4. Sample Collection and Return Analysis (WIP)

Narrative Analysis

Once the search locates a sample and the rover is positioned nearby the rover needs to verify this is a valid sample and, if it is, collect it, return to the starting platform, and deposit the sample. Then the search resumes.

A potential sample will be somewhere to the front of the rover within an area, based on our search analysis, about 3m wide and 2m deep and 1m ahead of the rover. In reality, the sample may be seen in an area deeper and wider than this, based on testing. This is not a problem.

With the search process having located a potential sample the rover turns toward the sample and drives forward while recentering its line of approach until the rover is a short distance, let's assume 0.5 m, from the sample.

At this point, the rover needs to verify the sample is valid. This means it is recognized as one of the ten samples and, also, that the robot has not previously collected this sample. (A rover in 2013 attempted to pick up the spot of sunlight shining through the canopy of leaves.) This last point is important because persistently attempting to collect a sample already collected can lead to an E-Stop of the rover. (If multiple robots are competing at the same time a sample collected by one of them will be replaced so it can be collected by another robot.)

A valid sample is then picked up by the rover. With the current lifter mechanism, the rover drops the lifter, drives forward to scoop up the sample, and then raises the lifter for carrying. The rover needs to verify that the sample is on the lifter. (A '13 rover thought it had the sample and returned empty handed to the starting platform.)

The sample is now taken to the platform. During the search, the rover has been tracking its location and bearing. At the start, the rover determined the location of the starting platform. Using its current location and the starting platform's location the rover can determine the bearing it must take to reach the starting platform.

During the drive to the platform, the rover must continually verify that the sample is on the lifter.

As a rover approaches the platform its path is guided by the home beacon. From 50m or more, the rover can see the red border marking the home beacon. Once the red rectangle is seen the rover can estimate the distance to the platform by measuring the size of the rectangle. The angle of approach can be estimated using the size of the sides of the rectangle as the nearer side appears larger and the ratio of their sizes indicates the angle of approach.

The actual destination location is a few meters in front of the platform. Obviously, the rover cannot mount from the rear since the home beacon is there. Nor can the rover approach from the sides since the current use of brushes to slide the sample off the lifter requires the rover to approach from the front. The home beacon also has large fiducial (figures) on the front and the back to make sure the rover's approach is correct.

Once the rover is positioned a short distance from the platform it needs to align itself with an assigned (virtual) slot. Each rover has a slot - left, middle or right -where it deposits samples. The reason for this is that collected samples are not allowed to touch one another during transport or when left on the starting platform. Since the rovers only collect one sample at a time there is no transport problem. To keep the samples separate on the platform there are three rows of brushes and three slots. This defines a grid of 9 spaces for samples. When (if?) the tenth sample is brought to the platform the rover can simply drive up and stop with it on the lifter.

The rover drives onto the platform, stops with the lifter behind a brush, drops the lifter, and backs away from the brush leaving the sample on the platform. Once it has backed up it raises the lifter, checks that the sample is gone, retries if the sample is still there, or backs completely off the platform.

Once clear of the platform the rover resumes the search by proceeding to an unsearched area.

Analysis Summary

The points in this section summarize the narrative analysis above. They omit some of the details of the specific 2013 implementation, such as the use of the lifter, which are design details and may change.

  1. Drive to sample, keeping it centered.

  2. Verify sample is valid, i.e. not garbage or light through a tree

  3. Verify sample not previously collected

  4. Pick up sample

  5. Attempt collection again if not successful

  6. Continually verify sample held

  7. Determine bearing to the starting platform

  8. Drive to offset point in front of the home beacon

  9. Detect home beacon, the approach angle, and the front / back of the beacon

  10. Turn to face the platform / beacon when at offset point

  11. Determine slot position for this rover

  12. Drive onto the platform in line with slot position

  13. Drop off sample

  14. Back off the platform