Digging the Dirt: Autonomous Mining Robots  

Mining is a multi-billion dollar industry and a critical part of the Australian economy. Much of the day-to-day work on mining sites is dull, dirty and extremely dangerous, and so is ideally suited to robots! Researchers at the Australian Centre for Field Robotics at the University of Sydney are developing autonomous vehicles to dig and remove dirt from around mining sites. The diggers have to be able to load dirt, drive around the site without hitting any obstacles or mining workers, and then unload the dirt, all without human intervention. To accomplish this, sophisticated software is needed. The challenge in this project is to implement software to guide an autonomous mining vehicle.

A schematic of the region to be cleared is shown below. You can consider it as a grid of n x n locations. The robot starts from the Base and needs to travel to each pile of dirt (shown in brown), while avoiding obstacles (shown in shaded grey). It can move horizontally or vertically, one step at a time. In the example below we have a 7x7 grid with the Base in location (7, 5) and the dirt in locations (1, 5), (4, 2), (5, 2) and (5, 3). Your task is to write a MATLAB program to transport all of the dirt back to the Base.

1.'Reading'the'configuration'of'the'mining'area'' Your program should read the file dirtgrid.csv, which contains 25 x 25 values representing the configuration of the area. The free locations are represented with 0 and the forbidden with 1. Dirt piles are indicated by the value 2. Your program should ask the user to enter the Base location, e.g. using x and y coordinates. If the location is out of range or forbidden, the user should be prompted to re-enter them until valid values are entered.

2.'Experimenting'with'initial'strategies'' First implement a simple strategy for finding and transporting the dirt, starting from the Base location. The robot does not know where the dirt is located, and so should move up and down each row (or column) until it hits a block. If it can't move because of a block it should turn right or left, then continue. The robot can revisit the same location, but can't move to forbidden locations. It must check every cell in the grid for dirt. Assume it can transport any amount of dirt in one go. Your program should ask the user to set a limit for the maximum number of steps the robot is allowed (e.g. 1000 steps), so that you do not allow your robot to wander indefinitely. Your program should display each step the robot takes formatted as columns of x and y coordinates.

• If the maximum number of steps is reached, it should display Maximum steps reached
• When the robot identifies a dirt pile (say at x = 3, y = 11 after 30 steps) it should display: Dirt loaded from position (3, 11) after 30 steps
• When the robot has completed its search (say after 65 steps) it should display: Area completed in 65 steps

• When you have loaded dirt into the truck, you should change the value of that cell to 0.

3. 'Showing the robot's progress- Your program should ask the user to select a strategy: B for the basic strategy or X for your smarter strategy (if you completed the extension in Part 4 below - or maybe you have multiple advanced strategies) and the maximum number of moves to make. It should then count the number of times the robot visits a particular square as it tries to find and transport the dirt. You should visualize the grid and the movement of the robot as a colour map using the MATLAB imagesc function using the visits matrix. Finally, use the getframe and movie2avi functions to create a movie of the visits matrix after each step so we can see the behaviour dynamically.

4. 'Extension: Smarter transportation!  Implement your own strategy that improves on the strategy in Part 2. The aim is to clear all dirt with the least number of steps. As an example, you could assume the robot has sensors that can evaluate how close the neighbouring locations are to a dirt pile, or that it remembers how often it has visited a particular location. You might also add an extension in which the robot returns to base via the shortest path. Alternatively you could limit how much dirt the robot can carry at time. Of course, you are free to develop any strategy you like. The extension will be marked on the sophistication (and success!) of your solution. You should include a brief explanation of your extension as a comment in the program.