Week One

Today, we kicked off the project by understanding the rules of the game. Specific details such as the dimension of each square on the map, how the robot receives the signal during the game, and what are prohibited in the game. After the rules are clear, we start to discuss how can the problem be solved and how would the logic transform into code. Robot design is also considered, and we came up with a unique idea of creating a walking robot. Team member Sandro starts to experiment with all the components that are provided. At the end of today's session, we decided to name our robot "Walker" to remind people it's unique design. Chris Sherlock and Sandor Magyar are the logician and programmer, Ryan Wang and Qilin He are the robot builder and media manager.

Week Two

Today, Chris downloaded the EV3 python and getting ready for coding. The rest of the team were still trying to work out the walking robot design. There were many problems within the design, most importantly it was almost impossible to balance torque using limited Lego pieces. After considering the material and time we have, the team decided to let go of the "Walker" design and moved on to the traditional wheel design, but the name "Walker" remained.

By the end of the session, "Walker 2.0" dynamical system was fully assembled, and the sensor was the last thing we need on the robot. Initially, we thought the robot was able to perform the task with one sensor; however, after discussing with an instructor, we realized two sensors were required to help the robot navigate through the map.

Week Three

Today we made the Walker move and turn! The Walker was able to move back and forth, and it could manually turn in all directions. However, the color sensors were not functioning 100% at all time, for some reason, it was not able to determine the edge of the board every run. Chris was still working on the coding, and Sandro made progress on the algorithm almost ready to solve the wumpus world puzzle. Sandor helped on the coding by giving suggestions for variables that the program would need. Ryan and Qilin keep editing the website for the Walker, and updated information and journals on it.

Week Four

Today we focused more on the testing of Walker on the board. Before solving the wumpus world puzzle, we first need to control the robot to complete a figure 8. Chirs wrote the code for detecting the edge of each square and move in loops of motion. By the change of colors, the sensor readings would let the Walker would know its own location from one cell to cell. Additionally, in order to keep Walker stay in a straight line, Chris built the code to let Walker know how to adjust when one sensor reaches the line before the other. The program would control one side of the motor running until both sensors hit the line. After that, we discussed the algorithm and logic of the puzzle. Since Wifi USB port was not set up yet, We have not tested the robot in the game at this point.

Week Five

Today we got the chance to connect the robot with USB port and upload the code. The Walker was able to make turns and follow the loops of code successfully; however, we encountered a problem when the robot could not make an exact 90-degree turn. Because of the subtle inaccuracy of the output power in the motors, the robot made an angle slightly off 90-degree. This made the robot walked straight toward the corner of the square. Once the Walker was close to the corner, its color sensors were able to detect the two sides of the corner, and therefore it would keep moving forward diagonally. Chris debugged the code, and the Walker completed the figure 8.

Unfortunately, we were way far behind of solving the wumpus world. The competition was called off since nobody was ready to solve the problem. But Chris still tried to finish up the code. We ran the robot on the board, lots of bugs appeared, but overall the robot completed the movement.

Round Up

None of the teams was ready for the robot competition. As the result, it was called off. But we did make a significant progress today. The robot was finally connected with WiFi USB port. This enabled us to update codes in the robot. Next, we had a simple program to let the robot do a figure 8 pattern. During the process, we made a small change to the body structure. The wires that connect the EV3 brick and light sensors were originally placed between the EV3 brick and the motors. This made the structure look tight and aesthetic. It didn’t work because tightness resulted in a bad connection with the sensor.

The Walker successfully did a figure 8, but its walking pattern was more like a B shape. We then refined the code which made the movement functions more smooth by using individual motors instead of a steering pair. This helped the robot move more accurately and freely. Although the wumpus puzzle was canceled due of lack of time, we still wanted to give it a try because we would like to test if our movement logical works or not.

In the testing process, there were a whole bunch of bugs popping up when the robot walked moved into the second square. We tried to fix them by any mean but they didn’t work as perfectly as we thought. Overall, the robot still performed most movement logic correctly.