Tuesday, May 20, 2008

Field Testing Report - Mojave Desert Studies Center

Dr. Christopher McKay, Planetary Scientist from NASA Ames Research Center. and the Northern California Chapter of the Mars Society (NorCal) Rover Team conducted a rover field testing trip to the Mojave desert from 21-23 March 2008. The purpose of the trip was to test the MAX 5 rover, a product of Senseta Incorporated, in field conditions planned for future Spaceward Bound activities and to introduce NorCal Rover team members to rover operations.

The aim of the field testing trip was to perform rover operations and gain insight into further planning of future operation at the Mars Desert Research Station (MDRS) in Utah and in outreach in the greater San Francisco Bay Area. It was also an opportunity for the NorCal Rover Team and Dr. McKay to evaluate the steps related to configuring and options for final purchase of the rovers as well as provide some time for NorCal Rover team members to gain insight into the obstacles which need to be resolved in order to meet the goals in the Spaceward Bound project proposal. Field testing was conducted at the Mojave Desert Studies Center (DSC) at Zzyzx.

The assembly operations and first part of testing took place in the hall of the DSC. First on a plain course later changed into one with artificial obstacles. The outdoors testing took place on several locations close to the DSC. The outdoors testing provided good ground for evaluating rover performance and rover impact on its environment. While the Mojave provides similar ground surface features to the target environment in the Utah desert it must be noted that ground coverage with desert plants is different and more challenging then the one for which the science rover is targeted.

Testing was subdivided into several categories:

Evaluate rover assembly and maintenance procedures in field conditions.

It takes some time to setup the rover and in a noisy wifi environment. Connections can drop between rover and laptop. This could use improvements in wifi robustness (software) and setup time (documentation). The communications issue could be a problem in a school but less so at MDRS. Once the rover is up and running it's fairly stable. Battery change has to be foolproof and simpler. We tested a rover with one camera, we need both of them operational as it will help in driving the rover. There's frame lag and this could use some research into compression and the practical limits of wifi. Battery lifetime could be improved by LiIon in future models.

Test rover maneuverability; in a controlled environment; in the field and by team members with different skill levels.

Driving the rover in the room was not that easy. Using the joystick as a controller is not intuitive at first. When driving there is a number of parameters on the screen which need to be tracked for a successful drive. Moreover in the room watching the rover interfered with the actual testing, training should probably conducted in adjacent rooms or with the operator turned with his back towards the rover. Certain basic rover moves should be introduced right at the start of training. Some of them could also be part of a standard library of moves, i.e. various turns and patterns. This would greatly aid the operator. Also keyboard support should be a high priority. We should come up with a standard rover course beginners should try to navigate.

Test rover capabilities in a desert environment resembling the MDRS Utah location.

During the afternoon we went to the desert and drove the rover in desert terrain. It was easier on that field than in the DSC room. Open ground is much easier for rover control. If both cameras where on it would help. A hammer was placed on the ground with the rover driving over it, not a problem for it at all. The impression is that stones half the size of the rover wheels or plants are not an obstacle. I would help to know the relative direction between the aim of the camera and the axial position of the rover, either by compass or an encoded mount which could be read by software.

Test the control software, laptops and communications operations with the rover.

The software runs smoothly on the laptop no issues there. However at times, perhaps due to wifi latency, the software does not respond in real-time anymore. In the desert it was not observed as a problem, but in a school or a surrounding with a lot of wifi traffic this could be a problem. A wifi sniffer and possibility a change of communication channels may be a solution. Also a wider angle on the camera field of view would be easier for rover control. Screen layout is important too, controls critical for driving the rover should be close to the camera view.

Field test the rover with a simulated science instrument load and evaluate its environmental impact on the desert surface.

Physical impact of the rover on desert terrain was studied. The goal was to find out how much load the rover can carry before disturbance of the ground is registered. The rover, weighing about 5 kg, was loaded first with 3 kg of weight and later with 6 kg without leaving signs on the ground. The loaded rover did also not leave signs on the terrain when going over obstacles. There is ample latitude in loading the rover with instrumentation without causing damage to protected environments. It's impact is less than that of a person walking the same terrain.

Evaluate practical communication range and power autonomy.

Through the tests it became obvious that communication latency could be an issue in controlling the rover, especially if it is compounded by interference from other wifi transmitters. There are some easy workarounds to mitigate this by switching channels or using higher gain antennas in a wifi crowded environment. Fortunately for science studies at desert location this may be less of an issue up to practical limits. The power autonomy is sufficient for meaningful rover sorties once the rover is in the hand of properly trained crews. At MDRS it would be advisable to prep one team member in rover operations ahead of time, this is something where good documentation, solid procedures and intuitive design come to bear. While LiIon may provide more power, it will not substantially change the autonomy of the rover as it stands today. Possible scenarios at MDRS may require the use of ATVs to bring the rover to more remote research locations. This rover evaluation was conducted in dry weather and terrain.

Test potential sciences packages

A Raman spectrometer and GigaPan camera assembly were tested in the field. The GigaPan processing is work intensive and although the results look promising more processing work needs to be done. It is not an instrument that will be easy to use by untrained people. The stitching software is far from perfect but is showing some promising results, processing times are long and it needs a lot of computing resources. Raman spectrometer testing was not conducted using the rover but using samples from the area. Results were satisfactory, issues regardin power and sample contact were discussed.

NorCal Rover Team, April 2008.
Photography (c) by Cornelia Knoepfel.

No comments: