Assignment 2 description
In Assignment 2, all problems are set in a Maze of the same type as Assignment 1.
The unity project, that can be found here: Assignment_2.zip Download Assignment_2.zip
You will solve five problems (P1-P5). Each problem is setup in a different "Scene" of the project.
A demonstration video of the five problems can be found here Links to an external site..
Problem P1: Short range multi agent search
This problem is relevant for UAV search, vacuum cleaning or lawn mowing, where the sensor footprint limits the search area.
Problem P2: Long range multi agent search
This problem is relevant for ground robot patrolling, where occluding obstacles limit the search area.
Problem P3: Vehicle Routing Problem (VRP)
With fewer items to visit, there is no need to search the entire area, instead ordering the items becomes more important.
Problem P4: Formation Sweep
In this problem you will move your cars in a formation, together with a "Replay Car" moving on a given trajectory. There are several potential benefits of formations:
- With moving search objects it makes sense to joint up in larger sweeps to reduce the chance of objects entering already searched areas.
- If localisation errors are significant, moving in formation reduces sensor overlap.
Problem P5: Shooter coordination
Given fixed positions of a set of enemies. Move your agents so that they win a shooter match against the enemies. Consider moving in cover behind obstacles, and in coordination with other agents.
Turrets and Cars
All problems above are set in an environment where you control a set of cars in relation to a set of gun turrets, as shown below.
In the search problems (P1-P4) the turrets have very low "health" and are immediately destroyed when in range and sight of the car. In the shooter coordination problem however, the turrets have only slightly smaller health values, and careful coordination must be achieved for the cars to survive.
Note that the turrets have a constant damage per second, whereas the car guns give extra damage at closer range.
The details of the problem setup can be found in the table below:
P1 | P2 | P3 | P4 | P5 | |
Problem type |
Vacuum cleaner robots | Indoor security robots | Resupply robots (bread trucks) | Formation sweep | Shooter Game |
Long range car gun |
Yes | Yes | |||
Long range turret gun | Yes | ||||
Weak turrets | Yes | Yes | Yes | Yes | |
Leader car to follow |
Yes | ||||
# turrets |
100 | 100 | 13 | 50 | 8 |
Access to turret positions allowed |
No | No | Yes | No | Yes |
Maze | P1 | P2 | P3 | P4 (very sparse) | P5 (sparse) |
To bring focus to path planning, rather than car control, we limit the max speed of the car to 20 for all problems except P4.
Progress report
The same routine as last assignment. Use this Google Sheet Links to an external site..
Links
Links/references to possibly (you decide) interesting material (see also the slides of Lecture 3)
- Killzone’s AI: dynamic procedural combat tactics
- Classical Formation Patterns and Flanking Strategies as a Result of Utility Maximization Download Classical Formation Patterns and Flanking Strategies as a Result of Utility Maximization
- Approximation Algorithms for Terrain Guarding Links to an external site..
- Vehicle Routing Problem (VRP) Links to an external site.
- Constructing spanning trees for efficient multi-robot coverage, Links to an external site.
- Obstacle Avoidance in Formation Links to an external site.
- A Control Scheme for Improving Multi-Vehicle Formation Maneuvers Links to an external site.
- Cooperative Control of Mobile Sensor Networks: Adaptive Gradient Climbing in a Distributed Environment Links to an external site.
- Behavior-based formation control for multirobot teams Links to an external site.
- P. Toth and D. Vigo, “The Granular Tabu Search and Its Application to the Vehicle-Routing Problem,” INFORMS Journal on Computing, vol. 15, no. 4, pp. 333–346, Dec. 2003. (sorry, no link)
- Solving the Vehicle Routing Problem with Genetic Algorithms Links to an external site.