AirPro: A Case Study that Steps into Augmented Reality
AirPro Diagnostics is a leader in automotive diagnostics scanning and advanced driver assistance systems (ADAS) calibrations. According to Benefit from AirPro’s Total ADAS Calibration Solution (airprodiagnostics.com), in order to reduce accidents on the road, manufacturers incorporate advanced driver assistance systems (ADAS) into the vehicles they build. However, crashes can still occur, and it isn’t unusual for components of the ADAS to become misaligned or damaged during an incident—damage that can’t simply be inspected visually.
When repairs must be done, it is crucial to calibrate a vehicle’s ADAS properly. Returning systems to their original specifications is crucial for safe operation. If calibration is off-kilter, the driver will depend upon unreliable information. NLP Logix Data Analyst, Austin Seymour states, “The importance of vehicle cameras being properly calibrated is vital because it impacts features such as lane-keeping assist (LKA) and automatic emergency braking (AEB) technologies, making it difficult to drive reliably, leading to accidents, injury, and even fatalities.”
Augmented reality is used to add computer-rendered graphics to real pictures and videos. Using augmented reality technology, AirPro developed the AUGGIE (a complete, portable, wireless, indoor/outdoor product) which is used to calibrate car cameras, restoring them to working order. Josh McFarlin, AirPro Diagnostics VP of Operations, states, “The Auggie is a forward-facing camera solution that improves the standard process for camera calibration, through accuracy, as well as time and cost savings.”
AirPro Diagnostics partnered with NLP Logix to tackle the challenge of using machine learning and augmented reality technology to teach the Auggie how to calibrate forward-facing cameras found in vehicles. McFarlin states, “The standard process to calibrate vehicle cameras was to take a target and place it 15-20 feet out in front of the vehicle, while taking precise measurements in very specific environmental conditions, such as a perfectly level floor with no other obstructions. In order to do this, repair shops were dedicating valuable floor space to this process. Using computer vision and machine learning we wanted to scale down this process to create a user-friendly product for the end user.”
For the Auggie to properly calibrate a vehicle’s ADAS, the following steps were identified:
- Intrinsic Auggie Calibration – The Auggie’s camera must be taught how to see the world. This process uses photos taken at different angles to allow the Auggie camera to understand its view of 3D space.
- Extrinsic Auggie Calibration – This allows the Auggie to realize its own position in 3D space
- Car Camera Calibration – With intrinsic and extrinsic calibration complete, the Auggie can then be placed on the windshield of a vehicle, where it begins to employ object detection to locate the vehicle’s camera.
“Over a six-month period, NLP Logix and AirPro collaborated to develop a cutting-edge deep learning detection algorithm along with the advanced mathematical calculations required for calibration. The software components were then layered together along with the hardware into a sophisticated system that ultimately delivers the solution,” said Seymour.
AirPro Diagnostics offers revolutionary remote calibration technology through the Auggie. McFarlin states, “As business evolves, we continue to keep up with retrains and ongoing learning. Thinking outside of the box has been key. We are fortunate to have NLP Logix as a partner. Together we have been able to work through road mapping, planning, and any unforeseen elements as we blaze a new trail as experts.”
The Auggie success:
- renders a digital version of the static target in a miniaturized version
- provides dramatically more room to work
- provides AirPro customers with expanded capabilities and peace of mind
- works in concert with most calibration-capable scan tools on the market
Learn more about AirPro Diagnostics here.
Learn more about modeling and integration here.