Team Four/Journal

Week 1

Electrical

The electrical design of this robot is fairly simply. It consists of two distinct wiring setups, one to transfer power and the other to transfer data. The wiring for the power transfer consists a 4-cell LiPo battery, two DC-DC converters, and two motor controllers. The DC-DC converters drop the [roughly] 16 volts from the battery to 8 volts needed for the Intel Edison and the 5 volts needed for sensor logic. The motor controllers are wired directly into the battery and supply the motors with power. The second wiring setup is for data transfer. This wiring is also very simple and consists data inputs from the multiple sensors on board. In order to keep our wiring organized we created our own plugs and labeled all of the wires.

Mechanical

The mechanical side of the first week also consisted of two separate parts. The beginning of the week was spent learning how to use the shop (including the laser cutter and 3D printers) and building the kit-bot. The kit-bot was given to us by the MASLAB staff and gives the software team a moving platform to work on while the final design is being created. This gives both the mechanical and software teams more time to create a better design. The second half of the week the mechanical team started working on the final design. This included prototyping a stacking-tower, a claw, and a carriage -- all of which will be implemented in the final design. This prototyping time was very useful as it allowed the team to learn how to use all of the shop machines fast and effectively which will be key in designing quickly.

Software

We started with getting the libraries up and running on the Edison. We tested and calibrated all of the sensors, including the motor encoder. Our drive loop for going straight with the help of the gyro works pretty well, but we’re still debugging the code to go towards a ball. For computer vision, we tried a number of things to detect the red ball. The one which proved most effective was the analysis of the distribution of values for all light intensities, as suggested in lecture. Currently, our program outputs the centre of a ball, its pixel area, its angle to the robots heading and distance.