Simulation & Coordination
of Mobile Robots

Object Reconstruction Using Large Quantities of Cube Robots

Consider placing a 3D environment scanner into an arbitrary room with everyday objects such as furniture and producing a 3D model of the room. This project investigates algorithms that allow thousands of small cube-shaped robots to autonomously rebuild a physical 3D representation of the scanned model.

The robots need to coordinate their movements, climb over one another and attach themselves together so that they can create a stable, Lego-like or Minecraft-like representation of the original environment. Simulation is used to investigate how such a large collection of robots could accomplish the task.

Coordinated Object Pushing by a Team of Robots

Teams of mobile robots can be useful when a single robot does not have enough power or agility to manipulate a large object. This project investigates algorithms for coordinating autonomous robots to push an arbitrary fixed-weight, 2D polygonal object through an environment containing polygonal obstacles.

The robots must position themselves efficiently around the object, follow a computed trajectory, rotate the object when necessary and continually reposition themselves so that the group can move the object smoothly and efficiently.

Weight Distribution for a Team of Lifting Robots

When a team of robots lifts a large heavy object, the load must be distributed appropriately. Good weight distribution can reduce the number of robots required while lowering the risk of overloading a robot or dropping the object.

We investigated algorithms for distributing the weight of a uniformly weighted 2D polygonal object by determining efficient robot placements based on the object's centre of gravity.

Hierarchical Robot Coordination

Individual mobile robots can be limited when a task requires coverage of a large area. We investigated ways of organizing robots into hierarchical teams in which small groups are managed by local coordinators, which are themselves managed by higher-level coordinators.

This structure allows global information to trickle down through the hierarchy while simplifying communication. If communication fails at one level, lower-level coordinators can continue operating using their most recent directives.

Simulation of Robot Mapping in an Unknown Environment

This project developed an algorithm that allows a robot with simple sensors, such as whiskers and a compass, to create a map of an unknown 2D environment containing obstacles.

The robot traces the borders of the environment and individual obstacles to produce polygonal representations. Because information is collected primarily while travelling along obstacle boundaries, theoretical bridge connections are introduced between polygons to establish relative positioning between objects.

Robot Colony Simulation

This project simulated a colony of robots that interact to perform various tasks. One task involved coordinated mapping: each robot generated a portion of an environmental map and the individual maps were subsequently merged into a global map.

PropBot Simulator

A fourth-year robotics course at Carleton made use of a limited number of PropBots designed for the course. This simulator allowed students to test robotic algorithms outside the laboratory before running them on physical robots.

The simulated robots were designed to closely correspond to the actual PropBots in their movements and sensors, providing a convenient environment for developing and testing software.