Reduce Vehicle Emissions with Predictive Maintenance by Pitstop Connect
Pitstop predicts vehicle emissions failures by analyzing connected car data with proprietary machine learning algorithms.
Pitstop has the capability to reduce the environmental damage caused by vehicle emissions. In Canada, GHG emissions from private vehicle operations has increased by 3% YoY and approximately 35% since 1990 . The Drive Clean program in Ontario has claimed to reduce 14.2Ktonnes in 2012. The Pitstop technology can increase this number to 21.3Ktonnes by regulating emissions in real-time and mandating vehicle repairs based on this data. The additional 7.1Ktonnes of reduction is calculated based on the premise that the savings would come from fixing 92424 vehicles immediately after failure. Ultimately this would save 2.96M litres of fuel simply by implementing a real-time mandate, where the user must fix the vehicle as soon as a failure is detected. This could reduce as much as 90% of passenger vehicle emissions in Ontario. This can all be achieved by leveraging the Pitstop Data Engine and focusing it on real-time vehicle emissions inspections.
How do you know that your solution is desirable to SMEs, and will reduce GHG emissions?
The Pitstop platform is constantly being marketed and sold to consumers, vehicle repair facilities and automotive OEMs. Currently, thousands of drivers in Ontario use the Pitstop platform and perform repairs at Pitstop certified repair centers. Within the Ontario market the Pitstop team is working to provide remote emissions testing capabilities to perform a Drive Clean inspection in real-time. The addition of enhanced real-time emissions monitoring algorithms and data visualization will enable commercial opportunities in selling to regulators and OEMs. Pitstop will own transparent unbiased insights on how vehicles actually perform in the field. This would enable Regulators & OEMs to mitigate issues such as the VW scandal. We already have contracts with dozens of dealerships, General Motors and Ford.
What actions do you propose?
Phase 1: Replicate Emissions test via mobile app & OBDII hardware (6 months, 2 backend engineers,2 mobile engineers)
Implement backend conversion functions the interpret the state of vehicle monitors and MIL light.
Implement Emissions Scan feature on mobile app. Design and mockup scan feature
Develop endpoints for mobile app to submit raw vehicle data and have the emissions test results returned
Write network calls on mobile app to submit and request data for emissions test.
Implement visualization of test results to vehicle owner on a pass or fail basis.
Phase 2: Quantify the impact of the emissions result (more than just a pass/fail) (8 months, 3 data scientists, 2 backend engineers)
Research the MOVES database(https://www.epa.gov/moves) from EPA
Determine how to leverage existing data profiles on Make Model Year emissions factors to estimate NOx, CxHy, CO2 output percentages based on OBDII input data
Integrate MOVES database into the Pitstop Data Engine (large project)
Prepare algorithms to estimate what the GHG emissions output is based on driving behaviour input (OBDII data).
Include algorithms on the PItstop Data Engine in order to output real-time estimates of how much NOx, CxHy, CO2 is being emitted from the vehicle
Develop endpoint to send data to mobile app.
Add visualization and enhanced results to the user experience within the mobile app.
Phase 3 : Increase emissions result accuracy of estimating NOx,CO2, CxHy (8 months, 3 data scientists, 2 backend engineers)
Purchase PEMS(Portable Emissions Monitoring Systems) devices to validate in real-time the accuracy of the MOVES estimation platform.
In accordance to conditions of inaccuracy (ex. MOVES is accurate when vehicle is in IDLE but not in motion) develop an action plan to collect missing data.
Determine the number of vehicles data needs to be collected on with the PEMS device.
Create a testing procedure based on terrain temperature and route used for data collection. Use previous research as ahttp://www.tandfonline.com/doi/pdf/10.1080/10473289.2003.10466245?needAccess=true
Equip various vehicles with the PEMS device and start recording data.
Train algorithms based on the newfound data collected.
Run validation tests to observe accuracy improvements in the model.
Phase 4: Showing the results data in a meaningful way (8 months, 3 data scientists, 2 mobile engineers)
Research how various GHG emissions cause specific environmental/human impact
Create a table that indicates mass of GHG relating to quantifiable impact on the planet. For example 100 hard accelerations cause 10 birds to die.
Implement this table into the Pitstop data engine
Develop endpoint to send data to mobile app.
Add visualization and enhanced results to the user experience within the mobile app.
Project with all 4 phases is prepared for launch at scale with a municipal, provincial or federal partner. With 8 engineers & 2 business analysts focused on the project.
Responding to Sandra Odendahl's comment on July 25th 2018-
Market: The business model relies on selling Pitstop technology to vehicle owners. How will you reach vehicle owners? Why would vehicle owners be motivated to purchase and/or use this technology? Is there a cost-benefit analysis associated with saving on gasoline that might help build a strong business case? What is the role for SMEs in this?
