|
Improvement on Navigation |
|
|
|
|
Written by Chin Fung Lai
|
|
Sunday, 30 January 2005 |
| There are a few navigation technologies that are under
development, e.g. GPS Self-Calibrating Pseudolite Arrays by Stanford University
Aerospace Robotics Laboratory (ARL)
, landmark-based and star tracking navigation system by Carnegie Mellon
University. One of the ways to have an accurate and efficient navigation
system is to set up a full GPS system around Mars. However this is very
difficult and expensive, an alternative way is to have pseudolites (pseudo-satellites),
which use small ground-based GPS transmitter instead of satellites. The
main problem associated with using pseudolites is that it required centimetre-level
accuracy of the pseudolites' position which is extremely difficulty to
achieve when placing them on another planet. Hence ARL has come up with
Self-Calibrating Pseudolite Array (SCPA) to overcome this problem.
|
|
The idea of SCPA is to create an array with a few pseudolites and receivers
which uses GPS signal to determine their own location relative to the others.
ARL shows in their experiments that the rovers can travel within the array
with centimeter level of accuracy. We can see from the diagram the architecture
of the SCPA, the ability of the Wireless Ethernet and notebook used in the
experiment can be built on to the main station on the planet or the rover
itself. As we can see, the rover is limited within the array where the pseudolites
covered. To allow the rover to have a longer range, the pseudolites can
be migrated on to server rovers. |
| This method is being developed by ARL called the Leapfrogging
Multi-Rover Navigation . The concept is to have four rovers mounted with
transceivers working together with leap-frog mode navigation. This is to
allow one rovers to move at one time and its relative position can be obtained
from the other three stationary rovers. The movement of a complete cycle
of this movement is shown on the diagram below. ARL is able to achieve a
drifting error (caused by the geometric errors which accumulate) of less
then 0.5% of the total distance traveled (40 m). A video clip of the real
rovers' demonstration on this method can be seen on ARL website . |
|
This method results in more mobility and flexibility of
the rover. However, the algorithm is much more complex and it required
more than one rover. In the experiment, four rovers are used, if for example
one rover fails, can the other three rovers still operate? The answer
is that it can achieve the same level of accuracy with at least two stationary
rovers. That is to say this navigation system will still works with at
least 3 rovers working.
Follow this link to view the video of ARL testing the system: http://arl.stanford.edu/~rover/videos/leapfrog_ames.mpg |
|
|
Last Updated ( Monday, 31 January 2005 )
|
