This article is about Spot's built-in obstacle detection system, which applies to the base robot. Payloads can provide additional obstacle detection capabilities (see for example Moving Object Detection).
Spot uses its perception system to automatically avoid collisions with obstacles. The perception system consists of five depth cameras: two at the front, one at the rear, and one on each side of Spot.
Notice
Although the camera images displayed to an operator may show an overlapping field of view, the perception system has gaps, especially at the rear corners of Spot.
Gaps in the perception system extend outward at a 45-degree angle from each hip. At Spot's front, the gaps taper and disappear about 2 meters from the robot as the front and side camera fields of view overlap. At Spot's rear, the gaps extend outward past the range of Spot's depth cameras.
Obstacles are mapped and remembered even when Spot’s movement brings the obstacle into one of the gaps. However, Spot may fail to detect:
-
Moving obstacles.
-
Obstacles that are hard to detect until Spot is very close. (For details on the limits of the perception system, refer to Nominal Operating Environments for Spot).
-
Obstacles that remain in a gap in Spot’s field of view during its entire approach path.
By default, Spot tries to keep a minimum distance of about 7.5 cm between itself and nearby obstacles. You can set an additional cushion of up to 50 cm using the controls in the Robot Controls panel (see Spot App Menus and General Controls). The obstacle avoidance cushion may prevent Spot from traversing doorways and other confined spaces.
Caution
Spot may collide with people or objects, even with its obstacle detection system enabled.
Operators and bystanders should assume that Spot may move unexpectedly at any time.
