We have written a report describing the realistic models and data structures we have defined for

• representing the map of a building being explored,

• describing the perception and mobility capabilities of the robots in a parametric form that is usable by our future software modules,

• expressing the motion plans to be computed by these modules, and

• defining a parametric model of a target's behavior.

http://underdog.stanford.edu/tmr/report.html

During the Second Quarterly TMR IPR meeting in Denver, we expect to demonstrate a software simulation of the next-best-view techniques we have developed for model construction (in two dimensions) by a team of mobile robots equipped with range-finders. We have also started refining these algorithms to work robustly in the presence of large uncertainty in robot localization.

We are continuing to work on our techniques for a team of robots to sweep an environment so as to find any set of targets that are hiding in it. The robots maintain a communication network at all instants during the sweep. We are planning software implementations of these techniques.

We have started developing some new target-tracking algorithms. The goal of our research is to ensure that multiple robots can track multiple targets. A principal feature of our algorithms is that we exploit the model of the environment constructed in the map-building phase to model the behaviour of the target; the targets being tracked will try to hide from the robots but we can use the current positions of the robots and targets in the map to predict where the targets are likely to hide. One key issue we are exploring is how we can dynamically change the target(s) each robot is tracking. We are also taking into account the fact that large localization errors might occur during the tracking because the robots might have to perform complex motions to track targets effectively.

The three Nomad SuperScout robots that we have purchased under the TMR grant have arrived. We are now experimenting with these robots and have started the process of porting our software to the robots so that we can effectively demonstrate our algorithms using real robots.

We have also received permission from the Army to purchase other essential equipment for our experiments (a laser ranging sensor and a laptop). We are now placing orders for this equipment.