ANSWER: Currently Pitstop integrates with a list of 15 different connected car platforms used by millions of drivers. These are adopted by drivers for many reasons such as fleet telematics, in-vehicle wifi hotspots, business tracking and predictive maintenance. Pitstops strongest distribution models have been working with dealerships and service centers who provide their drivers with Pitstop. Ultimately as vehicles need service we make it super easy to schedule appointments and get discounts for drivers. This has proven an increase in revenue for these shops by 15% and uptake in visits by drivers by 47%. Drivers use the platform so that they can trust their mechanic, which is hard to do without the comprehensive data Pitstop provides. When a vehicle is maintained by Pitstop's Artificial Intelligence suggestions it becomes 100% emissions compliant. The second model of distribution is to fleets of vehicle looking to reduce their maintenance costs. We work with major fleet tracking companies and plug our software ontop of their platform to provide this additional value. Currently, this is being tested across 50,000 vehicles in North America.
Partners: Have you already established a partnership with Drive Clean in Ontario, or is their partnership with Stratford? What about partnerships with automotive garages and service centers? How will you do this? Will PitStop somehow help this type of SME to go low carbon?
ANSWER: Yes Drive Clean Ontario is partnered with Pitstop directly. The city of Stratford is also partnered with Pitstop directly. Currently, Pitstop powers over 40 service center & dealership locations with major contracts signed to power over 400 locations by mid-2019! This is by working directly with major dealership groups and service chains (Speedy Auto, Autonation etc.) By increasing these service shops profit we by nature reduce emissions by ensuring vehicles are maintained correctly. For Fleet Telematics companies that we work with including Fleet Complete and Verizon we are deploying our artificial intelligence to provide suggestions to the millions of business owners who manage large numbers of vehicles. The benefit to them is reduced breakdowns and cost of maintenance. After following Pitstop's Artificial Intelligence suggestions it becomes 100% emissions compliant.
Team: An impressive team, but who else do you need on your team to manage partnerships and outreach? What about the skills required to develop a strategy and tools for online sales and marketing?
Currently, Pitstop is actively engaging in a round of financing. This is to focus on bringing on the key business development staff including the head of diagnostics at the largest Tier1 supplier globally. By bringing in veteran automotive executives the credibility of the Pitstop platform grows substantially and major enterprise organizations are willing to leverage it. This has been proven through our sales historically which has been driven by individuals like Rosa Saturno, Greggory Garret and Joseph LaCross. We believe the traditional sales approach coupled with strong digital marketing is how we will continue to scale and penetrate the automotive industry.
Who will take these actions?
The Pitstop team has extensive experience with predictive analytics in the automotive emissions domain. The team has already developed critical algorithms used by General Motors and other global OEMs. Drive Clean Ontario has also signed off on a pilot project alongside Dan Mathieson Mayor of Stratford to test the platform as apart of their smart city initiative.
Where will these actions be taken?
While partnering with the city of Stratford and Drive Clean Ontario Pitstop will scale to 20,000 Ontario drivers within by June of 2019. This would result in a reduction of emissions testing costs from $35,573 per vehicle down to $8,479 per vehicle in the province of Ontario.
What are the proposal’s projected costs?
Challenges include the marketing and launch of the product into the Ontario market direct to consumers as well as through retail dealership channels. The technology is 80% developed and the last part of integration into the Drive Clean program is required before launching.
Cost of Development: $75000
Cost of marketing : $100,000 (to acquire 20,000 drivers)
Cost of Integration for Drive Clean: $25,000
Project Risks and Risk Mitigation
Risk 1: Successfully passing the Drive Clean certification process.
Mitigation: The testing criteria has been proposed and the technical requirements for the Pitstop platform is well known. Our roadmap involves addressing these technical needs.
Risk 2: Integrating into the MOVES database successfully.
Mitigation: Pitstop has developed a big data environment to manage large data sets and querying them efficiently in real time.
Risk 3: Achieving r squared accuracies of over 95%
Mitigation: The Pitstop data engine has already successfully achieved this for 5 algorithms developed with automotive manufacturers. It is a challenge we are well versed in solving
Risk 4: Selling the technology to vehicle owners
Mitigation: The current business model surrounds distributing the Pitstop technology to vehicle owners through repair facilities. As well as selling the technology directly to vehicle owners through online sales channels.
Once the solution is built and implemented describe a path forward for it to scale to other users/companies.
