P19224: Biometric Driver Monitoring

Customer Handoff & Final Project Documentation

Table of Contents

Team Vision for Final Demo and Handoff

During this phase, any last minute prototype issues were ironed out before Imagine RIT and the customer handoff. The prototype was then tested to ensure that it was able to meet the Engineering Requirements set forth by our customer. The finalized prototype was demoed at Imagine RIT 2019 using the FSAE driving simulator. The prototype and all associated documentation was then given to the customer.

Test Results Summary

Our project was very successful in meeting the Engineering Requirements set forth by our customer. The only requirement that we could not test, was D2 - withstands G-Forces experienced while driving. This requirement could not be tested for due to a lack of Formula Car availability. The lack of Formula Car availability, also required that all testing be completed in the Formula Car simulator. The lack of a real car for testing could effect some of our test results, as more signal noise will be present when driving a moving car, rather than a simulator.
Performance vs. Engineering Requirements

Performance vs. Engineering Requirements

The full document can be viewed here: Test Plans and Results

Risk and Problem Tracking

Our final Risk Management document is shown below.
  1. Green - Risk was successfully mitigated
  2. Yellow - Risk was mitigated with certain parameters
  3. Red - Risk was not mitigated
Risk Management

Risk Management

The full risk management document can be found here: Risk Management

Final Project Documentation

Technical Paper

Our technical paper can be viewed here: Biometric Driver Monitoring Technical Paper

Imagine RIT Poster

Our poster can be viewed here: 2019 Imagine RIT Poster

Bill of Materials (BOM)


Confirm that all expenses and contingencies are afforded by the project financial allocation
Bill of Materials

Bill of Materials

The full document and actual spending can be viewed here: Bill of Materials

The project spending remained within the allocated budget as shown below:

Actual Spending

Actual Spending

CAD Schematics

All CAD schematics can be found here:

User Manual

The user manual can be found here: User Manual

Recommendations for Future Work

In-Car Testing

Due to a lack of Formula SAE car availability, the team was unable to test the system in a driving car. All testing was completed in the Formula SAE car simulator, which cannot simulate the car's actual movement on the track.

Pulse Oximeter

The pulse oximeter earclip sensor used in our system was an "Compatible Nellcor Veterinary SpO2 Sensor Animal Ear Tongue Clip 9 Pins Connector FDA/CE Approved" from Amazon.com for $29.99. This earclip sensor was chosen for its low cost, as other earclip sensors cost $160.00-$250.00. However, this earclip sensor is of extremely low quality and the signals that this sensor picks up are extremely noisy and not useful for data analysis. Future project iterations should consider purchasing and integrating a higher quality pulse oximeter earclip sensor into the system.

EMG Frequency Domain Analysis

One of the most important components found for fatigue analysis was the increase in EMG signal frequency as muscles fatigue. Since the EMG values are sent 100 times per second (as max or RMS) from each sensor, it may be feasible to implement an FFT for samples gathered within each period. This would facilitate finding the median frequency in each period and report it as an indicator of fatigue, or incorporate it into a greater fatigue algorithm. Additionally, this can be used to ideally filter unnecessary frequencies. GPU_FFT is probably the best library suited for this, as it offloads the work to the GPU while the CPU is already being rather heavily utilized, especially with the plotter working.

Fatigue Algorithm

Design an algorithm to measure driver fatigue using EMG signals. Values from this can be used to improve driver ergonomics and better determine when to switch out drivers. Preliminary muscle fatigue analysis can be found here: Fatigue Analysis.

Bluetooth Capability

Make the system bluetooth compatible to eliminate the need for wires connecting the driver to the system. This will eliminate the need for a push-pull connector and make it easier and safer for the driver to exit the vehicle.

Other Biometric Signals

Other sensors could be added to the system to measure various biometric signals from the driver. These could include ECG sensors to measure brain waves or electrodes to measure skin conductance. New biometrics could be also considered when designing an algorithm to measure driver fatigue.

CAN Optocoupler

It may be best to send CAN signals via an optocoupler to isolate the driver from ground to prevent improper EMG signals. Alternatively, the car battery ground could be tied to the ground on the system's battery.


The baseplate cools the Raspberry Pi CPU well (around 42 C at load), but there is still some room for improvement as the heatsink that was built did not make full contact with the CPU. Thermal paste may also be a better alternative to the thermal tape that was used.

Plans for Wrap-up

The final prototype and all associated documentation will be handed-off to our customer during the Customer Handoff Presentation. The EDGE website will also be finalized.

Home | Planning & Execution | Imagine RIT

Problem Definition | Systems Design | Preliminary Detailed Design | Detailed Design

Build & Test Prep | Subsystem Build & Test | Integrated System Build & Test | Customer Handoff & Final Project Documentation