The PART (Police and Ambulances Regulating Traffic) Program

By Viney Kumar

  • Share
  • Share URL
  • menu
  • <
  • >
  • 1Short Project Description 
  • 2Summary 
  • 3About MeAbout Our Team 
  • 4Question / Proposal 
  • 5Research 
  • 6Method / Testing and Redesign 
  • 7Results 
  • 8Conclusion / Report 
  • 9Bibliography, References and Acknowledgements 

The PART Program

The effectiveness of Emergency Response Vehicles (ERV) affects lives and properties that are at risk each day. Effectiveness is determined by the reaction time available for target vehicles blocking the ERV's path and the amount of traffic. The signalling method is the main factor impacting reaction time, though it is compounded by factors like background noise and sound-proofing in vehicles.   

The PART program provides a near fool-proof method of reaching a timely alert to a target vehicle enabling corrective action  and improving the success rate of the ERV in its mission.

The Part Program Summary

Recently, in India, I saw an ambulance, stuck in traffic, unable to move, leaving me wondering how many lives were being lost every day. Research shows sirens to be audible only within 100m or less and  only 26% can tell the direction of an ambulance without visual cues. This statistic is probably much worse in a noisy, traffic-choked place like India. An Emergency Response Vehicle (ERV) stands little chance of obtaining a clear path against such odds. This became the inspiration for my project.

Mobile phone technologies, especially smart phones have transformed productivity and communications. But what if we could use this creatively to help others? With GPS technologies, supported by the smartphone platform, picking up information on the go is a definite possibility and it offered the potential to supplement the sirens and save lives and property.

The PART solution concept is based on a Android-GPS wireless application founded on established research around poor pick-up rates of ERV sirens by drivers and the hypothesis that a wireless-GPS type alert  will be available significantly sooner (longer distance apart)  which will enable timely corrective action by vehicular traffic. This hypothesis, along with other support hypotheses, was successfully tested. 

The PART prototype (after seven design iterations) can now successfully receive an early-warning graphic and voice signal on a hands-free mobile phone, 800m ahead leaving enough reaction time for timely evasive action. Eventually I visualise drivers having mandated, pre-fitted, vehicle dashboard equipment. 



About Me

I'm Viney Kumar, and I am a Year 9 student at Knox Grammar School in Sydney, Australia. I love learning new things at school and science and technology in particular, apart from chess, music, squash and debating. I am driven to ask questions about the world around me and how things work, going back to building LEGO sets (at 5) and growing exponentially when I discovered the world of Computer technology and Smart boards (at 8). Some of the scientists that have inspired me include Thomas Edison, due to his perseverance, Ada Lovelace for her world-changing work in creating the first computer program and Tim Berners-Lee for his work on HTTP and the World Wide Web. When I grow up, I wish to study Physics as I want to investigate more about the world around me and potentially apply technology in shaping new solutions for mankind. The inspiration for my project came from my love of technology, and my experiences of seeing ERVs helplessly caught in traffic in India while lives are property are at risk. The Google Science Fair is an opportunity to apply my creativity to a real-life meaty science problem and make a difference. It is also an opportunity to learn more, chase my dreams and compete with others globally. Winning would mean all this to me, and more as it would validate my passion and beliefs and help me take my project to the next level.                                                                                                                                                                                                                                                                                                                                                                                                   


Sirens are the primary alert mechanism used by ERV's worldwide but have severe limitations. Studies in the US show that a driver in a vehicle travelling at 60km/hour cant hear  the ERV's siren until it is about 100 metres away (even assuming no car-radio or air-conditioner noise) and in noisy conditions only within a distance of 1-2 metres.  A study from USDoT points out that only 26% of people could correctly determine the direction of an Ambulance from inside a closed car. Coupled with the added issues of traffic jams and high road noise there is a low probability of a driver in a developing country complying to an ERV siren

To make a difference in saving lives, property and fighting crime I decided to investigate the following question: How can I create a more effective early-warning system for Emergency Response Vehicles? I brainstormed a range of creative ideas and then hypothesised that an Android-GPS wireless solution should  be able to reach the target vehicle significantly quicker (i.e. leaving enough time to pullover) compared to published ERV siren response test statistics  which would prove my hypothesisI also hypothesised that the problem in decoding the direction of the approaching ERV would eliminated or greatly reduced by providing a graphic image of the approaching ERV and its destination. The above early-warning system would significantly increase the information available, the reaction time, and the ability and motivation of drivers to take timely action. 



