Intelligent Control For Unmanned Vehicles

1
Intelligent Control For Unmanned Vehicles

• By
• Karl Murphy, Marilyn Abrams, Stephen Balakirsky,
  David Coombs, Tsai Hong, Steven Legowik, Tommy
  Chang and Alberto Lacaze
• Presented By
• Pete Gross

2
Topics

• Control Structure
• Detection of Obstacles
• Route Mapping
• Route Planning
• Vehicle Level

3
Intro

• Task is not easy to accomplish
• Vehicle Contains
• Electric actuators
• Kalman filter
• Inertial
• Dead Reckoning
• GPS Sensors
• LADAR

4
Control Structure

• Sensor modules process
• State of vehicle
• Surrounding environment
• Data updates state and surroundings
• LADAR
• Detects obstacles
• Placed in a map

5
Obstacle Detection

• LADAR used for obstacle detection
• 128 by 64 pixel map
• One per second
• 50 meter range
• Pixel classification
• Ground
• Obstacle
• Cover

6
Route Mapping

• Elevation and Features placed in map
• Map
• Oriented towards the north
• Scrolls as vehicle moves
• Features
• Integrated over time
• Confidence is computed
• False detections filtered

7
Software Architecture
8
Route Planning

• Shortest Path
• Two types of Paths
• Straight
• Curved
• Best Path Selected
• Removes blocked segments
• Finds path with least cost

9
Paths

• Curved
• Extends 20 meters
• Segments are spirals
• Straight
• 20 to 50 meters
• 4 Hz path replanned
• Small objects not detected

10
Vehicle Level

• Models the Terrain
• Plans Paths of up to 500 m
• Multi-layered Map
• Cost to Traverse Segments Computed
• Changes in Cost Change Vehicle

11
Conclusion

• Enables speeds of up to 35km/h
• Avoids Large Obstacles
• Still a Large Area for Improvement

12
Questions

• ?s