8
Projects

Projects. 1. Path planning – Input: 2D map, initial location, destination Output: waypoints (or false if there is no path). – Input: 2D map, initial location,

Embed Size (px)

Citation preview

Page 1: Projects. 1. Path planning – Input: 2D map, initial location, destination Output: waypoints (or false if there is no path). – Input: 2D map, initial location,

Projects

Page 2: Projects. 1. Path planning – Input: 2D map, initial location, destination Output: waypoints (or false if there is no path). – Input: 2D map, initial location,

1. Path planning

– Input: 2D map, initial location, destinationOutput: waypoints (or false if there is no path).

– Input: 2D map, initial location, number of destination points with priorityOutput: Path that visit all of the destination points in optimized order.

– Extending the algorithm to big maps, using algorithms such A*,D*

– Find the optimal map presentation (optimal grid size) – depending the map (dense obstacles or not).

Page 3: Projects. 1. Path planning – Input: 2D map, initial location, destination Output: waypoints (or false if there is no path). – Input: 2D map, initial location,

2. Fuzzy logic

– Build behavior that execute fuzzy controller (use open source library) for a specific robot (e.g. RV400).

– Detect obstacle – stop and report.– Robust controller for different type of robots– Implementing Edi Smukler’s work – any time

algorithm for ordering the controller roules.– Recognition impasses (number of choices).– Given waypoints – execute the controller on this

set of waypoints.

Page 4: Projects. 1. Path planning – Input: 2D map, initial location, destination Output: waypoints (or false if there is no path). – Input: 2D map, initial location,

3. Fence patrol

– Multi robot fence patrol include overlap– Removing robot – other allocate the task again– Using fuzzy controller to maintain distance (from

fence) and speed.– Different speed in different segments– Attending events – each event includes deadline

and time to execute.

Page 5: Projects. 1. Path planning – Input: 2D map, initial location, destination Output: waypoints (or false if there is no path). – Input: 2D map, initial location,

4. Fence patrol

– The same as previous but the algorithm that consider intruders.

Page 6: Projects. 1. Path planning – Input: 2D map, initial location, destination Output: waypoints (or false if there is no path). – Input: 2D map, initial location,

5. Circular Fence Patrol

– The same as previous

Page 7: Projects. 1. Path planning – Input: 2D map, initial location, destination Output: waypoints (or false if there is no path). – Input: 2D map, initial location,

6. Path planning – Multi robots

– Input: 2D map, initial location, destinationOutput: waypoints (or false if there is no path).

– Path corrections without conflicts between the robots paths.

– Dynamic missions -- allocating to robots missions, each mission has start point and end point.

Page 8: Projects. 1. Path planning – Input: 2D map, initial location, destination Output: waypoints (or false if there is no path). – Input: 2D map, initial location,

7. Navigation

– Input: robot location, destination pointOutput: robot arrive to the destination point (e.g. avoid obstacle from right if there is a circle – backtrack and take left).

– Ariel Felner’s article implementation.– Implementation of one of the well known

navigation articles .