My research focused on two areas (apart from tools): the effiicacy of ERV sirens and to check if there had been any work on alternate solutions.

The reaction time available for a driver commences from hearing the warning sound of the siren, which causes individuals to turn and search for the source, to ultimately locate it and to decide and take evasive action. However by this time it is all too late (particularly in a traffic choked situation in a developing country).

I found that there had been limited but sufficient research done showing why sirens were ineffective (though the research was done in the context of the accidents caused by ERVs). Stephen Solomon in the reference " Emergency Vehicle Accidents, Prevention and Reconstruction", 1999,  states that it was shown that an ambulance siren (115 decibels) can't be heard in a  target vehicle ( with no radio or air-conditioner noise) travelling at 60 kmph until the ERV is within 100 metres. He concludes that, with the airconditoner and radio on, the siren cannot be heard inside a target vehicle until the ERV is less than 1-2 metres away. There was also a 'marked reduction in the audible intensity of the siren between the two vehicles when the approach angle was altered by an intersection (90 degrees), this angled approach resulting in no siren sounds being audible within the passenger vehicle'. An EMS Services organisation (EMMCO) has established that for a ERV travelling at 25mph (40kmph), 'the warning time is reduced to 7 seconds'

The American College of Orthopaedic Surgeons have in a study demonstrated that 'the siren sounds of an ambulance proceeding at 100 kmph barely precedes the ambulance, so that vehicles ahead of it cannot respond to its warning'. The study showed that the distance for getting the attention of a motorist travelling at 100kmph to be within 2 meters of the ambulance's front bumper.  A research study by the US department of transport (USDoT) indicates 'that over a sirens effective frequency range, the average signal attenuation (through closed windows) resulted in a maximal siren effective distance of siren penetration of only 8 to 12 meters at urban intersections'. This study further suggests that ambulance siren tones are non-directional, and that only 26% of people in a closed car in the audible range, can tell the direction of the siren noise.   

This research and conversations with ERV staff shows that there is a definite need for early-warning systems for the detection of ERV's. These studies helped to shape my project, as they gave me information on the range of ERV sirens,  their limitations, and provided a baseline (in time and distance) to compare with my proposed solution. 

I also did a Google search for a wireless-GPS solution similar to what I had in mind, but it seemed none existed, which confirmed I could proceed ahead with my project.

For Research references pls. see Section 'Bibliography, References & Acknowledgements'


Method / testing and redesign

After completing the initial background research for my problem, and formulating my hypothesis, I used the Engineering Design Process to build the solution. 

1.     Identifying Functional Requirements

2.     Conceptual Design -  identifying, evaluating options & selecting a feasible design

3.     Program and Build prototype

4.     System Test and Redesign

Preparation For Testing

Starting with a very basic initial Requirement to generate an early-warning audio-visual signal that was an improvement on the siren, I then looked at three conceptual design options (described in the pack attached) and chose the option that would be easiest to prototype and demonstrate the concept, leaving the complexity of the final production delivery model to a later stage.  The final choice was to use the Android platform with the prototype deploying the locational sending and tracking programs respectively on two mobile phones simulating an ERV and a target vehicle. It was decided to develop the warning when the ERV was a distance of 500m and 200m (later revised to 500m and 800m) and to use an audio warning with a graphic image. 

I was now ready to commence Coding and Build. While programming the solution I started with the idea of using Google maps natively on the Andoid platform, but then moved away from it on the advice of my mentor who suggested using Web Maps API for a few good reasons. This worked well and made the application more accessible by non-android platforms.


Step By Step Process

1.  The program was unit tested (to remove bugs).

2. Simulation testing was done (through loading the programs into a mobile phone and checking for the generation of the warning messages and location statistics).It took seven iterations of redesign before my test results would confirm that the model was working properly.

3. After the program was working properly, I tested using a real-life simulation with two mobile phones in two cars, the first representing the ERV and the second the target vehicle or recipient. In these rounds I tested my two initial hypotheses by gathering measurements and comparing them against published research benchmarks for the existing system. My independent variable was the type of signalling method used. My dependent variable was the elapsed time between the first warning of approach and intersection ( in the second experiment it was accuracy). The experiment occurred on Clissold Rd in Wahroonga NSW.

