Studded Tire Detection

ENSC 45 - Studded Tire Detection

Will Cooper, Patrick Hinshaw, Patrick McCarthy, Brian Spink

Intro

• Project Goal: To detect and count the total number of cars and determine which have studs.

• Original scope• Automated stud detector prototype

• New scope• Developing a method for stud detection

Milestones

• MATLAB algorithm to detect cars• MATLAB program to detect studded tires

MATLAB Algorithm Flowchart

• Signal Extraction• Collect all relevant

signals from a set of data

• Signal Analysis• Perform 3 types of

analysis for stud determination

• Calculation of Results• Calculate the total

number of cars and studs

Extraction Method

Input a threshold

Extraction Method (Continued)

Data Analysis

Percent Analysis

PDF Analysis

PDF Analysis

Results

• Text file with relevant data• Time of signal• Values of each detection method• Total number of cars / studs detected

• MATLAB demo of tireDetect.m

Error Analysis

Caveats

• SNR / undesired signals• Data Collection Inconsistencies• Rudimentary signal extraction• Reliability of detection algorithms• Inconsistent errors

Improved Analysis Possibilities

• Higher quality equipment• Sensitive sample extraction• Weighted voting scheme• Additional test methods• More detailed error evaluation• PPV• NPV

Budget

• Polsen SCL-1075 microphone -$50• ATR-6550 microphone -$55• Bat detector microphone -$50• Set of studded tires -$100• MATLAB Licenses -$300 _________ Total : $555

Project Continuation

• All research items will be passed down• MATLAB programs and experimental data• Microphones• Research articles

• Suggestions• Collaboration with other methods• Video implementation

Conclusion

• Developed an algorithm to detect studded tires• Total car & stud counter• Formatted text file for output

• Sound is a viable means of detection

• State of studded tires in Washington State• Recent talks of a ban