Autonomous Observers
Stanford CS Robotics Lab
Motion Planning to Maintain Visibility of Moving Targets
We consider the problem of computing robot motion strategies that
maintain visibility of a moving target in a cluttered workspace. Both
motion constraints (as considered in standard motion planning) and
visibility constraints (as considered in visual tracking) must be
satisfied. Additional criteria, such as the total distance traveled,
can be optimized. The general problem is divided into two categories,
on the basis of whether the target is predictable. For the
predictable case, an algorithm that computes optimal, numerical
solutions has been developed. For the more challenging case of a
partially-predictable target, two on-line algorithms have been
developed that each attempt to maintain future visibility with limited
prediction. One strategy maximizes the probability that the target
will remain in view in a subsequent time step, and the other maximizes
the minimum time in which the target could escape the visibility
region. We additionally consider issues resulting from our
implementation and experiments on the mobile robot system.
Related Paper
Motion Strategies for Maintaining
Visibility of a Moving Target by Steven M. LaValle,
Héctor H.
González-Baños, Craig Becker, and Jean-Claude Latombe.
To appear in Proc. 1997 IEEE International Conference on Robotics and
Automation. (Postscript version)
Computed Examples for Predictable Targets
A simulation is shown of the optimal state trajectory for a tracking
problem. The observer strategy, including the initial position, is
chosen to maintain visibility and minimize the total distance
traveled. The target positions are shown as black circles, and the
target trajectory starts in the upper left. The observer positions
are shown with white circles. The line-of-sight is shown between the
observer and target at each stage.
This example shows an optimized observer trajectory that maintains a
specified distance range, and the initial observer position is fixed.
In this example, the target can moves at
twice the maximum speed of the observer. In the optimal trajectory of
the observer, there are only two stages in which the target is not
visible.
Experiments for Partially-Predictable Targets
Two Nomad 200 robots are used for the experiments. One is equipped with
a camera that is mounted on a rotary base.
A sample execution is shown of the on-line planner. The black disk is the
observer.