Book a free consultation
Location tracking Start-up

Navigating using AI. Vehicle positioning without GPS.

A start-up approached Exometrics to help develop a way for vehicles to navigate without using GPS. Here’s how we helped them.

95%

accuracy compared with GPS

The opportunity

A start-up had the novel idea of using environmental sensors, such as thermometers and barometers, to detect position. They approached Exometrics for help developing a powerful new positional system.

  • Thermometers
  • Barometers
  • GPS
Aerial view of a multi-level motorway interchange

Our solution

Millions of simulated paths. Deep learning that reads pressure, not GPS.

Simulation

A city road network, rebuilt as a simulation.

Our data engineers and data scientists created a simulation of a city road network using the Google Maps API and OpenStreetMap (OSM).

Deep learning

Trained on millions of paths. Pressure sensors instead of GPS.

We simulated millions of navigation paths and used them to train sophisticated deep learning models that learn to use pressure sensors instead of GPS.

The programming language and the deep learning framework behind the project.

  • Python
  • TensorFlow

Results

  • Our team developed a way to detect position using air-pressure changes measured by barometric pressure sensors
  • We used deep learning to create virtual pressure maps of a few cities around the world

95%

accuracy compared with GPS

Need a similar solution?

Get in touch for a free consultation with our experts.

Book a free consultation
View all case studies