Fairness and Safety

Most of the testing was done on computers, in a controlled environment which means that those tests were definitely fair. In the real life tests, I controlled variables such as the speed of the vehicle, the type of phone used, time of day, and the same testing route, which made the experiment fair. Safety was ensured by testing at safe hours, and making sure that there was one person driving and another looking at the phone, to avoid distractions.









Observations, Findings, Trends and Patterns

During Preliminary simulation testing I had the following results and made a few adjustments to my design:

I loaded the early iterations of the application onto Android phones to check the generation of the warning messages and location statistics.I found that the program generally functioned as it was supposed to. However, in the early versions, I noticed a huge unacceptable lag between the location change  warnings. To counter this, I adjusted the speech warnings on the program to 500 m and 800m instead of 500 m and 200 m.

I also added in a timer mode to update the ERV location at fixed (and smaller) time intervals instead of updating based on the change of location. This ensured that even if the target vehicle was stationary it still received an ERV update resulting in greater accuracy.

I also added a menu so that the choice of update mode (timer or location) can be made.

During final System testing I tested my two hypotheses: (1) to explore the warning reaction time (PART vs Siren) and (2) to test the ERV location reporting accuracy

Test 1 - Warning Reaction Time (Table 1)


The Time between the first warning and the point of intersection or crossover of the vehicles has been measured and averaged (see Column 800m to target). When the  available reaction time with PART (67 secs) is compared with the published  siren reaction times of only 7 seconds (with windows closed, radio loud and cell phone on), it indicates an increase in the available warning reaction time by 857%

If the Siren reaction times with only the windows closed and the radio at a medium level is taken as a basis of comparison, the PART result (67 sec) showed approximately 388% more warning reaction time than the siren (13.7277 seconds

Test 2 - ERV Location Reporting Accuracy (Table 2)


The purpose of the secondary hypothesis is to test how accurately the app pinpoints the ERV location. The address were observed to be 100% accurate while the latitude and longitude generally differed by 1/100th to 1/10,000th of a degree. Existing siren response tests indicate that only 26% of people inside a closed car can even tell the direction of an ERV while its proximity can only be gauged auditorily and visually from a short distance (generally always below 100 metres). This means that from the results of both tests I can deduce that PART has far greater location accuracy than a siren

To summarise: The PART system consistently outperformed the performance of a siren-only audio warning in both tests proving that PART was vastly superior to a siren based warning system.



Conclusion /Report


  • The PART system consistently outperformed the warning reaction time provided by a siren-only audio warning. The Warning reaction time showed an increase of between 388% and 857% (in a low noise and high noise environment respectively)
  • The address was observed to be 100% accurate while the latitude and longitude generally differed by 1/100th to 1/10,000th of a degree.Even when GPS latitude and longitude was out by a few degrees between the actual ERV location and polled location the reverse geolocation to street address remained consistent.

  • The siren reaction time of 7 seconds is a function of multiple noise pollution, the car internal noise and relative speed together. In contrast, the PART warnings appear almost instantaneously as they are a function of the relative speed of the ERV, the system update frequency and system distance setting of the first warning (which can be reset).The siren variables are uncontrollable while the PART variables are controllable.

Results vs Hypotheses

  • I thought that the results supported the primary hypothesis, which stated that there would be more time for a vehicle to pull over.

  • This was because the results indicated an 857% increase in the response (warning reaction) time when the windows were closed, the radio was loud and the ERV was travelling at 40 km/h.

  • My secondary hypothesis was also proven during the testing due to the dramatically increased accuracy from GPS as opposed to auditory signals.


  • I think my methodology and toolsets were sound and the results are completely reliable, though the accuracy of my results could have definitely been improved with more repetition (i.e. ensure greater consistency of test conditions). However, the results were accurate enough to ensure a fairly close range of results and clearly shows the trend.

Results vs Hypotheses

  • I thought that the results supported the primary hypothesis, which stated that there would be more time for a vehicle to pull over.

  • This was because the results indicated an 857% increase in the response (warning reaction) time when the windows were closed, the radio was loud and the ERV was travelling at 40 km/h.

