NIST’s new standard test method for evaluating the performance of optical tracking systems. These systems are similar to home gaming devices that track a player’s movements. Image: Roger Bostelman/NIST
Optical trackers are growing in importance, with robotics, VR, medtech, film production and flight training thirsty for improved technologies. Thanks to a new standard test, we could be about to get them.
Operating a piece of machinery that relies on tracked movement, special awareness and other robotic capabilities can prove immensely challenging, with a moving environment in need of constant tracking.
Human eyes work together, constantly scanning the environment, feeding information back to the brain to give us special awareness, depth perception and general vision.
Optical tracking systems attempt to replicate this, with cameras recording surroundings, and variations of LiDAR reading where objects are in relation to the tracking system.
In stationary scenarios, this often works well, but when moving – say an Amazon delivery drone, picking up and putting down objects from production lines, at doorsteps etc – this can prove tricky.
Pinpointing position in 3D space, often under varying conditions, is critical. Now a new standard test method to measure optical tracking systems’ accuracy has emerged to underscore the industry.
In the US, researchers have created a tool to measure how well an optical tracking system can define an object’s position and orientation, known as its ‘pose’ – within this, there are six degrees of freedom: up/down, right/left, forward/backward, pitch, yaw and roll.
“The tools are two barbell-like artefacts for the optical tracking systems to locate during the test,” said Roger Bostelman of the National Institute of Standards and Technology.
“Both artefacts have a 300mm bar at the centre, but one has six reflective markers attached to each end while the other has two 3D shapes called cuboctahedrons [a solid with eight triangular faces and six square faces].”
Optical tracking systems can measure the full poses of both targets by transporting them along two straight lines, up and down, and left and right. By turning three ways, the pitch, yaw and roll aspects are picked up, while the routes allow the systems to track the X, Y and Z axis measurements.
The setting is loaded with cameras and infrared readers, allowing operators to measure optical tracking systems’ accuracy throughout.
Bostelman said that the new standard can evaluate the ability of an optical tracking system to locate things in 3D space with unprecedented accuracy.
“We found that the margin of error is 0.02mm for assessing static performance and 0.2mm for dynamic performance,” he said.
