self localisation: establish own position with respect to a frame of reference
controllergenerate left velocity and right velocity and from that the angle of the robot
does that remind you of anything? a machine learning algorithm?
obstacle avoidance: short term goals
reinforcememt learning for controller: positive reinforcement
Occupancy map
artificial potential field
grid based techniques
Read the article on autonomous mining robots
how are obstacles detected?
fixed roads
base station has map and each vehicle has GPS
what is this virtual bubble?
artificial potential field
GPS denied (see notes on GPS)
if you put a new obstacle, then map regenerated and potential field
disadvantages?
After this chapter you should understand
A robot is an embodied intelligent system capable of
Unlike traditional software, robots interact with the physical world.
Examples include
Sensors
↓
Perception
↓
Localisation
↓
Mapping
↓
Planning
↓
Control
↓
Motors
What is around me?
Where am I?
What does the world look like?
Where should I go?
How do I move there?
Unlike software,
Robotics is about making good decisions with imperfect information.
Typical sensors
Typical algorithms
Probabilistic Robotics Thrun, Burgard & Fox
Modern Robotics Kevin Lynch
Robotics, Vision and Control Peter Corke