This is an early test of the “simple” object avoidance feature that will be included in Rover-3.3.
The Rover is an AION Robotics R1 with a Lightware SF10/B lidar.
There is also an NVidia TX1 running APSync which provides live video and telemetry to the ground station. MP conveniently provides a “radar view” on the GCS so the user can see in real-time precisely how far away object are.
LightWare has a wide beam “Flash LiDAR” (no moving parts) with a 16 degree field of view nearing production. This has 16 beams, each 1 degree wide, and it can identify the range and angle to an obstacle or multiple obstacles simultaneously. It has serial and I2C interfaces, in-field firmware upgradability, a maximum measuring range of 20m outdoors and drivers for a one or two axis servo scanner. It produces live radar images or can transmit pre-processed results as seen in this test video.
For a large surface like a wall the LW316 will detect it at 20m in outside conditions. Smaller obstacles produce less signal so they will be detected closer.
It should be around that price although the first units might be a bit higher until we ramp up production. Of course the LW316 is much easier to interface to the PixHawk or other controller and it works in the dark. Also, don’t forget that it can drive a low cost “Bob” scanning mechanism so that it can look around and prevent the rover from driver off a cliff or getting jammed under a low obstacle.
I’m still getting the build finished, but just wondering if there is a way of changing the behavior of the ‘simple’ stop. In your video the rover stops, but then begins again if the obstacle moves out of the way. Is there a setting so that the rover requires operator intervention to continue?