The Pitstop platform is already being used by thousands of vehicles across Ontario. This feature addition enables us to instantly turn on emissions reductions for these users. For future user adoption and scale we are partnered with strategic investors and companies listed below:
Dan Mathieson - mayor of Stratford
The City of Stratford has historically been the living embodiment of its motto, “Industry and Arts,” with automotive manufacturing and Stratford Festival tourism forming the core of the City’s economy. Our investments have earned Stratford international recognition as one of the Intelligent Community Forum’s Top 7 Smart Cities for three consecutive years. We are excited to be working with innovative companies to test new technology.
One of the companies that we look forward to working with is Pitstop – a Toronto-based telematics company that is revolutionizing automotive diagnostics through mobile applications and artificial intelligence. One of Pitstop’s current projects is aimed at reducing the environmental impact of carbon emissions. Pitstop has designed an IOT telematics device and a smartphone application that allows government mandated vehicle emissions tests be automated through this platform, which significantly reduces GHG emissions. This platform will be integrated with the ability to book appointments with car dealerships. The City Stratford is excited to partner with Pitstop to test this platform on vehicles in our community to help build a case study for implementation.
2. ROUSH - Tim Werner director of the emissions laboratory
Tim and his lab have been performing emissions validation cycles for car manufacturers across the globe. Their equipment and test facility is open for Pitstop to leverage when looking validating algorithms and insights.
3. Drive Clean Ontario - Remote OBDII program
Drive Clean Ontario has offered Pitstop the opportunity to integrate its software to perform regulated Drive Clean emissions tests in Ontario.
How will your solution lead to change on a larger scale over time (i.e. 3 to 5 years out)? How many businesses can potentially be affected by your solution?
Pitstop has a contract to apply its predictive algorithms for 300,000 vehicles within the US for the year of 2019 through the third largest dealership group. This emissions technology would instantly be added as a feature to the platform to impact the emissions output of 300,000+ vehicles.
What business and funding model have you considered for your solution to become sustainable?
The platform sells its diagnostic tool for $50 a unit through dealerships, fleets and service centers (covering initial costs). Every emissions test performed will reimburse $10 a test by the Ontario Drive Clean government. Additional revenue streams include a lead charge for every appointment booked to a dealership, reselling anonymized data about vehicle failures to car companies. Already Pitstop generates $30,000-$50,000 per service center location a year to gain access to these service leads.
What impact will the proposed actions have on reducing greenhouse gas emissions?
The major performance metric is cost per Ktonne of GHG reduction. The current Drive Clean Cost per Ktonne of CO2 reduction is $35,573 and with Pitstop this will be reduced to $8,479. If the Drive Clean Program were to increase the frequency of their inspections, the cost to run the program will also increase linearly. The reason is because every test as a set cost of $24.50, whereas with Pitstop $24.50 is spent once for 2000 tests. The key factors in getting the cost reduction is by providing an OBDII telematics device to a vehicle owner paired with the Pitstop smartphone application to enable real-time tests (upto every minute vs. 2 years)as well as providing instant feedback on how to fix these issues alongside connecting the vehicle owner directly to a trusted repair facility. The use of software, cloud computing, machine learning, smartphone and low cost hardware enables this platform to be accessible today.
What are other key benefits?
The goal of this platform would be to take emissions testing to the next level in which the exact value of PPM GHG emissions output is read per vehicle in real-time. This requires comprehensive data collection to benchmark vehicle sensor readings to GHG output. Machine learning models then are applied to the dataset to learn how vehicle sensors correlate to PPM output. This ultimately allows the Pitstop Data Engine to estimate the PPM of NOx, CO2, CO, CxHy in real-time as a vehicle is driving through software & an OBDII device. Never has this been done before without equipping a vehicle with thousands of dollars of additional hardware to the tailpipe of the vehicle in order to record these measurements, known as PEMS (Portable Emissions Measurement Systems).
About the Authors
Shiva Bhardwaj - project lead, is an Electrical and Computer Engineering graduate from the University of Waterloo. Shiva has specialized in Embedded Systems, OBDII Communication, and System design gained from working for companies like NVIDIA and Blackberry. Shiva served as a technician in his father’s repair shop since the age of 12 and was a certified emissions mechanic. He has already launched and distributed a company called ShockLock technicians across the globe. With Pitstop he has expanded the platform to capture 4 million miles of connected car data and 2 Million vehicle failure cases.
Christopher Mah- Data Scientist PHD in Mathematics (University of British Columbia)
Chris has extensive experience with data analytics and machine learning. He has a background in neuroscience and many years of experience in the interpretation and analysis of real time sensor data. He also has experience as a car salesman at a Toyota dealership, bringing some practical insight to his theoretical background. Chris works on Pitstops data science strategy and implementation.
Karol Zdebel - Mobile Engineer
Karol has a strong background as an android developer. With an extensive background developing software for various industries Karol has a deep understanding of mobile application distribution.