SARTI
Semi-Autonomous Robotic Technology Integration
Technology
- Testing
Technology
Robotic Research is currently developing several technologies
that will continue to enhance the autonomy of ground robotics systems.
Some examples include:
Registration
Registration can provide:
- Viable extension to organic sensing to achieve
higher during speeds;
- Realistic Leader-Follower implementations where
GPS is not viable, such as under tree canopy;
- Map building capabilities that are currently not
exploited;
- Integration and fusion of different sensing modalities
into the platforms;
- Registration to apriori maps.
|
Ground vehicle sensor to ground
vehicle sensor registrations on MOUT scenario
|
Ground vehicle sensor to ground
vehicle sensor registration in Forested areas (GPS, INS, registred
data)
|
|
Aerial to Aerial registration
in Urban areas
|
Ground to Aerial Registration
|
Moving Obstacle Detection
- Needed in most scenarios (road, urban, off-road);
- Moving obstacle detection from moving platform
mostly studied for particular situations:
- Vehicle detection in highways or roads (CMU);
- Human detection from fixed camera (UMD).
The proposed approach:
- Is applicable to most range sensors (LADAR, stereo,
Radar);
- Moving obstacles from moving sensor;
- Low computational requirements (as opposed to optical
flow);
- Low false positives;
- Does not need to have all pixels collected simultaneously
to create an image (can be used with LADAR);
- Integrates seamlessly with current software
on autonomous ground vehicles.
GDRS color range LADAR image showing a person
Moving obstacle detection
For more information: info@RoboticResearch.com
|
