ºÝºÝߣ

ºÝºÝߣShare a Scribd company logo

Applying cloud think infrastructure to real time vehicle parameter estimation

?Download as PPTX, PDF?
?
Dawn Chia
Dawn Chia

The document discusses applying CloudThink infrastructure to estimate vehicle parameters in real-time. It aims to identify mass, rolling resistance, aerodynamic resistance, and powertrain efficiency by linking vehicles to CloudThink and processing their data intelligently. This could help with tire pressure sensing and autonomous vehicle odometry to improve fuel efficiency, tire lifetime, distance estimates, and autonomous vehicle performance and costs. The document also discusses using CloudThink's open platform and secure API to estimate vehicle mass in real-time using an electric vehicle as a case study.

1 of 3
Download to read offline
Applying CloudThink infrastructure to real-time vehicle parameter estimation Erik Wilhelm, Sanjay Sarma, Lynette Cheah, Francisco Pereira Goal: to identify vehicle characteristic parameters (mass, rolling and aerodynamic resistance, and powertrain efficiency) in real-time by linking light-duty vehicles to the CloudThink infrastructure and intelligently processing their data. Applications: 1. Tire pressure sensing: underinflated tires reduce fuel efficiency and tire lifetime, yet loss coefficients can be used to under inflation algorithmically replacing the expensive tire pressure monitoring systems are only beginning to penetrate the market. 2. Autonomous vehicle odometry: understanding vehicle mass and rolling resistance will significantly increase the accuracy of the distance travelled estimate, improving the performance and lowering the cost of autonomous vehicles. Research Question: How can applying advanced parameter identification and machine learning technology to analyse and classify data from cloud-linked vehicles be used to improve the design of current and future transportation technologies relevant to Singapore?
Applying CloudThink infrastructure to real-time vehicle parameter estimation Erik Wilhelm, Sanjay Sarma, Lynette Cheah, Francisco Pereira Real-time Electric Vehicle Weight Estimation Mitsubishi iMiEV Secure, flexible API Highly accurate mass identification Signal Analysis Identified Mass - Driver Only Time-series Data - Driver Only 50 40 1800 1600 2 accel (m/s ) torq/10 (N-m) error (%) 30 20 10 1400 Mass (kg) Scaled values 2000 tau (N-m) spd (km/h) mass/100 (kg) force/100 (N) 60 Open-source Hardware Ident. Mass Cum. ave. True mass 1200 1000 800 0 600 -10 400 -20 200 -30 0 192 194 196 198 200 202 Time (s) 204 206 208 0 10 20 30 Event 40 50 60
Applying CloudThink infrastructure to real-time vehicle parameter estimation Erik Wilhelm, Sanjay Sarma, Lynette Cheah, Francisco Pereira CloudThink open platform design: Research Question: What is the most effective way to design an open platform for moving vehicle data to the cloud? To perform this analysis, the design choices made during CloudThink¡¯s development will be catalogued and codified such that a statistical analysis can be made.

More Related Content

Applying cloud think infrastructure to real time vehicle parameter estimation

  • 1. Applying CloudThink infrastructure to real-time vehicle parameter estimation Erik Wilhelm, Sanjay Sarma, Lynette Cheah, Francisco Pereira Goal: to identify vehicle characteristic parameters (mass, rolling and aerodynamic resistance, and powertrain efficiency) in real-time by linking light-duty vehicles to the CloudThink infrastructure and intelligently processing their data. Applications: 1. Tire pressure sensing: underinflated tires reduce fuel efficiency and tire lifetime, yet loss coefficients can be used to under inflation algorithmically replacing the expensive tire pressure monitoring systems are only beginning to penetrate the market. 2. Autonomous vehicle odometry: understanding vehicle mass and rolling resistance will significantly increase the accuracy of the distance travelled estimate, improving the performance and lowering the cost of autonomous vehicles. Research Question: How can applying advanced parameter identification and machine learning technology to analyse and classify data from cloud-linked vehicles be used to improve the design of current and future transportation technologies relevant to Singapore?
  • 2. Applying CloudThink infrastructure to real-time vehicle parameter estimation Erik Wilhelm, Sanjay Sarma, Lynette Cheah, Francisco Pereira Real-time Electric Vehicle Weight Estimation Mitsubishi iMiEV Secure, flexible API Highly accurate mass identification Signal Analysis Identified Mass - Driver Only Time-series Data - Driver Only 50 40 1800 1600 2 accel (m/s ) torq/10 (N-m) error (%) 30 20 10 1400 Mass (kg) Scaled values 2000 tau (N-m) spd (km/h) mass/100 (kg) force/100 (N) 60 Open-source Hardware Ident. Mass Cum. ave. True mass 1200 1000 800 0 600 -10 400 -20 200 -30 0 192 194 196 198 200 202 Time (s) 204 206 208 0 10 20 30 Event 40 50 60
  • 3. Applying CloudThink infrastructure to real-time vehicle parameter estimation Erik Wilhelm, Sanjay Sarma, Lynette Cheah, Francisco Pereira CloudThink open platform design: Research Question: What is the most effective way to design an open platform for moving vehicle data to the cloud? To perform this analysis, the design choices made during CloudThink¡¯s development will be catalogued and codified such that a statistical analysis can be made.