  • My secondary hypothesis was also proven during the testing due to the dramatically increased accuracy from GPS as opposed to auditory signals.

Future Work and questions

  •  I have developed a significantly better model of an early-warning system for ERV's, providing vast improvements over the siren in the areas of available reaction or response time for vehicles and locational accuracy.

  • Timely cleared pathways will save lives, property and reduce crime. This will definitely have a positive impact on our world.

  • Some of the future work necessary includes running the application on a vehicle dashboard device, hands-free. More experimentation and testing would make it ready for the road. With suitable promotion and publicity with ERV organisations, transport authorities and vehicle manufacturers, wide acceptance is expected.


Bibliography, references and acknowledgements



De Lorenzo, R and Eilers, M Lights and Siren: A review of emergency vehicle warning systems, Annuls of Emergency Medicine, December 1991

Killeen, John.The theoretical and practical aspects of visual warning methods in use on emergency vehicles,  1999

Solomon, Stephen, Emergency vehicle accidents - prevention and reconstruction, 1999

US National Safety Council newsletter, April 1997


Internet References

1. Ambulance visibility research:

2. EMMCO study:

3. FEMA publication:

4. Queensland Regulations on ERV:

5. New South Wales Australia Regulations on ERV:

6. India article on ERV:

7. Science Buddies: Engineering  Design Process:

8.. Science Buddies: Engineering Design Steps

9. Functional Spec:

10. Android Advanced programming references:

11. Android toolsets:

13. Advanced series  tutorials (162 tutorials):

Android Developer API Guides



14. Stack Overflow




14. CODE PROJECT by Hamdy Ghanem


15. Wikipedia


16. The Android Developer's Cookbook: Building Applications with the Android SDK: Building Applications with the Android SDK (Developer's Library)




2 Application Basics: Activities and Intents 23

3 Threads, Services, Receivers, and Alerts 51

4 User Interface Layout 79

5 User Interface Events 117

8 Networking: Using Web Content 204

10 Location-Based Services 251


Accessing location data via GPS

Implementing threads, services, receivers, and other background tasks

Interacting with other devices via SMS, web browsing, and social networking


17. Sams Teach Yourself Java In 24 Hours, By Rogers Cadenhead


18. Head First HTML5 Programming

by Eric Freeman & Elisabeth Robson

  • Publisher: O'Reilly Media; 1 edition (October 18, 2011)
  • Language: English
  • ISBN-10: 1449390544
  • ISBN-13: 978-1449390549


Chapter five: Making your HTML Location Aware: Geolocation



19. Google Maps API

Geocoding Addresses with Google Maps API



20. GPS and GPRS

Free GPRS and GPS tracking software (top 5 free apps)   What is GPS real time tracking   How GPS devices pin point the location of moving objects

Location predictors and queries for tracking moving objects   more on how a GPS receiver works

Android metro view app    

21. Sydney bus location tracking applications     ArrivoSydney   TripView on apple   Tripgo



This project has been a fruit of labour inspired and supported by a number of individuals whom I have been fortunate to be associated with.  

Firstly I would like to thank Mr Warren Liang, Freelance Consultant & Software Designer for his mentoring and guidance throughout this project. He partcularly assisted in reviewing my work and providing feedback comments and recommending adjustments I could make during the Design and Build phases. 

Ms. De Ridder, Head of Science Faculty and the Science Department at Knox Grammar School, was a great source of support and encouragement and provided me an opportunity to share my work with the Science Club at school which led to many students being inspired to take up the challenge next year. She also gave me experienced and valuable advice on how I should conduct my research. 

I would like to thank my parents, for their invaluable support during the testing and for encouraging my work from concept to completion of  my project. I would also like to thank them for nurturing my ideas and putting up with my barrage of questions throughout my project. They have supported me in many ways including driving the family car during system testing - I would like to acknowledge all of their efforts from the bottom of my heart. I would like to thank my Grandparents, who have followed my project's development closely through their weekly conversations on skype from India and  for always believing in me and this idea.

Finally, I would like to thank Marcel Lima, my friend and schoolmate who has been a great sounding board and an eager volunteer during the project. His enthusiasm and friendship is something I will always treasure.






Back to Dashboard