Some Misty I and some Misty II prototypes may generate inaccurate maps due to depth sensor calibration flaws. The depth sensors in later prototypes have received additional calibration.
Good to know.
is there a fast and reliable method to decide if there are calibration errors?
GET /api/alpha/slam/depthimage while directed at a wall and measure curvature of the depth image?
I will ask our team and post a method here if I learn anything. For now I can offer that we refined the calibration process and improved Misty’s visor to prevent the depth sensor from getting knocked out of calibration after manufacturing the first set of Misty II prototypes. So, the depth sensor calibration is likely to be a little off for most Misty I robots as well as for the earliest Misty II field trial prototypes.