (V) Max Current (mA) Max Power (W)

ATmega32U4 5 200 1

BlueSMiRF Gold 3.3 100 0.33

ADXL 193 5 2   0.01

Total Power (W)

1.39

32Department of Electrical and Computer Engineering

Lower frequency response• Vibrations

Mapping to graph

• More data points

Cost

Single-Axis VS. Multi-Axis

33Department of Electrical and Computer Engineering

Impact Data Collection Power circuit Sensor array Processing and transmission of sensor array

User Interface Cumulative linear acceleration User preferences Risk display Stability

Data Analysis Sub-concussive impacts integrated Statistical analysis for entire system

Proposed CDR Goals