Swift Navigation and PolySync Announce Technology Partnership

New PolySync Core Driver for Swift Navigation’s Piksi Multi GNSS Receiver is Assisting in Paving the Way to a More Streamlined and Cost-Competitive Autonomous Driving Sensor Suite

Swift recently shipped its newest product, Piksi Multi, a multi-band, multi-constellation high-precision GNSS receiver, perfectly apt for the autonomous vehicle market. Ushering in a new era of precision GPS affordability, Piksi Multi represents a revolution in centimeter-accurate GNSS capabilities for the mass market. Swift Navigation solutions utilize real-time kinematics (RTK) technology, providing location solutions that are 100 times more accurate than traditional GPS, at a fraction of the cost.

PolySync provides Core, a development framework and runtime system for autonomous car development that turns algorithms, sensors and actuators into plug-and-play apps. Software built with PolySync Core moves seamlessly between computers, teams, vendors and continents helping automotive OEMs and Tier 1 companies to get to market faster with one codebase from R&D to production. Swift Navigation’s Piksi Multi GNSS receiver is the latest pre-built driver in a suite of popular autonomous driving sensors from PolySync. When purchasing the PolySync Core middleware, this Piksi Multi driver is free.

“Piksi Multi and PolySync are quite literally paving the way for autonomous vehicles: Piksi Multi is making centimeter-level localization more accessible at an attractive price point, while PolySync is easing integration of sensors within the autonomous vehicle sensor suite with its PolySync Core and pre-built drivers, such as the new Piksi Multi driver,” said Timothy Harris, CEO and Co-Founder of Swift Navigation. “This allows integrators to focus on their own navigation algorithms and simplify development of self-driving cars, lowering barriers to entry to emerging automotive teams.”

“PolySync Core supports a wide range of sensors in the self-driving car industry, including cameras, radar, LiDAR and IMU,” said Josh Hartung, CEO and Co-Founder of PolySync. “We are thrilled to add Swift Navigation’s high-precision, centimeter-accurate GNSS sensor driver to our line of autonomous vehicle sensor drivers.”

PolySync and Swift Navigation ran three test vehicles at the Self Racing Cars event at the Thunderhill West Track in Willows, CA on April 1-2, 2017: PolySync drove its vehicles autonomously using open source car control (OSCC), and Swift’s $595 Piksi Multi GNSS Receiver performed on par with a Novatel product, sold at roughly 10x the price.

For Swift Navigation, the Self Racing Cars event was a great opportunity to test Piksi Multi functionality in a dynamic, real-world environment with multiple testing scenarios in three vehicles, including one from Swift and two from PolySync and its Udacity Self-Driving Car Engineering Student Team.

How PolySync managed the GNSS data lifecycle in the autonomous driving application:

  • Piksi Multi calculated and sent data to PolySync Core
  • The PolySync Core driver parsed, processed, abstracted and published the data to the Core bus
  • The Udacity team’s PolySync node subscribed to the platform motion (GPS/IMU) data and image data from the PolySync Core bus
  • The Udacity team’s node used the image data and the GNSS/IMU data to perceive the environment and predict the steering wheel angle and throttle position to accelerate the vehicle or brake pedal position to decelerate the vehicle
  • The Udacity team’s node published control commands to the PolySync Core bus
  • The OSCC node subscribed to control commands from the PolySync Core bus and translated them to actuation commands that were executed on PolySync’s Kia test vehicle.

“We used the Swift Navigation GNSS signal as part of our neural network solution for getting around the track. It performed great and had higher accuracy than expected,” said Anthony Navarro, Udacity Student and member of Team Soulless at the Self Racing Cars event.

Swift Navigation’s own test vehicle captured two sets of data, both times with human drivers. The first data set was collected at low speed to illustrate precise mapping of road boundaries. This type of data supports applications such as auto-steering. The second data set was collected to capture more generalized product performance data at higher rates of speed. Similarly, the PolySync test vehicle collected data using its new Piksi Multi driver and a team from Udacity’s Self-Driving Car Engineering program utilized the new Piksi Multi driver during the autonomous driving session on the racetrack.