Complete coverage relies on a path planning algorithm that will move one or more robots, including the actuator, sensor, or body of the robot, over the entire environment. Complete coverage of an unknown environment is used in applications like automated vacuum cleaning, carpet cleaning, lawn mowing, chemical or radioactive spill detection and cleanup, and humanitarian de-mining. The environment is typically decomposed into smaller areas and then assigned to individual robots to cover. The robots typically use the Boustrophedon motion to cover the cells. The location and size of obstacles in the environment are unknown beforehand. An online algorithm using sensor-based coverage with unlimited communication is typically used to plan the path for the robots. For certain applications, like robotic lawn mowing, a pattern might be desirable over a random irregular pattern for the coverage operation. Assigning directional constraints to the cells can help achieve the desired pattern if the path planning part of the algorithm takes the directional constraints into account. The goal of this dissertation is to adapt the distributed coverage algorithm with unrestricted communication developed by Rekleitis et al. (2008) so that it can be used to solve the complete coverage problem with directional constraints in unknown environments while minimizing repeat coverage. It is a sensor-based approach that constructs a cellular decomposition while covering the unknown environment. The new algorithm takes directional constraints into account during the path planning phase. An implementation of the algorithm was evaluated in simulation software and the results from these experiments were compared against experiments conducted by Rekleitis et al. (2008) and with an implementation of their distributed coverage algorithm. The results of this study confirm that directional constraints can be added to the complete coverage algorithm using multiple robots without any significant impact on performance. The high-level goals of complete coverage were still achieved. The work was evenly distributed between the robots to reduce the time required to cover the cells.

Coverage control has many applications such as security patrolling, land mine detectors, and automatic vacuum cleaners. This Thesis presents an analytical approach for generation of control inputs for a non-holonomic mobile robot in coverage control. Neural Network approach is used for complete coverage of a given area in the presence of stationary and dynamic obstacles. A complete coverage algorithm is used to determine the sequence of points. Once the sequences of points are determined a smooth trajectory characterized by fifth order polynomial having second order continuity is generated. And the slope of the curve at each point is calculated from which the control inputs are generated analytically. Optimal trajectory is generated using a method given in research literature and a qualitative analysis of the smooth trajectory is done. Cooperative sweeping of multirobots is achieved by dividing the area to be covered into smaller areas depending on the number of robots. Once the area is divided into sub areas, each robot is assigned a sub area for cooperative sweeping.