The SLAM is one of the major subsystem which involves the functionality of planning and mapping of the paths. As the robot is entirely autonomous, it plans for the paths to maximize the coverage of the arena while avoiding all the objects and extinguishing the fires, and produces the map of its coverage. The arena overview including the obstacles can be found in the appendices.

​

For the localisation, we designed the track of travel to be a zig-zag path. From the location of the robot being placed in, it should find one of the corners in the arena and set the flag of starting SLAM to true. To allow the robot to travel to the desired corner, we have programmed it to face in the direction of Northwest at start-up using the readings from the magnetometer. Problem with magnetometer is that the readings can be altered by external magnetic field, or any metallic object around the MPU. This problem will cause the direction of robot to face roughly in the direction of Northwest, while adjustments are made after ensuring the robot is facing precisely in the desired direction. After the desired corner is reached, SLAM can start and the zig-zag localisation technique is utilized. When the robot reaches the top wall of the arena a flag “StafeRight2” is set to true and false if on the bottom wall of the arena. This will make sure the zig-zag path is traversed correctly. When the localisation finishes, the MPU sensor data is extracted to the excel worksheet to generate the map.

​

The mapping is achieved by means of plotting the position of the robot at constant time intervals. The velocity vector of the robot is broken down into two components, velocity in x and velocity in y, where the magnitude is obtained from the robots linear speed and angle acquired from MPU reading. Next position of the robot is calculated by integrating the x and y components of the velocity vector over time and adding it to the previous coordinate or position. The coordinates are read through the serial monitor where it is saved into a text file at the end of the run. The text file containing all the coordinates are then exported to an excel spreadsheet and plotted using a line graph. The plotted line should indicate the complete path of the robot, and there should be some overlapping part when the robot carries out object avoidance.